Method of controlling operation of a multiple-station network

The present invention relates to a method of controlling a multi-station wireless communication network and a multi-station wireless communication network.

A communication network is known from the international patent application No. WO 96/19887 in which individual stations send messages to other stations using intermediate stations for transmitting the message data. In order to be able to send a new message to the network via one intermediate station chosen from among several possible, or to forward the message in the same way, each station usually has to contact several other stations at any time. In order to optimize the operation of this type of network, it is necessary to adapt the cooperation of individual stations according to established criteria to minimize competition or interference between stations, while maximizing data transmission rate with minimum transmission power.

The method according to the invention is characterized in that at least one calling channel different from the at least one data channel is defined, in which the stations transmit probes to other stations, to which any station responds, probe signals are received from other stations and the monitoring is carried out. probe signals transmitted by other stations are selected at intervals at each station and according to first predetermined criteria, Calling channel for transmission of probe signals to other stations, broadcasts broadcast probes from each station are transmitted at intervals on the selected calling channel, and other stations receiving broadcast probes from that station respond directly or indirectly with at least one station indirect, by which the availability of a given station is indicated as destination or intermediate stations and the direct or indirect responses of other stations to these broadcast probes are assessed according to second predetermined criteria, thereby identifying other stations with which the station communicates with optimally.

Preferably, each of the other stations receiving the probe signals from a given station modifies their own probe signals, thereby including data indicating the quality of communication between the station and them, and the given station responds to these data to change at least one parameter of its transmission. thus communicating optimally with the desired number of other stations in the network.

Probe signals from a given station are preferably transmitted, including data identifying other stations that the station has detected as available destination or intermediate stations.

Probe signals are preferably transmitted from a given station, including data indicative of the quality of communication between the station and each other designated station.

Preferably, the probe signals are broadcast as probe signals to all or more other stations.

Preferably, the probe signals, additionally comprising addressed probe signals, are addressed to at least one other station with which the station transmitting the addressed probe signals requests a connection.

Preferably, the addressed probe signals are transmitted more frequently than the broadcast probe signals.

Preferably, addressed probes, including a time information corresponding to the lifetime of data indicative of communication quality between the station and each other identified station, are used by the receiving station for the addressed probes when selecting other stations to join.

Preferably, probe signals including power gradient information corresponding to the combined transmit power required for each identified station to reach those other identified stations to which each identified station connects are used by the station receiving the probes in selecting other stations to combine.

Preferably, chasing signals are sent from the originating station to the destination station along multiple paths to the destination, thereby generating power gradient information used by network stations in routing for data transmission from the origin station to the destination station.

Preferably, a power gradient message is sent from the destination station to the originating station, containing data corresponding to the total power required for transmitting the data message from the origin station to the destination station via the optimal route.

Preferably, all network routed messages are fed with the power gradient information corresponding to the total transmit power required to achieve the individual

To stations via a message en route through the network, whereby an optimal route is selected for messages on the network.

Preferably, stations receiving probing signals from a given station respond by transmitting response signals to a given station, the given station compares the number of response signals received from different stations with a predetermined value, and changes at least one transmission parameter if the number of response signals is not corresponds to the second value until the number of response signals received by the given station corresponds to the predetermined value.

Preferably, a plurality of calling channels are defined with each calling channel apart from the first having a higher data rate than the previous calling channel, and selecting a different calling channel having a different data rate than the previous calling channel according to a second predetermined criteria, if the response signals do not correspond to a predetermined value.

Preferably, when the first predetermined criteria include the calling channel data rate and / or the calling channel transmit power, the calling channel is selected according to the highest available channel data rate and / or the lowest available channel transmit power.

Preferably, when the second predetermined criteria include the calling channel data rate and / or the calling channel transmit power, the other calling channel is selected to have an increasingly lower channel data rate and / or an increasing channel transmission power.

Preferably, the predetermined value, which is compared with the number of response signals, is calculated to correspond to the desired number of neighbors available to a given station as intermediate or destination stations for the station to communicate optimally with the desired number of other stations in the network.

Preferably, a plurality of data channels are defined, each data channel except the first one has a higher data capacity than the previous data channel, and each station transmits data to its neighbors on the selected data channels after determining the availability of these neighbors.

Preferably, when the data channels correspond to the individual calling channels, a data channel is selected for data transmission that corresponds to the selected calling channel.

Preferably, when multiple data channels correspond to a single calling channel, the data channels are monitored for activity by the stations and the data channel detected as idle is selected by the station wishing to send data, thereby optimizing the use of the data channel between the stations.

Preferably, the probe signals are transmitted by each station on the calling channels containing information indicative of the intention of a given probe transmitting station to move to the selected data channel, which is then marked as active, for another station to communicate with the station on the selected data channel.

Preferably, the probe signals are transmitted regularly by stations trying to establish communication with other stations, and the other stations receive probes corresponding to a random number of probe signals, the random number being equal to or less than the number of probe signals transmitted.

Preferably, each station controls the interval between the transmission of the probes by the probe clock, and the probe clock determines the interval between successive probes that is longer than the duration of the probe signal, and transmits response signals at intervals between successive probes.

Preferably the interval between transmitting successive probes at each station varies according to whether or not the station has data to send, and the probe timer determines a first, relatively short interval between successive probes when the station has data to send, and a second relatively long interval between successive probes when the station has no data to send.

Preferably stations are designated as valid and these stations send probes containing data identifying them, and other stations receive these probe signals, modifying in turn their own probe signals to include data identifying important stations so that even remote stations important sources obtained this data.

Preferably, gateway stations, certification certification stations, and sometimes originating or destination stations are used as the designated valid stations.

PL 193 013 B1

Preferably, you distribute the updated software for controlling the stations by downloading the updated software to the selected station and you distribute the parts of the updated software to the other stations until each other station has the complete updated software.

Preferably, the updated software is distributed in update blocks containing version data and block number data to compile the updated software from the plurality of received update blocks.

Preferably, the date and time for applying the updated software are indicated by at least one of the update blocks including time data.

The communication network of the invention is characterized in that each station comprises a processor for determining at least one calling channel different from the at least one data channel for transmission by the probing signal stations to other stations for response by any station, receiving probe signals from other stations and checking the signals. probes broadcast by other stations and selection at periods, at each station and according to first predetermined criteria, the calling channel to transmit probes to other stations, a transmitter to send probes to other stations at slots on the selected calling channel, and other stations that receive probes from the station respond directly or indirectly through at least one intermediate station. to indicate the availability of a given station as destination or intermediate stations, and a receiver operating in conjunction with the processor to evaluate the direct or indirect responses of other stations to the probe signals according to second predetermined criteria, to identify other stations for optimal communication with the station.

It is an advantage of the invention to provide a method of controlling a communication network and a multi-station communication network that regulate communication between the stations to optimize the operation of the network without undue competition or interference between the stations. The invention adjusts the operation of each station so that, at any given time, stations can send data or receive data from a number of neighbors at the highest possible data rate but with the lowest possible transmit power, thereby reducing interference with other stations.

The invention is illustrated in the drawing examples in which Fig. 1 is a diagram of a multi-station communication network showing how a origination station transmits data via a plurality of intermediate stations to a destination, Fig. 2 is a flowchart showing a channel adaptation operation and method. according to the invention, Fig. 3 is a flowchart showing the software update mechanism according to the invention and Fig. 4 to 7 - block diagrams of a communication network for practicing the invention.

Figure 1 shows schematically a communication network having a plurality of stations each having a transceiver for transmitting and receiving data from any station within the coverage of a radio network or other network where the user stations communicate with each other via intermediate stations.

The network users are the subscribers who are charged with the bill for using the network. Alternatively, these types of networks are used by security forces such as the police or the military.

In Fig. 1 originating station A connects to five nearby stations B through F and sends data to destination O via intermediate stations B, I and M.

To maximize network performance, each station has a number of neighbors to connect to in case the station needs to send or receive a message. On the other hand, if a station is sending data to a selected neighbor, it is desirable that the transmission causes minimal interference with other stations, otherwise the resulting competition between stations reduces data throughput on the network.

A communication network includes a plurality of stations attempting to connect via multiple channels having different frequencies, carriers, coding, e.g., different extension codes, different antennas, time slots, and so on, or any combination thereof. To optimize channel reuse, stations try to maintain a limited number of direct neighbors, typically five. A neighbor is defined as the station to which the station communicates.

The station limits the number of visible neighbors by changing the transmission frequency, code changing the PN sequence, increasing the data rate and reducing transmission power. Other stations congregate at fixed calling channels where they find stations communicating using a probe signal. Once another station has been found and one of the stations has data to send, the stations may move to the less used data channel.

When there are a number of stations nearby, they will eventually use higher bit rates and lower transmit power. From time to time, stations check the calling channels with less

Data rates to assist any remote stations that cannot use the higher data rates. In the same manner, a station that is on a lower rate calling channel periodically checks all rates greater than its current rate to find possible sets of higher rate stations.

Figure 2 shows a flowchart where several different channel adaptation clocks according to the invention operate at a given station. Each timer is checked sequentially, however, there may be separate processes or events that are checked simultaneously. The various channels and their associated clocks will be described below.

Each station sends Probes at regular intervals determined by the Probe Timer to try to find other stations. If any station receives a probe, it responds randomly. A random response is typically one response for every one to four probes received. In other words, other probe receiving stations respond to a random number of probe signals with the random number equal to or less than the number of probe signals sent. This prevents competition with other stations nearby.

The time between the probes as determined by the probe clock is used to respond to other stations on each one to four probes received. Since the time between the probes is longer than the duration of the probes, the station responds with a small data packet. However, the maximum length of the response packet is not longer than the normal probe clock period.

Each station changes the Probe Timer randomly between probe transmissions to avoid collisions with other stations. If a station starts receiving a transmission from another station, it will enter a new Probe Timer period.

When a station has data to send, it sends the probes at intervals proportional to the rate of transmission used by Probe 1. However, if a station has no data to send, it typically uses a period five times longer than Probe 2 when it has data. This gives the stations that have the data more ability to communicate. Since other stations reset their probe clock every time a transmission is detected, they may never poll if they have no data to send. Thus, each station forces the probe after a time that is at least five times the normal period.

A station that has data to send sends probes five times as often as a station with no data. A station with no data resets its probe clock each time it hears another station's probe. Since the station without data uses a longer period, transmission is never possible. Thus, the station with no data resets its probe clock each time it hears another station, unless the last transmission was before Probe Timer 2, in which case it sets the Probe Timer to Probe Timer 1. The data station also uses the Probe Timer period. probe 1, therefore a station without data obtains the possibility of sending a probe signal. After sending the probe, it reverts to using probe timer period 2.

A broadcast probe, sent by a station without data, shall be addressed to all stations. So every station responds. However, if a station has data to send, it will convert broadcast probes to probes addressed to stations for which it has addressed probe data. Addressed probes pass sequentially through all identifiers for which the station has data. Only the station addressed by the Addressed Probe responds, and no other station responds, so the addressed station always responds immediately.

When turned on for the first time, the station starts polling with the lowest transmit power and highest baud rate on the calling channel. This avoids interference with other stations nearby.

Each time a different station responds to the probe, the responding station is counted as a neighbor. If the required number of neighbors is not obtained within a predetermined period of time set by the adaptive clock, the station increases the probe transmission power by 10 dB and further increases the probe transmission power until it obtains the required number of neighbors. If it reaches its maximum transmit power before reaching the required number of neighbors, the station moves to the next lower transmission rate, i.e., the previous Calling Channel, but remains at its maximum transmit power. It continues to lower the data rate until

The PL 193 013 B1 will obtain the required number of neighbors. If it never reaches the maximum number of neighbors, it remains at the lowest baud rate and maximum transmit power.

Each time a station changes to a different Calling Channel, it will reset the adaptive clock, also if it changes its probe transmit power.

In a network of mobile stations, stations are in motion all the time, therefore the number of neighbors is constantly changing. If the number of neighbors exceeds the required number, the station starts increasing the rate on the next calling channel. It continues to increase its transmission speed until it no longer exceeds the required number of neighbors. If it reaches the maximum number of neighbors, it will begin to lower the probe transmission power by 10 dB until it reaches the minimum transmission power or it no longer exceeds the required number of neighbors.

Each time a station changes its rate, it switches to a different calling channel, thus avoiding the interference of lower transmission rates with higher transmission rates.

When a station responds to another station on the calling channel, it limits the length of the data packet to the probe clock period, thus preventing other stations from probing during its reply. If the responding station has more data to send than fit in the small packet, it indicates in the packet header that the other station must move to the specified data channel.

There are a number of specific data channels for each Calling Channel. The station requesting the change randomly selects one of the available data channels. When another station receives the request, it immediately moves to that data channel, where the two stations will continue to communicate until neither of them has data to send or until the maximum dwell time set by the data clock has elapsed.

When a station transitions to the data channel, it loads the data clock and remains on the data channel for as long as the data clock will allow. When the data clock expires, the stations revert to the calling channel and begin probing again.

For each Calling Channel, there are a Previous and Next Calling Channel, in addition to the Lowest Rate Calling Channel that only has the next Calling Channel and the highest Rate Calling Channel that only has the previous Calling Channel. As the number of neighbors in an area increases, the stations move to higher rate calling channels. However, stations that are further away from this area will not have as many neighbors and therefore remain on the lower rate calling channels. In order for stations to stay connected, they must check the previous and next calling channel at regular intervals.

The check clock is set when a station first enters a calling channel. The check clock period is proportional to the rate of each calling channel for check clock 1. When the check clock expires, the station first determines whether it is currently checking or if it still needs to check. If she did, it would move down to the previous Calling Channel from the one she was controlling. If the station did not control, it would go to the highest calling channel. This channel becomes the current control channel.

When the station is on the Control Channel, it resets the Check Clock. Check clock period 2 is significantly shorter than the period used to access a Calling Channel. When the Check Timer expires, the station goes to the previous Calling Channel and becomes the new Check Channel.

The station continues in this manner until it reaches the initial calling channel. At this point, it goes down one channel below the Calling Channel. If there is no previous Calling Channel, it will terminate the check and reset check clock l to a larger value. If there was a Calling Channel, the station would repeat the normal checking steps. After this last check, it returns to the original Calling Channel.

That is, the station periodically checks all calling channels above its current calling channel and one channel below its current calling channel. The upper channels take less time to check as they will usually run 10 times faster than the current channel. However, checking the calling channels below the current one takes a certain amount of time and therefore only checks one level down.

Not only does Calling Channel Control keep stations on different Calling Channels in contact, it also helps stations on lower calling channels spot more neighbors and thus helps them transition to higher calling channels.

PL 193 013 B1

For each Calling Channel, the baud rate is typically 10 times that of the previous Calling Channel. From the bit rate for the calling channel, the duration for all clocks is calculated using the multiplication factors. The absolute values ​​of the factors are listed below, but these values ​​are given as examples only and may vary considerably. Additionally, the correct values ​​change dynamically with changes in network traffic and the number of stations.

<td>Clock</td><td>multiplication factor</td><td>Example for 8k speed</td>

<td>Probe clock 1:</td><td>10 x probe packet time (data in Tx queue)</td><td>300 ms</td>

<td>Probe clock 2:</td><td>5 x probe clock 1 (no data in Tx queue)</td><td>1500 ms</td>

<td>Adaptive clock:</td><td>100 x probe clock 1</td><td>30,000 ms</td>

<td>Data clock:</td><td>5 x probe clock 1</td><td>1500 ms</td>

<td>Check clock 1:</td><td>30 x 1 probe clock (currently without Control)</td><td>9000 ms</td>

<td>Check clock 2:</td><td>2 x probe clock 1 of the current control channel</td><td>60 ms</td>

<td></td><td></td><td>(80k check)</td>

In a network using the method according to the invention, the following limitations / options are usually implemented:

A station may never connect at a baud rate lower than the current channel's baud rate, but may connect at a higher baud rate if the bandwidth permits.

A station never answers another station whose received signal-to-noise ratio falls below the required level. However, it does respond if there is no previous channel to go down to. For example, if it is on the 80 kbps channel, it does not correspond to the station whose received signal-to-noise ratio is inappropriate. This forces the other station to go down to 8 kbps. However, if it is already at 8 kbps, there is no other channel to go down to and therefore it responds.

When switching channels, a station must always wait for the Probe Clock time before polling so that its probes do not interfere with transmissions from other stations.

When responding to a certain station, the packet length is always less than the delay of the probe clock to prevent response packets from being interrupted by hidden terminals. When station A responds to the probe from station B on the calling channel, the length of the reply packet measured in time is less than the period of the probe timer 1. This is to prevent station C from transmitting in the reply packet. This happens when station A and station C hear each other, but station B and station C do not hear each other. Station C resets its probing period when it detects that station A is sending a probe. Since it cannot hear station B, it does not reset its probe period when station B responds, so it sends a probe after the probe timer expires. The probe from station C would damage the reply packet from station B if it were longer than the probe clock period. However, if the response packet is shorter, it would reach station A without damaging it before station C sends a probe.

If a station has more data to send than is sent in the probe timer period, the station sends when it can and requests that the other station go to the data channel. Thus, two stations should not connect for more than three transmissions, i.e. consecutive reply transmissions on the calling channel. For example station 1 Tx probe -&gt; station 2 Tx data -&gt; station 1 Tx data or station 1 Tx data or station 2 Tx data request a data channel if there is more data to send.

The probe clock period is not always the same, it always has a random change added, usually 50% of the clock period. This ensures that a number of stations do not transmit simultaneously and therefore will never receive each other. For example, at 8 kbps, the probe clock for data in the Tx queue typically varies in the range of 300 to 450 milliseconds.

When a station has no data to send, it typically tries to reach five of its neighbors. However, if it has data to send, it tries to get more neighbors, typically 15. The station would probe at a faster rate and therefore is more likely to get more neighbors. If it does not get more neighbors, it increases the transmit power. In higher intensity networks

In view of the traffic, the number of neighbors required cannot be increased as this would result in too much competition.

Stations can monitor other stations moving to data channels. This is an indication of the available data channels.

A second receiver can be used to scan the data channels to find clean data channels with little background noise.

When the station probes, it provides information in the header of the probe signal data packet as to which of the monitored data channels are clean. When the next station responds and wishes to switch to the data channel, it then combines its own information with that of the other station to make a better choice as to the data channel used.

When a station transmits data, it does not use a power level much higher than that used for probing. For example, if a station probes at 0 dBm to obtain the required number of neighbors, it cannot, for example, respond at 30 dBm as this would cause interference with more distant stations. The amount by which the power used for data transmission exceeds the probing power is a parameter set for the entire network.

Noise and traffic can be monitored simultaneously on multiple calling and data channels using multiple receivers.

Probe packets and data packets may be sent simultaneously on multiple calling and data channels using multiple transmitters.

The network may have more than one calling channel for one rate and multiple data channels for one rate.

In a first embodiment of the invention, two types of probe signals are used. The first type of probe is a broadcast probe, which lists the best stations that a given station detects. The number of stations listed is usually in the order of ten. Each station in the list has a number associated with it that indicates how well the probing station has heard the stations in the list. The second number indicates how well the stations in the list detected the probing station, it is taken from the broadcast probes of the other stations. So the third station knows immediately how well the probing station has heard the other station and how well the second station has heard the probing station.

This arrangement eliminates the need to respond to probes because when a station hears its own identifier in the probing signal, it knows that the probing station hears it and with what quality. When it sends its own probe, it appends the ID of the station it just heard. The other station hears its own ID and closes the loop. So by simply sending out probes each station close to the other stations knows what stations it can hear and how well. By monitoring other probes, it also knows which stations the probing station detects and with what quality. This information is then used to set the number of neighbors.

Each broadcast probe from each station lists all the stations it has detected. Since all stations that can hear the probes see themselves in the list, the station sending the probes does not need to do so often. In the sounding method of the first embodiment, a station needs to get a reply from every other station to know if it can hear it. Now all neighbors know that the probing station can hear them because they are in the list. When they in turn send a Broadcast Probe, all other stations will know they have been heard if they appear in the list.

The second type of probe in this embodiment is an addressed probe. When a station has data to send to or by the second station, it inserts probes addressed between broadcast probes with much higher repetitions. These addressed probes force the addressed station to respond. Thus, when a station has data to send, it sends a short probe addressed at faster periods, thereby increasing the connectivity to the desired station. The addressed station knows the probing station has data to send, otherwise it would not address it. The addressed station then chooses to transition to a data channel over which the two stations transmit data.

If the station does not see its own identifier in the probe list and the list is not full, then it should randomly reply to the sending station with the power level required to return to the station in question. This prevents the remote station from ever seeing any neighbors because they will all probe at a lower power level.

The probe addressed from each station also lists the station from which it has received the data that it will send to the addressed station. There are a number for each station ID in the list

PL 193 013 B1 indicating how up-to-date the data under consideration is. So any other station hearing the probe knows it has a route back to the data source and knows how long it takes for the data to reach it. This information is then used for route selection.

If a station hears that two stations are sending addressed probes with the same origin ID but different message delays, it determines which one is smaller and therefore which route is better to choose. This is a gradient towards the origin identifier. When a station wants to reach the origin identifier, it uses this information to route the segments. If conditions change, the station dynamically selects a new route for the segments under consideration.

The station always knows how much power is needed to return to the other station. So it knows what power to use so that the probes are heard by all the neighbors.

For example, if a station wants to reach five of its neighbors, it polls with the power required to reach all five of its nearest neighbors. In the first method of probing, the station simply increases the power in steps of 10 dB until the required number of neighbors is obtained. However, since it uses 10 dB steps, it can significantly exceed the required number of neighbors. Then it would drop its power by 10 dB and be below the required number. Now the situation is that the station knows that if it drops another 10 dB, it will lose the required number of neighbors. Instead, the station calculates the power it needs to probe to reach the required number of neighbors and will not go below this power even if the required number is exceeded. The required power always changes as conditions change.

The station tries to keep the number of direct and indirect neighbors to a minimum. For example, if it tries to maintain one direct neighbor and at least five indirect and direct, it will calculate the power required to reach the direct neighbor. If it reaches four other neighbors via this neighbor, it will achieve five direct and indirect neighbors. Otherwise, it uses more power to obtain two direct and seven indirect neighbors until it has a number not less than required.

Part of the demodulation process at each station includes forward error correction. If the forward error corrector detects errors that cannot be corrected while receiving the packet, it notifies the main code that an error has occurred. The master code then stops receiving the packet. As a result, the station is not bound by receiving a packet that is corrupted. It also helps a station to receive another packet early from another station. Sometimes, in a network of the type under consideration, one station transmits at a higher level than the other station, thereby damaging the packet. The receiving station is able to detect a damaged packet, interrupt reception and start receiving a stronger signal.

In a second embodiment of the invention, a method similar to the first method is used. However, instead of using the time since the detection of the message segment for route selection, this method uses the total or aggregate transmit power required for route selection. The total transmit power required is the power required by each intermediate station to reach the next station from the start to the destination. Each intermediate station also includes a fixed transition factor, usually of 3 dB, to prevent reverse routing.

The station sends out broadcast probes at regular intervals. The broadcast probe includes a list of other stations detected by the current station, or stations that have a flag set, e.g., a busy flag in motion. A station is considered busy in traffic if it is currently either sending message data or receiving message data. For each station, the list also includes the total transmit power required to reach the station and a number of flags indicating the type or status of the station, e.g. gateway, certification authority / network operator, busy traffic, and so on. Markers are used to refine route selection.

The first part of the list contains the identifiers of the immediate neighbors or stations that the current neighbor has detected with its receiver. The transmit power required, which is included in the list, is the power that the current station would use to reach these stations directly. The second part of the list includes each station that has a tag set, that is, a gateway station that is busy, and so on. The transmit power required, contained in this part of the list, is the minimum total power required to reach these stations through the current station. It is possible that the total transmit power required to reach one of the direct neighbors through the intermediate station is less than the direct transmit power required. This usually occurs when there is a direct route and an alternate route to the same destination station, with the alternate route using less combined power.

With reference to Fig. 1, if station I detects both stations M and L, it has the required direct transmit power for both stations. However, when station I detects the probe of station M, it sees the list for

Station I then has the direct power required for station L, as well as the total power required for station L via station M. It is possible that the total required power to transmit to station L via station M is less than the direct power required. transmission to station L.

Each time a station detects another station's probe, it calculates the power it would need to return directly to that station, remembers the required direct transmit power for each detected station, and checks the list included in the other station's probe. In this list, he sees what power would be required by the remote station to reach one of the stations included in the list. The local station adds the linear power requirement for the remote station to reach one of the stations on the list to the power required by the local station to reach the remote station, and then adds an additional switch factor to that number. This new sum is the total power that the current station announces in probe signals to other stations.

The local station first converts the direct power required to reach the remote station from dBm to watts. And then it converts the power required by the remote station to reach the station on the list from dBm to watts. The local station adds these two numbers to get the new value in watts, which is then converted back to dBm. At this point, a transition factor, typically 3 dB, is added. The new sum is the total power required by the local station to reach the station advertised in the remote station list.

From the data rate of the caller's channel, the duration for all clocks is calculated using the multiplication factors. The absolute values ​​of the factors are listed below, but only as examples, and may vary considerably. In addition, the values ​​change dynamically with the change of traffic in the network and the number of stations.

<td>Clock</td><td>multiplication factors</td><td>Values</td><td>Example for 80k speeds</td>

<td>Full Tx Queue</td><td></td><td></td><td></td>

<td>the sounder clock</td><td>= PrbF x MaxTxDur</td><td>= 1 x X</td><td>= 247 ms</td>

<td>Empty Tx Queue</td><td></td><td></td><td></td>

<td>the sounder clock</td><td>= PrbE x PrbF x MaxTxDur</td><td>= 10 x 1 x X</td><td>= 2471 ms</td>

<td>Channel adaptation period</td><td>= Adpt x MaxTxDur</td><td>= 5 x X</td><td>= 1235 ms</td>

<td>Increase of the required power TX</td><td>= Txlnc x MaxTxDur</td><td>= 5 x X</td><td>= 1235 ms</td>

<td>Duration of the data channel</td><td>= Data x MaxTxDur</td><td>= 5 x X</td><td>= 1235 ms</td>

where MaxTxDur = max packet duration + RxTx redirection + delay Tx = = X = 247.1 ms

The maximum packet size is set to 1023 bytes. The length for all clocks increases if the maximum packet size is increased. The effect of increasing the maximum packet size is to reduce the number of probes in a given period of time and therefore the connection to the neighbor is slower, which in turn increases the propagation delay of the data carried over the network. On the other hand, if the maximum packet size is reduced, the amount of data that can be sent between probes on the calling channel is reduced. This in turn increases the propagation delay of the data carried over the network. By weighing the amount of data that is sent on the calling channel against the number of probes over a given period of time, a maximum packet size is determined.

If a station has more data to send than can fit in a packet of maximum size, it requests the other station to switch to the data channel. The two stations can then send more data to each other as long as they remain on the data channel. They can stay there as long as they have data to send, or for the duration for the data channel, whichever comes first. If the station moves to the data channel and does not find the other station, it will revert to the calling channel.

Stations on such a network tend to move, and so may move away or even out of range. The local station therefore needs to know how to increase the power required to reach the destination station and possibly remove from the list a destination station that is no longer nearby. Path loss between two stations changes quickly, especially due to Rayleigh fading. Each time a station transmits to another station, it has to do so at optimal power and determines the required power each time it hears another station's transmission. Typically, when two stations are busy sending data to each other, they will have multiple transmissions per second. Every time they recalculate

The power required to reach the second station, and the power changes typically follow changes in the Rayleigh decay cycle.

This has a good effect on optimizing the power required for each transmission, but causes path selection problems as it is possible that with each dropout the alternative path seems temporarily better. The station must therefore maintain the two necessary direct transmit power values. One value is used for each transmission to determine the power required for transmission, and the second value is used for routing. The latter value does not change according to the fast path loss changes associated with Rayleigh fading, but instead has a more damped effect, changing according to station movement and out-of-range travel. Typically, the latter value is envisaged to be announced in the station list in the probe signal.

This suppressed power requirement is achieved by reducing the rate of increase in power required to reach any station on the list. All stations on the network increase the power required at the same rate. The rate of increase is directly related to the data rate on the calling channel. Each time a new probe is detected, the local station determines whether the new calculated required power is less than the required power it has in its own list. If it is smaller, it lowers the required power in the list. It does not reduce power in one step, but reduces it in smaller steps each time it detects a station, thus suppressing the effects of Rayleigh fading.

If a station no longer detects probe signals from a particular station, it continues to increase the power required to reach that station. Eventually, the required power reaches a predetermined maximum value, requiring the local station to remove the second station from the list. This value is typically 125 dBm.

At this point, the station has a list of required transmit powers to reach other stations either directly or indirectly. For each station given in the remote station list, there is an entry indicating the total transmit power required by that remote station. In Fig. 1, station I has directories of direct transmit power required for stations L, M, N and B, it has intermediate directories for all stations detected by the last four stations, i.e. stations A, G, H, J, L, M, N and O. L, M, and N stations appear in both direct and indirect directories as they are shared neighbor stations. If station I wanted to route data to station M, it could choose to transmit directly to station M via stations L or N. Station I would determine for which route the required power is the lowest and would use this path to route the message segments to station M.

At this point, station I has no direct route to station O, and if it does not want to connect to station O, it does not need this route. However, if station O is busy moving from station A, the segments to and from station O will traverse station I, and in this case station I detects these traversing segments. The header of each segment indicates the total transmit power required to return to the beginning of the segment. When station M first receives a segment from station O, it places the required direct power in the segment header before handing over to station I.

When station I receives the segment from station M, it adds the power required to reach stations M to the segment from station O and also adds a transition factor to this value. The segment now includes the total power required from station I to station O via station M. The procedure is repeated at each transition until the segment reaches station A.

Based on the power required for each station, any station in the network has a power requirement gradient towards any other station in the network. The station simply routes the message segments to the lowest power required.

When originating station A wants to connect to destination 0 first, there is no gradient since neither station is busy in traffic. To create the gradient, originating station A sends a special chasing message to destination O. This message is routed from originating station A by shifting it to the required power gradient. At each station, the message is split into two and routed in two different directions. Thus, the message flows through the network in the direction from station A.

Once the chaser message has reached a station that has a gradient towards destination 0, it is routed to the destination. When it reaches destination O, station O sends back an ETE end-to-end acknowledgment message to station A. This message automatically has a gradient directed back to originating station A because the chaser message has formed a gradient. The ETE message has a higher priority than the chaser message and therefore moves faster on the network. When any station receives an ETE message, it routes it to the originating station along the path where the chasing messages were sent. This is used to stop the flow of a prosecuting message.

PL 193 013 B1

The chaser message is very small and has a short duration of activity. Thus, even if the chaser message flows through the entire network, it has a minimal effect on network capacity.

When station A first sends the chaser message, it is also marked as busy in traffic and keeps this flag for the duration of the chaser message. Any other station detecting station A's probes sees that the tag is set and lists station A for its own probes. Any other station that detects the second station with station A in the list with the busy flag set in traffic also recommends station A with the busy flag set in traffic. Every time station A sends a new message, it will reset the traffic timer to be the same as the active time message being sent. If station A stops sending messages, the traffic marker time will eventually expire and is no longer set.

A station places a different station in the list of probes each time it updates the required power as more suitable than it was before, i.e. if the station detects a neighbor station or detects the station in the list of another station, it determines whether the required transmit power to the station in question is more appropriate than it was before. value that exists in the internal list. If more appropriate, it changes the required power in the internal list and also places the station on the next list of probes. As a result, the list of probing signals is small.

If station A no longer has a busy flag set in traffic, it is no longer placed in the list of other stations. Other stations will slowly increase the transmit power required to reach station A, and since it is no longer in traffic, they will not receive the updated values. Finally, the transmit power for the other stations required to reach station A will reach a predetermined value, requiring them to remove station A from their lists.

In addition to the busy flag in traffic, there are a number of other flags, usually to designate essential stations in the network with which any other station might need to connect from time to time. The gateway tag is used to indicate a vital station that is used as a gateway to another service, for example for Internet access. If a station has a gateway to the Internet, it has a tag set to indicate that it is an Internet gateway. Each time the neighbor station detects a more suitable power required for the gate station, it lists it with the gate tag set. Any other station that detects a gate station in its neighbor list in turn places the gate station in its own list. The gateway station directory is carried over the network. Finally, all stations have gateway stations in their lists. So all stations that need access to the Internet know where they need to route data from the Internet.

There can be more than one Internet gateway on the network. The station requires only one gateway, therefore the station only places one Internet gateway on the list. It always selects the one with the lowest transmit power required. The effect is that stations will always have a better power gradient to the nearest gate and will usually only list the closest gate. If the nearest gate becomes unavailable, they will automatically reject that gate as the required power increases to the point where other gates seem more appropriate.

In the same way as a gate tag is used, other tags such as a certification body can be used. Certification authorities or network operators are stations in a network that maintain and issue authority certificates used for network security. All stations on the network use a set of private and public keys to verify and encrypt data received from and sent to other stations. As with the Internet gateway, a station only needs to know the route to one certification authority.

When the destination station receives a message from the originating station, it starts a gradient timer that elapses halfway through the initial message activity time. Active time is the amount of time a message is valid. This new gradient timer then takes effect before the received message expires. If a new message is received from the same origin station, the gradient timer is set to the new value. This means that as long as messages are received from a specific origin station, the gradient clock never expires.

When messages from the specified origin station stop arriving, the gradient timer expires. Then the gradient message will be sent to the origin station. Since the message contains the total power required for reverse transmission to the station that is sending the gradient message, the transmit power required gradient will be updated from the destination station that originally received the messages back to the originating station.

The purpose of the gradient message is to ensure that a new gradient is created when no messages are received from the originating station. Messages may stop coming because they don't

PL 193 013 B1 has more messages either because the gradient is no longer valid. This is caused either by the simultaneous movement of a number of mobile stations or by switching them on or off simultaneously . Typically the network does not need a gradient message because the gradient improves itself dynamically between the two stations that are in motion, which is a fault tolerant mechanism.

You can also append a tag to the last message sent from the station to prevent sending a gradient message, i.e., the originating station informs the destination station that there will be no more messages. The destination station then knows that no gradient message is needed.

Stations do not send routing information with the messages they send and do not send routing information among themselves, but send information about the required transmit power and information about which stations are in motion or are gateways, etc. The task of each station in the network is dynamically making a route selection decision for each message or message fragment in the transmission queue.

The station at any time dynamically changes the route of the message fragment from one transmission queue to another and maintains a separate transmission queue for each of its neighbors. The route selection at each station is not correlated with the route selection at the other stations and is based solely on information about the required transmit power and route selection status flags. Since the required transmit power and the status flags may change at any time, the selected route for the message also changes. The station routes the message as it is based on the current information available. Thus, if a new route selection opportunity arises that appears to be better than the current route for a particular message, the station dynamically re-routes the message.

The variation in transmit power required changes the routing gradient for each station in the network. At any point in time, a transmit power required gradient for a particular station indicates the best route for the message. The station always routes the message down the gradient to the destination station, except for the chaser messages which are directed up the gradient of the starting station and simultaneously down the gradient of the destination station.

As described above, each time a station detects one of its neighbors it calculates the transmit power required to reach that station. The local station reduces the transmit power required value it previously stored for its neighbor if the new transmit power required is less than the previous value. However, it does not reduce it to the new value in one big step, but in small steps each time it hears a neighbor station.

Thus, the more often the local station hears its neighbor, the lower the transmit power required value will be. If it does not hear the second station for some time, the transmit power required value starts increasing at regular intervals until it finally reaches a level where the local station removes its neighbor from its internal list.

If a neighbor station is transmitting a lot of data on behalf of other stations, it spends a lot of time on the data channel and little time on the calling channels. As a result, neighbors will not hear it very often, so the transmit power required to reach that station is high. This causes the neighbors to choose alternative routes for the data as appropriate. As data is now forwarded via alternate routes, the station that was sending a lot of data has less data to send. Thus, the method dynamically spreads or distributes the traffic load among the neighbors based on how often the stations are heard. The gradient of the required transmission power always moves away from areas with high traffic congestion to areas with low traffic congestion. The transmit power required gradient automatically balances route selection and traffic congestion.

If a neighbor station hears another station frequently and over a long period of time, it always has a suitable transmit power required gradient towards that station. However, if a station hears the other station frequently, but only for short periods of time, it has an average transmit power required gradient. The longer the time periods, the better the gradient is. Thus, the station tends to have a better gradient towards the station it hears frequently and long.

Although the station hears the second station frequently and for long periods of time, the second station is far away and in this case the received transmissions are not strong. The required transmit power is calculated from the path loss to the second station. The weaker the received signal, the greater the path loss and thus the greater the transmit power required. The greater the transmit power required, the greater the transmit power required gradient. As a station routes towards areas with less transmit power, it tends to route to other stations that are closer.

PL 193 013 B1

The better a station hears another station, the better the transmit power required gradient is. Since the gradient is based on the required transmit power taking into account the background noise of the distant station, it is also better towards stations that have little background noise. Stations that have a lot of local interference will have a lot of background noise. The station has a better transmit power required gradient towards stations that hear well and that have little background noise. Thus, areas of high background noise are always avoided in routing the data.

The route selection is based on fast and slow Rayleigh decay. In the case of a slow Rayleigh fade, a station hears the other station frequently, long and well when there is little path loss between the two stations. The result is a good route through the station during a period of low path loss. If the fade starts to get bigger, the path loss increases and the path gets worse because the transmit power required gradient is steeper. The required transmit power is greater than the previously stored value, and the value will not be updated. However, as the transmit power required is increased at regular intervals, it automatically deteriorates slowly. Thus, route selection is dynamically dependent on slow Rayleigh fades.

In the event of a rapid Rayleigh fade, the station detects the second station frequently, but in short bursts. This results in an average transmit power required gradient. The gradient is suppressed by the slow gains and the fact that the transmit power required is only reduced in small steps. This means that the station is not dependent on the fast Rayleigh fade, but has an average effect. Thus, if there are three stations in remote locations, one mobile, one in the Rayleigh valley, and one at the Rayleigh peak, the station in the valley has a weak transmit power required gradient, the station at the peak has a good transmit power required gradient, and the mobile station has an intersecting gradient. . This is equivalent to the situation-dependent availability of these three stations, i.e. the valley station offers the wrong route and requires a lot of power to reach. The station at the top offers a good route and requires little power to reach. A mobile station requires low power some of the time and more power at other times, and as such provides an average route.

The transmit power required gradient used for routing takes into account the average station power that the station uses during fast Rayleigh fading. Although the routing averages the effect of a fast Rayleigh fade, the station continues to dynamically change transmit power while it is currently transmitting a data packet. Thus, the actual transmit power used when the station transmits the data packet matches the fast Rayleigh curve. The station maintains a separate required transmit power used for actual radio transmissions that is not suppressed in the same way as the required transmit power used for route selection.

The transmit power required gradient is continuously optimized and improved as long as data messages are flowing. The data messages from the originating station maintain a new gradient back to the originating station from any station along the route in which the data messages flow. End-to-end acknowledgment ETE messages, returning from the destination station, maintain the new gradient back to the destination station. Each station on the route announces that the destination and origin stations are busy in traffic. In turn, each neighbor of each station on the route also announces that those stations are in motion, and so on. Thus, there is still an optimizing gradient directly along the route and also along the sides of the route. Each time a new message flows along the gradient, the gradient is re-optimized. Should any station directly along the route become inactive, move away or become congested, the gradient automatically optimizes around that station.

In the case of a very common destination for data messages, such as a GPS based vehicle tracking center, a lot of messages are sent to the destination station and in turn a lot of ETE messages are sent from the destination station. These ETE messages propagate in all directions to all vehicles that send the GPS position update to the tracking center. This means that a specific central station has a very optimized and widespread gradient in the network. If a new vehicle turns on the tracking device for the first time or after a long time off, it immediately routes to the central station without the need to send a pursuer message. It is known that the central station always has optimized routes, so there is also no need for it to send gradient messages.

To simplify the updating of software in the network, a mechanism is used that uses information from direct neighbors. When a new software update becomes available, it is loaded onto the hard drive of any station on the network and then sent in blocks from that station to its direct neighbors. In turn, the direct neighbors send them to their neighbors until the entire network is updated.

PL 193 013 B1

Figure 3 shows the flowchart of a software update protocol. When the first station has an update on its hard drive, the user issues a command telling the station to start broadcasting the new update on its probes. When the direct neighbor detects the probe, it sees the new update. The neighbor station then requests a software update message from the first station. The station sends the first update block. When the direct neighbor receives an update block, it waits for a certain period of time before requesting the next block. The reason for the wait is to prevent software update messages from blocking the entire network.

The neighbor station continues to request update blocks from the first station until it has got all the blocks. Each time it receives a new block, it starts broadcasting on its probes which version and block number it has. Then the neighbor station starts sending updates to the other neighbor stations even before it has the entire update. This starts a bucket run for software updates. It is possible that one of the neighbors gets the update earlier than itself. If this other neighbor is closer in transmit power than the first station, the neighbor even starts an update request from the closer neighbor.

The station always requests an update from the nearest neighbor. This means that if a station is mobile, it will request updates from other stations each time. The size of the software update blocks is so small that when the station is mobile, it does not carry the delayed long message on the network.

If a station moves to an area where its neighbors do not have a complete update or no update, it waits for the software update to come along with it on the network. If a station returns to an area where the update is more advanced than it has, it continues where it left off previously.

The first part of the software update tells you when to update. This is the date and time that the station needs to download an update and replace the existing software. Until then, the station will simply hold the update. After updating the current version of the software, it keeps the update until a newer version arrives. This allows the update to be sent to each station that may have been disabled during the software update period.

It is also possible to include data on which identifiers should manage the update, which allows a partial network update to test the new software version. The update is still sent to every station on the network, but only some stations will perform the update. If the update works as planned, a special update block is included at the end of the software update. The station detects that there is a new block to join. When it receives a new block, it checks for new update IDs in the block. This allows another station to be updated without sending all the new software.

The following is a table defining the structure of the probe signal and data packet used in the method and system according to the invention, together with a glossary with explanations.

Probe and data packet format

<td>Variable</td><td>Long guest bit</td><td>It enables</td>

<td> 1</td><td> 2</td><td> 3</td>

<td>Admission</td><td> 64</td><td>Modem test sequence (101010101010 etc.)</td>

<td>Syncl</td><td> 8</td><td>The first sync character used to lock the Zilog</td>

<td>Sync2</td><td> 8</td><td>Second sync character used to lock Zilog</td>

<td>Sync3</td><td> 8</td><td>Third sync character checked by the software</td>

<td>Package size</td><td> 16</td><td>Packet size from Sync3 to last CRC</td>

<td>Size control</td><td> 8</td><td>Packet control = MSB XOR LSB of the packet size</td>

<td>Protocol version</td><td> 8</td><td>Protocol version</td>

<td>Package type</td><td> 8</td><td>Packet type (e.g. probe, data, key, etc.)</td>

The table continues

<td> 1</td><td> 2</td><td> 3</td>

<td>Sender ID</td><td> 32</td><td>The ID of the transmitting station</td>

<td>Recipient ID</td><td> 32</td><td>Receiving station ID (0 = Broadcast)</td>

<td>Package number</td><td> 16</td><td>Package number</td>

<td>Tx adp power</td><td> 8</td><td>The current power of the transmitting station in dBm</td>

<td>Tx adp path loss</td><td> 8</td><td>Path loss measured at the transmitting station in dB</td>

<td>Activity of adp Tx</td><td> 4</td><td>The current level of activity of the broadcasting station</td>

<td>Tx adp antenna</td><td> 8</td><td>The current antenna configuration of the broadcasting station</td>

<td>Adp Tx Bkg RSSI -1</td><td> 8</td><td>RSSI of the transmitting station in dBm -&gt; current modem-1</td>

<td>Adp Tx Bkg RSSI</td><td> 8</td><td>RSSI of the transmitting station in dBm -&gt; current modem</td>

<td>Adp Tx Bkg RSSI +1</td><td> 8</td><td>RSSI of the transmitting station in dBm -&gt; current modem + 1</td>

<td>Jump noise adp Tx</td><td> 8</td><td>The excursion frequency and the level at the transmitting station</td>

<td>Activity of adp Rx</td><td> 4</td><td>The required level of activity for the receiving station</td>

<td>Adp Rx channel</td><td> 8</td><td>Rx and Tx channel required for receiving station</td>

<td>CRC header</td><td> 16</td><td>16 bit CRC for header data</td>

<td>Neighbor route check marks</td><td> 8</td><td>Bit 0 - in motion, bit 1 - gate, bit 2 - cert.</td>

<td>Neighbor data size</td><td> 16</td><td>The size of the route selection data in bytes = 3 + 4 (update) + IDs * 6</td>

<td>Neighbor software update</td><td> 32</td><td>Software update version (16) and block number (16)</td>

<td>Neighbor data</td><td>x</td><td>Neighbor station * (32 (ID) +8 (TxPowerReq) +4 (ModemReq) +4 (markers)</td>

<td>Package data</td><td>x</td><td></td>

<td>CRC</td><td> 32</td><td>32 bit CRC for the whole packet, including header</td>

The introduction is a modem trial sequence of alternating ones and zeros.

Sync1-Sync3 are three sync characters used to detect the beginning of a valid packet.

Packet size is the total size of the packet from Sync3 up to and including the last byte of CRC. The maximum packet size allowed on the probing channel is determined by the probing rate, i.e. the station does not send a packet that is longer timed than the period between probes on the probing channel. The maximum packet size that is allowed on the data channel is determined by the length of time the station remains on the data channel.

Size control is used to check the packet size variable to avoid receiving a packet of incorrect length.

The protocol version is used to check which protocol version is in use. If the software cannot handle this version, the package is ignored.

The packet type specifies the type of packet that is sent. The next packet immediately follows the current packet if the most significant bit is set.

The recipient ID is the identifier of the station to which the packet is addressed.

The sender ID is the identifier of the station currently sending the packet.

Packet number: Each sent packet is given a new sequence number that is not used in any way by the protocol and is only information for the systems engineer. Each time the station is reset, the packet number starts with a random number to avoid confusion with older packets.

Adp Tx power is the actual power of the transmitting station, given as absolute power in dBm, in the range -80 dBm to +70 dBm. The field allows values ​​from -128 dBm to +127 dBm.

The adp Tx path loss is the path loss measured by the transmitting station. Path Loss = (Remote Tx Power - Local RSSL) of the receiving station's previous transmission.

PL 193 013 B1

A value of 0 is used to indicate that the transmitting station's RSSI has been blocked. The path loss is used as a correction factor at the receiving station the next time the receiving station transmits to the transmitting station.

Tx adp activity is the level of activity of the transmitting station, measured as activity = watts * time / (bandwidth * success) averaged over time.

The adp Tx antenna indicates the current antenna configuration used by the transmitting station. Each of the 255 possible configurations describes the entire antenna system, i.e. the transmitting and receiving antenna.

Adp Tx Bkg RSSI is the actual background RSSI at the transmitting station for the modem that is currently transmitting and allows values ​​from -255 to -1 dBm. The value sent is the absolute value of the RSSI and the receiving station must multiply the value by -1 to get the correct value in dBm. A value of 0 is used to indicate that the channel is not available or is greater than or equal to 0 dBm. The value of 0 dBm cannot be used for adaptation purposes.

Adp Tx Bkg RSSI-1 is the same as above but for the previous modem.

Adp Tx Bkg RSSI + 1 is the same as above but for the next modem.

The jump noise adp Tx is the lower 3 bits for the jump frequency in Hz, 0 = none, 1, 5, 10, 50, 100, 500 and&gt; 500, and the next 5 bits for the amplitude in dB.

The activity of adp Rx is that if a station has a high level of activity and interferes with other stations, they will use this field to force the active station to drop its level of activity. If a number of stations request a drop in activity, the interfering station responds and lowers the level of activity. If no station requests a drop, the active station begins to slowly increase its activity level. Thus, if the station is in a very remote area, it continues to increase its level of activity to establish communication. If it is in a very busy area, other stations will keep their activity low.

In the preferred embodiments of the invention, the station always tries to maintain five neighbors, so other stations should not need to request that the station decrease activity. However, the feature is applied for cases where stations cannot lower their power or increase their baud rate any further, but still interfere with too many other stations.

The adp Rx channel allows 255 predefined channels. These channels are fixed for the entire network. Each channel has an associated polling rate which can be disabled and makes it a data channel. Each channel has a minimum bit rate associated with it. The channels will have specific Tx and Tx frequencies and may be specified as other means as well, e.g. satellite, Diginet, ISDN etc.

The transmitting station requests that the second station move to a data channel that has polling turned off when it has more data to send to the receiving station than fits in a packet size acceptable for the probing channel.

The header CRC is a 16-bit CRC check for the header data, the sum of all bytes in the header, which is only checked when the packet CRC is in error. This is a way to determine which station sent the packet. If the packet CRC is in error and the header CRC is correct, the data provided in the header should be used with caution as the header CRC is not a very good means of error detection.

The neighbor routing checkboxes given below are not included in the header CRC as they cannot be used if the packet CRC is not correct. This makes route selection less error-prone.

Neighbor routing markers are used to refine the routing. They provide additional information about the current station. The currently specified bits are 0-set if the current station is busy in traffic, 1-set if the current station is a gateway to the Internet.

Bit 2 - set if the current station is a certification authority and bit 3 - reserved.

Another 8-bit byte can be added if more flags are required.

Neighbor data size determines the size of the route selection data in bytes, which includes the neighbor route flags and the size of the neighbor station data, which is 3 bytes. Another 4 bytes are appended if the Neighbor software update field is included. An additional 6 bytes are appended for each neighbors included in the data section of the neighbor station. Neighbor software update must be announced if the neighbor station data is given.

Neighbor software update is the current version of the update software available in the current station, i.e. the upper 16 bits of the field and the current block number available, i.e. the lower 16 bits of the field.

PL 193 013 B1

Neighbor Data is a list of neighbors for which the current station has route selection data. Whenever the current station receives updated route selection data for stations that is better than the data it had previously, it updates its own data and places the station in the list on the next probe. The data section has four subfields for each station in the list.

Station ID is a 32-bit field with the neighbor identifier.

TxPowerReq is an 8 bit field indicating the total or direct transmit power required to reach the station ID of the current station.

The Req modem is the modem required by the current station to reach the destination station.

Markers are markers giving additional information about the route selection for the destination station: bit 0 - in motion, bit 1 - gateway, bit 3 - certification authority, bit 4 - direct neighbor. The last bit indicates that the station in the list is a direct neighbor of the current station.

The packet data is composed of one or more segments. Segments are of any type, derived from, or intended for any identifier.

CRC is a 32-bit CRC check for the whole packet. If this CRC is in error, the entire data packet is discarded, however, header data can be recovered if the CRC of the header is correct.

The format of the message segments

<td>Variable</td><td>Position bit</td><td>Long guest bit</td><td>Description</td>

<td>Segment type</td><td> 0</td><td> 4</td><td>Message fragments of segment type = 0x00</td>

<td>The segment type that is being confirmed</td><td> 4</td><td> 4</td><td>Confirmed segment type (used for confirmation type segment)</td>

<td>Target ID</td><td> 8</td><td> 32</td><td>Target ID of the message fragment</td>

<td>Start ID</td><td> 40</td><td> 32</td><td>The start ID of the message fragment</td>

<td>Message number</td><td> 72</td><td> 14</td><td>Send message number 1 -&gt; 16383</td>

<td>Message submission number</td><td> 86</td><td> 2</td><td>Message Submission Number 0 -&gt; 3</td>

<td>Message size</td><td> 88</td><td> 16</td><td>Total number of bytes in the message 0 -&gt; 65,535 (+1)</td>

<td>The beginning of the fragment</td><td> 104</td><td> 12</td><td>Initial message fragment sent 0 -&gt; 4 095</td>

<td>End of the passage</td><td> 116</td><td> 12</td><td>Final message fragment sent 0 -&gt; 4 095</td>

<td>Fragment priority</td><td> 128</td><td> 8</td><td>256 levels (0 -&gt; 255, 0 = highest priority)</td>

<td>Fragment expiry time</td><td> 136</td><td> 24</td><td>In milliseconds 0 -&gt; 16777215 (4.66 hours)</td>

<td>Fragment creation time</td><td> 160</td><td> 24</td><td>In milliseconds 0 -&gt; 16777215 (4.66 hours)</td>

<td>TxPowerReq for the starting ID</td><td> 184</td><td> 8</td><td>Tx power required to reach initial ID</td>

<td>TxModemReg for the starting ID</td><td> 192</td><td> 8</td><td>Tx modem required to reach initial ID</td>

<td>Segment data</td><td> 200</td><td>x</td><td>The rest of the segment contains the message fragments</td>

Segment Type indicates the type of segment that is being sent. The types include:

message segment that contains the message data, sent from the originating identifier to the destination identifier and forwarded from the intermediate station to another intermediate station, segment acknowledgment used for segment message acknowledgment, sent from the intermediate station that has just received the segment message from another intermediate station, ETE message the segment sent from the target ID to the starting ID when the target ID receives the segment message, also transferred from an intermediate station to another intermediate station,

Segment ETE confirmation used to acknowledge the segment ETE, sent from the intermediate station that has just received the segment ETE from another intermediate station.

The segment type acknowledged is used to indicate the type of segment being acknowledged by the current segment.

The target ID is the target ID for the current segment.

The start ID is the start ID for the current segment.

The message number is the number of the sent / acknowledged message.

The message submission number is the message re-submitted after a specified period by the starting identifier if no ETE is received from the target identifier. This field shows the current show number.

The message size indicates the total number of bytes in the message.

Chunk start indicates which chunk start number is being sent. The message is broken into pieces of 16 bytes each.

Fragment End Indicates which fragment ending number is sent.

Fragment priority indicates the priority of the fragment.

Fragment's TTL indicates the relative TTL, which is the number of milliseconds left until the fragment is no longer valid. Each intermediate station is responsible for reducing this value. Typically the station converts this value to absolute time and converts it back to relative time just before transmission. Thanks to this, there is no need for all clocks in all stations to be synchronized.

Fragment creation time is the initial number of milliseconds the fragment should be active. This number does not vary and is used by the destination and intermediate stations to determine how long it took for a fragment to reach them from the originating station.

TxPowerReq for the Origin ID is the total transmit power required to reach the Origin ID from the station currently sending the segment.

TxModemReq for the starting ID is the lowest modem number used at the intermediate station to reach the starting ID from the station currently sending the segment.

The segment data contains the actual message fragments.

Figures 4, 5, 6 and 7 show the basic equipment used to implement the method and system according to the invention. The description below relates specifically to the operation of the equipment.

Based on the decision to transmit, the main processor 149 selects a power level, baud rate, and packet duration to use and sends this packet to the serial controller 131 and simultaneously via the peripheral interface switch 147 to the transmit / receive switch 103 in transmit mode, and turns the transmitter on with delay. Zilog 131 sends packet data along with header and CRC through PN sequence coders in block 128 or 130 depending on the selected rate.

The host processor 149 includes in the data packet, as one of the information fields, data corresponding to the transmit power used, which is the same transmit power that is sent to the PIC power control block 132, which is in turn used to control the power control circuit 141, which is in turn, it controls the gain control block and the lowpass filter 143, which uses the feedback from the power amplifier 145 to control the drivers 144 and 142. The gain detection and feedback method allows the exact power level to be determined based on an instruction from the power control circuit 141.

Before turning on the power amplifier, the synthesizer 138 selects a transmission frequency, and then the power amplifier 145 receives the instruction via power control unit 141 and the amplifier turns on.

If power levels below the minimum provided by the power amplifier 145 are required, the switchable damping circuit 102 may be turned on to provide additional attenuation of up to 40 dB. Thus, the processor instructs the power amplifier to turn on the attenuator 102, resulting in an output power level range of minus 40 dBm to plus 50 dBm. When the amplifier is turned on, the processor obtains forward and reverse power information from the low power detection circuit 101 that is sent by the analog-to-digital converter 146 and used by the host processor 149 to control the transmitted power level. This information is then stored in dynamic RAM 150 to provide information on the forward power level and reflected power generated as compared to the requested level.

PL 193 013 B1

The amount of transmit power output is affected by the performance of the transmit power control loop through blocks 145, 144, 142, and 143 and the switched attenuator 102. Additionally, any mismatch in antenna 100 also causes changes in reflected and forward power. The relative power obtained at the output for the different required levels is stored by the processor in RAM, providing a table comparing the desired and actual power output levels. It is used to allow the processor to use a more precise power level field in information provided in later transmissions, within messages or probe signals. Since the power level varies from minus 40 dBm to plus 50 dBm, there are effectively ten different power levels in the 10 dB periods that can be transmitted. Thus, the table stored by the processor has ten power levels, and the requested power level and the actual power level will be within this range.

Every other station in the network then receives the transmitted signals through the antenna 100. The received signal passes through a low power detection circuit 101 and a switched attenuator 102, which is initially set to 0 dB attenuation, followed by a 2 MHz bandpass filter 104 which removes the interference out of band and goes to preamplifier 105 which amplifies the signal before being mixed downstream by mixing circuit 106 to an IF signal level of 10.7 MHz. This signal is filtered by a bandpass filter, amplified in the IF 108, and filtered and amplified in blocks 109, 110, 111, and 112.

Final filtering occurs at blocks 114 and 115, and at this stage, the signal is measured at block 116 using a narrowband RSSI function that is used by host processor 149 to determine incoming signal strength.

If desired, this enables the processor to request from the power control circuit PIC 132 to turn on additional attenuation up to 40 dB at the receiver. It is only necessary to enable additional attenuation if the signal exceeds the measurement range of block 116. Otherwise, leave 0 dB attenuation for attenuator 102, allowing full receiver sensitivity to receive weak signals. The incoming transmission is measured simultaneously in two bands, namely 8 kHz and 80 kHz. The 80kHz band is measured by disconnecting the 10.7MHz LF signal downstream of the 150kHz ceramic filter 109 and using the 150kHz ceramic filter 121 and the NE604 integrated circuit 120. It also has an RSSI output that is received via the interface by the main processor 149.

The wideband and narrowband RSSI signals are measured by an analog-to-digit converter 146, which then passes the data to the host processor 149, which has a lookup table, retrieves information from the analog-to-digital converter, and derives the receive signal power from the previously scaled data in dBm, typically from minus 140 dBm. up to 0 dBm. This information is typically generated using the output of a scaled signal generator by inputting it to the input of the receiver and then selecting different signal strength levels and instructing the processor via keypad 209 what power levels are input. The information is then permanently stored in static RAM or high-speed RAM 150.

The receiving station accurately records the power level of each incoming transmission, then reads the address of the incoming transmission and the embedded transmission power level. When comparing them, for example, the transmit power level plus 40 dBm at the receiver is measured as minus 90 dBm and is used to calculate the 130 dB path loss. The path loss varies from 0 dB to a maximum of 190 dB (+50 - (- 40) = 190). The minimum path loss that can be measured is dependent on the transmit power of the transmitting station and the maximum signal that the receiving station measures. Since in this arrangement the maximum received signal is 0 dB at the antenna terminal 100, a path loss of 0 dB can be measured as long as the transmit power is less than 0 dBm. Otherwise, for example with a transmission power of 50 dBm, the minimum track loss that can be measured is 50 dB. This is improved by adding further steps in the switchable damping circuit 102 or by using a different receiver circuit. If the switched attenuator 102 is fully switched and the output of the analog-to-digital converter indicates that the RSSI signal is at its highest level, the receiving processor marks the transmission related data as blocked. This means that the path loss is less than that which can be measured.

The processor continuously measures the background signal and interference while receiving, and provided no transmissions are detected on any modem at any baud rate, controls and measures noise and interference in dBm and generates an average that is stored in static RAM. Once a transmission is detected, the most recent noise measurement is compared to the signal strength to determine the signal-to-noise ratio. In each transmission, background noise received before

The transmission is broadcast within the transmission message or probe as the next field along with the transmit power. Other stations in the network take and determine from transmission not only the path loss, but also the noise base of the distant station just prior to its transmission. The receiving station, because it knows the path loss and the noise base of the remote station, knows the power to transmit to obtain the desired signal-to-noise ratio at the remote station.

The required signal-to-noise ratio is usually based on the operation of the modem and the number based on the duration of the packet and probability of success. This required signal-to-noise ratio is stored in the database by the processor and is continuously updated based on successful transmissions to the various target stations. For example, if the station receives the transmission and calculates that the path loss is 100 dB and the remote station has a declared noise base minus 120 dBm to provide the required signal-to-noise ratio, e.g. 20 dB for 8 kilobits per second, it transmits at a power level minus 20 dBm. This required signal-to-noise ratio is different at 80 kilobits per second in that the noise base is greater in the wider 150 kHz band compared to 15 kHz and the operation of the 80 kilobits per second modem is different from the 8 kilobits per second modem.

Thus, the receiving station knows that if, for example, the declared wideband noise base is minus 110 dBm and the path loss is still 100 dB, but the required signal-to-noise ratio is, for example, 15 dB, it requires a transmit power of plus 5 dBm. The station receiving the transmission knows what power level to use to respond to the originating station.

With the procedure outlined above, the station determines the transmit power required to reach its neighbors and then places the required transmit power in a list of neighbors that it includes in its probe signals.