Method and device for data transfer

The invention relates to a method and a device for in particular wireless data transmission with possibly even battery-operated transmitters and / or receivers, preferably a data transmission for consumption recording and / or evaluation. In accordance with the proposed method, at least one transmitter sends out data in the form of data packets by means of a time control at specified transmission times. At least one receiver is switched on by means of a time control during a reception time window for receiving the data or data packets transmitted by the transmitter, the time of transmission generally being within the reception time window and course differences in the time controls of the transmitter and the receiver being taken into account.

The invention thus relates in particular to a data transmission system with which one or more transmitters regularly transmit data packets to one or more assigned receivers. The transmitters each have a time control, which determines the times of the data transmissions in a known manner deterministically or pseudostochastically. The receivers also each have a time control, which enables them to predetermine the nominal target transmission times of the assigned transmitters using the same method as the transmitters.

Such a data transmission system optimizes the power consumption of the receivers by specifically switching them on only at the expected, nominal transmission times (target transmission times) of the assigned transmitters, the reception time window being extended by a time reserve forwards and / or backwards to avoid possible gear deviations between to take into account the timing of the respective transmitter and receiver since the last data transmission. So that the possible rate deviations do not increase in the course of time to such an extent that the transmission times and the reception time slots diverge, synchronization often takes place in such data transmission systems between the receivers and the transmitters assigned to them, which also includes the operation of the receivers over long periods enables reception time windows matched to the transmission times. This means that data transmission systems can also be implemented with battery-powered receivers.

In practice, such data transmission systems are used, for example, in systems for recording consumption of heat, cold, water, gas or electricity in residential or commercial buildings. For this purpose, battery-operated consumption recording devices are typically used which regularly transmit consumption values ​​and possibly other process data by radio to data collectors serving as receivers in the radio range. For reasons of economy, battery life spans of more than 10 years can be achieved with the consumption recording devices. The data collectors themselves are typically also battery-powered in order to save the costs of cabling for mains supply during installation. Comparable battery life is also sought here. In addition, such data transmission systems are used in particular in the context of wireless sensor networks and receivers of time signal transmitters. In the introduction to the<patcit id="pcit0001" dnum="DE19905316A1"><text>DE 199 05 316 A1</text></patcit> Such a system is described as state of the art, in which the transmission time of the transmitter is varied stochastically to avoid multiple collisions with another transmitting consumption measuring device, the last, high-resolution measured temperature of the heat transfer medium being used as a quasi-statistical influencing variable. However, there is the problem that the transmission time cannot be predicted due to the statistical variation and the receivers must remain permanently ready to receive.

For energy-saving data transmission from transmitters to receivers, especially in wireless sensor networks with battery-operated transmitters and / or receivers, the <patcit id="pcit0002" dnum="DE19905316A1"><text>DE 199 05 316 A1</text></patcit> for time synchronization, therefore, a system in which data packets are transmitted from transmitters to receivers, the transmit times being selected by the transmitters in such a way that the receivers can calculate them in advance, on the part of the receiver the algorithm, the random code number and the start time of the pseudo-stochastic random generator of all transmitters assigned to the receiver are known. The receivers are then switched on specifically for the short time interval of the data transmission and go to receive. When calculating the expected next transmission time, the measured time interval between the last and the penultimate transmission of the respective transmitter is taken into account by a correction factor. The reception time windows are extended by pre-run and post-run tolerance periods in order to intercept gait deviations that have arisen in the meantime from transmitter and receiver clocks. In a special variant, it is proposed to lengthen the length of the reception time window disproportionately depending on the time interval from the last and penultimate data reception. This enables the receiver to initially keep the time window for a first reception attempt relatively short and only to make the time window of the next reception attempt so large in the relatively rare case of an actual non-reception that the accumulated time difference between transmitter and receiver is covered with great certainty and the transmitter is "caught" again.

The <patcit id="pcit0003" dnum="DE102005056932A1"><text>DE 10 2005 056 932 A1</text></patcit> describes the effect of temperature drift on the timers of the transmitters and receivers in a radio transmission system. Since the RF quartz frequency of the radio transmitter does not change abruptly, but exhibits an approximately time-linear behavior when the temperature change is usually only slow, the relationship between the time interval since the last reception and the change in the transmitter frequency is at least approximately linear. The relationship between the deviation of the transmission time from its nominal value and the actual time interval of the current transmission event from the last transmission event is, however, at least approximately quadratic as an integral over the clock frequency of the radio transmitter. The accumulated time error since the last radio telegram reception will therefore only correspond (as the mean value over the frequency change assumed to be linear in time) to just half the time error that would result from the product of the frequency change and the time interval since the last reception. This can be taken into account by the receiver's reception control. The approach of<patcit id="pcit0004" dnum="DE102005056932A1"><text>DE 10 2005 056 932 A1</text></patcit> is therefore to neglect the temperature dependence of the time error between the relatively short transmission intervals and to assume a time-linear behavior of the frequency change of the quartz crystal. In the case of battery-operated devices, it is also proposed to correct the transmission time of the radio transmitter.

From the <patcit id="pcit0005" dnum="US5771180A"><text>US 5,771,180</text></patcit> is known for realizing a real time clock based on an oscillator, counting the oscillator signals and correcting them with correction values ​​which are permanently stored in the memory of the real time clock. For this purpose, the temperature of the oscillator is measured and the difference to a reference temperature is formed. The digital counter that counts the oscillations of the oscillator is then corrected on the basis of this difference. When determining the correction values, the entire temperature range of interest is measured and the correction value associated with each temperature is stored in the memory.

In the <patcit id="pcit0006" dnum="EP1223673A2"><text>EP 1 223 673 A2</text></patcit> a similar method is known in which correction values ​​are determined on the basis of a known temperature-dependent error curve of the quartz crystal, wherein symmetry properties of the error curve are used.

The correction of the quartz crystals themselves is comparatively complex since the corrections have to be made continuously. This is particularly disadvantageous in the case of battery-operated devices, since, due to the corrections which are to be applied continuously, a considerable battery power is consumed overall in each transmitter, even if the individual computing steps in the processors provided for this purpose require only comparatively little energy.

The Indian <patcit id="pcit0007" dnum="DE19905316A1"><text>DE 199 05 316 A1</text></patcit> The prior art described ensures by measuring the time interval between two previous transmissions and determining a correction factor therefrom that the manufacturing tolerance and aging of the time-determining components, ie in particular the quartz crystals, are compensated for in the transmitters and receivers. Short-term time deviations, which are essentially caused by fluctuations in the temperature of the time-determining components in the transmitters and receivers, are only taken into account in general by an increase in the tolerance period that is disproportionate to the time interval from the last successful reception. Since the clock gear deviation caused by temperature changes in the transmitters and receivers must be measured quantitatively and based on the sign of the deviation using a relative worst case, i.e. the maximum possible gear deviation and gear deviation direction in the operating temperature range of the transmitters and receivers, there remains potential for savings in this regard Receiving time window of the receiver and thus the required receiver operating energy not used. A receiver pulse operation proposed in the aforementioned publication also reduces the average power consumption of the receiver during the reception time window, but requires an additional preheader in the data telegram, which must be at least as long as a reception pulse period.

It is therefore an object of the present invention to further shorten the power consumption times of, in particular, battery-operated receivers in a data transmission system, in order to enable longer operation for a given amount of charge, such as is made available by a battery. In the case of mains-powered receivers, too, the amount of charge needed to save electricity should be reduced. The power consumption of the transmitters should not be increased, as would be necessary, for example, in the case of receiver pulse operation due to the preheader preceding the data packets.

This object is achieved with the features of claims 1 and 11.

In the method described at the outset, provision is made in particular for the temperature of the transmitter and thus of the timer of the time control to be taken into account when determining the reception time window in the receiver. In closed data transmission systems, for example the sensor networks for consumption detection already mentioned, the timer and / or the temperature characteristic of the timer of the time control in the transmitter (s) is usually known. By taking the temperature of the transmitter into account, knowing the temperature characteristic of the transmitter, the reception time window can be shifted in such a way that it lies optimally around the actual transmission time. A sensible and optimal choice of the time position and duration of the reception time window is thus achieved by taking into account the temperature of the transmitter or the time-determining elements of the time control of the transmitter in the receiver.

It is crucial that by taking into account the temperature of the transmitter at the time of transmission, the actual time of transmission relative to the timing of the receiver in the receiver can be predicted more precisely. This enables the reception time window to be better coordinated with the actual transmission time. Due to this higher accuracy, the reception time window can be shortened compared to the prior art, which leads to the desired power saving, which has an effect particularly in battery-operated receivers. But otherwise the solution according to the invention also helps to save electricity and reduces the standby times of mains-operated devices, which, taken together, account for a considerable amount of electricity consumption.

According to the invention, more information is taken into account in the proposed method for data transmission than in the prior art, in that in the receiver an additional information or assumption about the temperature of the time control or its timer is used in the transmitter to determine the reception time window. By taking the temperature in the transmitter into account, it is possible to predict the transmission time of each assigned transmitter more precisely and thus to optimally match the timing and duration of the required reception time window. This minimizes the energy consumption in the receiver.

According to the invention, it is additionally possible to also take the temperature of the transmitter into account when determining the time of transmission. This corrects the time of transmission so that it corresponds to the time of transmission selected in the transmission scheme. In this case too, it is possible to shorten the reception time window compared to a case in which the temperature of the transmitter is not taken into account. In this variant, the temperature of the transmitter is thus also taken into account when determining the reception time window in the receiver, it being possible in principle to only correct the transmission time based on the temperature in the transmitter.

In a first simple, but already effective embodiment of the present invention, the temperature of the transmitter is estimated as a function of the typical installation location with its typical ambient temperature range. Knowing the installation location, this estimate can also be made in the receiver alone, without the temperature in the transmitter actually having to be measured. The type of device in which the transmitter is located is decisive for the installation location. Heat cost allocators or heat meters are typically located in warmer environments than cold water meters or gas meters. The average ambient temperatures in the area of ​​such meters are based on experience and can therefore be used to estimate the temperature of the transmitter depending on the typical installation location.

According to the invention, additional information can also be taken into account, for example the season, the outside temperature or the room in a house in which the transmitter is installed. It is therefore likely that transmitters located in heat cost allocators will not be exposed to other temperatures in summer than receivers located in a comparable room. Then the temperature of the transmitter and the receiver can be assumed to be approximately the same. In the case of high outside temperatures, however, it can be assumed that transmitters installed under the roof or in an attic apartment have significantly higher temperatures than a receiver located on the ground floor or basement of a house. Using these or comparable considerations, the typical internal temperature at the location of the transmitter and the receiver can thus be estimated in order to determine a typical temperature difference between the transmitter and the receiver. Since additional information, for example whether a heater is switched on or off, is usually available on the receiver side, it is advisable in this case to use the additional information on the receiver side, especially since it is also used there cumulatively for all transmitters can be. Since the installation location is generally known, it can be taken into account in a simple manner both when adapting the reception time window and when adapting the transmission time.

The typical time behavior of the ambient temperatures can also be taken into account in the estimation in order to calculate the temporal position and duration of the next reception time window. It is known, for example, from a radiator that a heating process takes place faster than a cooling process. A radio heat cost allocator installed on a radiator will typically heat up faster at the start of a heating process than it cools down after the radiator is switched off (for example by turning off the thermostatic valve). Such processes can also be used according to the invention in connection with information as to whether a heater is switched on or off.

More precise information about the temperature of the transmitter can be obtained if the temperature of the transmitter and in particular of the time-determining element of its time control, ie typically of the watch quartz, is measured. The precise measurement allows a more precise coordination of the reception time window and the transmission time. This variant is particularly advantageous if temperature sensors are already present in the transmitters because of their intended purpose. This is the case, for example, with radio heat cost allocators, which typically each have a radiator and a room air-side temperature sensor. Depending on the construction and arrangement of the temperature sensors, the temperature of the watch quartz can be estimated as a weighted average of these two sensor temperatures.

Alternatively, of course, you can also attach your own temperature sensors that are specially designed to directly measure the temperature of the time-determining element in the time control or to simply estimate it. In a first variant, the temperature values ​​can therefore be measured in the transmitter and, for example, also transmitted to the receiver, which then uses the temperature values ​​when adapting the subsequent reception time window. With heat cost allocators or heat meters, these temperature values ​​are usually transmitted anyway, so that there is not even a separate data transmission effort. In a second variant, the measurement of the temperature in the transmitter can also lead to an adjustment of the transmission times. Since the receiver knows that a temperature-dependent correction of the transmission times takes place in the transmitter, the reception time windows can be shortened overall and optimally arranged around the nominal transmission time. This also saves a considerable amount of energy in the receiver. Of course, these two variants can also be combined.

According to the invention, the temperature of the receiver can also be measured, the temperature of the time-determining element, in particular its time control, ie typically the clock quartz, being determined. The temperature information in the receiver can also be used to optimally set up the reception time window to the time of transmission, ie to shift it if necessary. Due to the greater accuracy, the reception time window can be further shortened even when this variant is used, as a result of which the reception time window can be determined even more precisely. In particular, the temperature information in the receiver can be used to estimate the temperature difference between the transmitter and the receiver if the temperature information of the transmitter is also available. This variant is also advantageous if, for cost reasons, for example, it does not make sense to measure the temperature in each transmitter. From the temperature information of the receiver, conclusions can be drawn about the temperature of the transmitter and the particularly relevant temperature differences between the transmitter and receiver, possibly in combination with the installation site. Since, in typical data transmission systems, there are generally considerably more transmitters than receivers, the total costs are reduced by this last-mentioned variant, in particular if separate temperature sensors had to be provided in the transmitters.

The temperature of the time controls in the transmitter and / or receiver is expediently measured regularly, ie periodically in predetermined time periods. As a result, longer-term systematic influences, which are related, for example, to the temperature differences associated with the change of the season or to aging-related influences of the timers, are continuously recorded.

If a temperature characteristic of a timer is not known or if individual deviations from a generally applicable temperature characteristic are to be recorded, it is possible according to the invention for a particularly precise adjustment to determine and parameterize the temperature characteristic of the time control of a transmitter and / or receiver beforehand by means of measurements. This allows a particularly precise adaptation and definition of the reception time window to be achieved.

Furthermore, it is possible to determine the time deviations of the actual from the nominal transmission times in order to take the temperature of the transmitter into account. According to the invention, this can also be done standardized to the time interval to the last received data packet. Since the main influence in the shifting of the data telegrams is the temperature of the transmitter and / or receiver or whose time controls are, the temperature in the transmitter is also indirectly taken into account by taking into account the actually measured deviations from the nominal transmission times.

It can be advantageous to determine a distribution density function of the time deviations and to use this as the basis for determining a reception time window. For this purpose, a nominal reception probability can be determined using the distribution density function and a time window can be dimensioned so large that the desired reception probability is exactly achieved on the basis of the distribution density function. According to the invention, the time window can be selected to be minimally short, so that the area of ​​the highest probability in the probability density function is used for the time and the duration of the time window.

In order to increase the overall reception probability, after one or preferably several, for example two or three, unsuccessful reception attempts, the reception time window for the subsequent reception can be divided into two sub-reception time windows, one of which is before and one after the nominally provided or the reception time window to be applied using the above-described method. Possibly. it makes sense to set up a small overlap with the reception time window that is actually to be provided, so that no time gaps arise between the different reception time windows. If successful reception has taken place in a first sub-reception time window, the use of the second sub-reception time window can be omitted according to the invention.

The temporal length of the proposed sub-reception time windows do not necessarily have to correspond to the length of the nominal reception time window, but can be determined individually, for example on the basis of a desired probability, according to the distribution density function. With this variant of the invention, it is possible to restore data transmission between a special transmitter and the receiver despite a reduction in the reception time window even in the event of unsuccessful reception attempts if, for example, this did not occur due to unforeseen influences on the timer of the transmitter and / or the receiver.

According to claim 12, the invention also relates to a device for data transmission, in particular for wireless data transmission with possibly even battery-operated transmitters and / or receivers, with at least one transmitter, which has a time control for transmitting the data or Has data telegrams at specified transmission times, and with at least one receiver, which has a time control for switching on the receiver during a reception time window for the reception of the data or data packets transmitted by the transmitter. In order to maintain this synchronized data transmission, at least one time control has a correction device to take into account the gear differences of the time controls of the transmitter and the receiver. According to the invention, this correction device is set up to take into account the temperature of the transmitter or of the time-determining elements of its time control when determining the reception time window in the receiver. In particular, the correction device of the transmitter and / or of the receiver is set up to carry out the above-described method or parts thereof. For this purpose, the correction device can have a microprocessor suitably equipped in terms of program technology or have access to one. The microprocessor can then be arranged, for example, in the transmitter or receiver or in a device equipped with the transmitter or receiver, such as a consumption recording device.

At least one temperature sensor can preferably be provided on the time control of the transmitter and / or the receiver. This sensor can also be a temperature sensor arranged in the device that has the transmitter or receiver, which allows conclusions to be drawn about the temperature of the time control or its time-determining elements. A typical device for this is a heat cost allocator.

According to the invention, the method and the device are particularly suitable for use in measuring devices for recording consumption or consumption costs. In principle, it is suitable for all wireless transmission networks and receivers of time signal transmitters.

Further features, advantages and possible uses of the present invention also result from the following description of exemplary embodiments and the drawings. All of the described and / or illustrated features, alone or in any combination, form the subject matter of the present invention, regardless of their summary in the claims or their references.

Show it:<dl id="dl0001"><dt>Fig. 1</dt><dd>schematically a data transmission system implemented in a house according to the present invention;</dd><dt>Fig. 2</dt><dd>a transmitter and a receiver of the in <figref idref="f0001">Fig. 1</figref> represented system;</dd><dt>Fig. 3</dt><dd>the time relation of transmission time and reception time window in the prior art and according to the present invention in comparison;</dd><dt>Fig. 4</dt><dd>a typical characteristic curve for the temperature response in a watch quartz serving as a time-determining element;</dd><dt>Fig. 5</dt><dd>the arrangement of a reception time window relative to a distribution density function and</dd><dt>Fig. 6</dt><dd>the arrangement of two sub-reception time windows relative to a distribution density function after unsuccessful reception attempts.</dd></dl>

<figref idref="f0001">Fig. 1</figref> schematically represents a house 1 with a data transmission system realized for consumption recording with several battery-operated transmitters and receivers. The transmitters and receivers are integrated in consumption recording devices. In the example of the house 1 shown, these are heat cost allocators 3.1 to 3.5 arranged on radiators 2.1 to 2.5. The radiators 2.1 and 2.2 are on the ground floor, the radiators 2.3 and 2.4 on the first floor and the radiator 2.5 in the attic of house 1.

In the basement of house 1 there is a cold water connection 4 with a water meter 5 which, like the heat cost allocators 3, is each equipped with a battery-operated transmitter for transmitting the consumption data. In another basement room, a data collector 6 is provided, which has a receiver for receiving the data or data telegrams transmitted by the transmitters of the heat cost allocators 3 and the water meter 5.

As an example of the devices participating in the data transmission system, in <figref idref="f0001">Fig. 2</figref> a heat cost allocator 3 and the data collector 6 shown in detail. The heat cost allocator 3 has a transmitter 7 with an antenna 8 for the wireless transmission of data messages. The transmitter 7 is equipped with a time control 9, which serves to send out the data at fixed transmission times. For this purpose, the time control 9 has a clock quartz 10 as a time-determining element or timer. Furthermore, temperature sensors 11 are provided in the heat cost allocator 3, which measure the room air temperature and the radiator surface temperature in order to determine therefrom the radiator overtemperature which is important for the heating cost detection as the difference between the radiator surface temperature and the room air temperature.

The data collector 6 has a receiver 12 in order to be able to receive the data transmitted by the transmitter 7. For this purpose, the receiver 12 is equipped with a time control 9 comparable to the transmitter 7, which also has a time-determining element 10, in particular a watch quartz. According to the embodiments of the invention described here, a temperature sensor 11 is also arranged on the watch crystal 10 of the receiver 12 in order to measure the temperature of the watch crystal 10 of the receiver 12.

The timing control 9 of the receiver 12 switches on the receiver 12 during a reception time window for receiving the data, the timing control 9 having a correction device 13 or being connected to it, which takes into account the difference in the timing of the timing controls 9 of the transmitter 7 and the receiver 12 . The correction device 13 can, for example, be integrated in a computer, for example a microprocessor, which is provided in the data collector 6 anyway.

According to the invention, the correction device 13 is set up to take into account the temperature of the transmitter 7 or of the watch crystal 10 and its timing control 9 when determining the reception time window in the receiver 12. As a result, the time of transmission for the transmission of the data by the transmitter 7 of the heat cost allocator 3 can be predicted more precisely than in the prior art, which the temperature of the time control 9 or of the quartz watch 10 is not taken into account at the time of transmission.

In the prior art it is therefore necessary to make the reception time window longer since the inaccuracy is greater over the actual transmission time. This is in<figref idref="f0002">Fig. 3</figref> presented in a comparison of the prior art with the method proposed according to the invention.

In the upper part of the <figref idref="f0002">Fig. 3</figref> the temporal correlation of the transmission times 14 to the reception time windows 15 are shown schematically. The timer of a transmitter causes the transmitter to send a data telegram of a certain length at the nominal transmission time 14, which is indicated in the illustration by the width of the transmission time 14. After a certain time interval, a second transmission takes place at a further transmission time 14. The transmission times 14 are determined by a time control of the transmitter, which in particular has certain fluctuations depending on the temperature. The receiver must therefore open a reception time window 15 which takes into account certain temperature-related fluctuations. The uncertainty based on temperature fluctuations about the start of the transmission time 14 is shown by the arrow 16. Since the receiver cannot predict the transmission time 14 more precisely than indicated by the arrow 16, the reception time window 15 must be dimensioned in time so that a data telegram that is transmitted within the fluctuation range of the transmission time 14 represented by the arrow 16 is reliably received by the receiver can be.

According to the invention, the temperature of the transmitter 7 or of the time control 9 with the time-determining element 10 is now taken into account when determining the reception time window 15. Since the temperature of the transmitter 7 at the time of transmission 14 is at least approximately known, the overall uncertainty about the time of transmission 14 is reduced, which is reflected in a shorter arrow 16 which indicates the possible fluctuation range of the time of transmission 14. In the example shown, in which the transmitter 7 corresponds to one of the peak temperature of the characteristic <figref idref="f0002">Fig. 4</figref> has also been taken into account, a temperature change of the transmitter 7 starting from the apex of the in <figref idref="f0002">Fig. 4</figref> The characteristic curve shown can only be shifted towards a reduction in the clock frequency, so that the transmission time 14 can only be delayed, but cannot be anticipated. Therefore, taking this additional information into account, the overall fluctuation 16 of the transmission time 14, which is anyway smaller, only has to be taken into account with regard to a later data transmission. The range of fluctuation represented by arrow 16 is therefore, according to the invention, substantially smaller than in the prior art, so that the reception time window 15 can be made substantially shorter overall. This contributes to significant energy savings for the receiver 12.

Various variants of the present invention are explained again in detail below.

In a first embodiment of the method according to the invention, the type and the typical installation location of the transmitter 7 with its typical ambient temperature range are taken into account when dimensioning the reception time window 15. For example, in a system for wireless consumption recording in buildings 1, as in<figref idref="f0001">Fig. 1</figref> is shown, a data collector 6 is assigned electronic heat cost allocators 3, water meters 5 and possibly gas meters or other meters which transmit their measurement data to the data collector 6 by radio. Since the heat cost allocators 3 are mounted directly on the radiators 2, they are exposed to a larger temperature range than the water meters 5 or gas meters or the receiver 12 arranged in the basement. As a result, the difference in clock speed between transmitter 7 and receiver 12 and thus the deviation of the nominal or target transmission time 14 from the actual transmission time 14 within a transmission period can be greater for the heat cost allocators 3 than for the water meter 5 or a gas or electricity meter. In the receiver 12 of the data collector 6, the information about the type or the mounting location of the recording device to be received is stored, so that the receiver 12 can provide a shorter time window for a data packet of a water or gas meter 5 when calculating the temporal position and the duration of the next reception time window 15 than for an expected data packet of a heat cost allocator 3, that is typically subject to greater fluctuations. Due to the fact that the different types or Installation locations of detection devices 3, 5 can thus be reduced overall the reception standby time of the receiver 12. Additional information (for example summer, winter or spatial arrangement in the house under the roof or on the ground floor) can also be used to make average assumptions about a typical temperature, which, in addition to reducing the length of time of a reception time window 15, also shift towards an actual transmission time 14 can result.

In heat cost allocators 3, the use of 32.678 kHz watch quartz crystals as time-determining elements 10 of the time controls 9 is common. From these watch quartzes 10, a parabolic characteristic curve of the oscillation frequency versus temperature is known, as shown in FIG<figref idref="f0002">Fig. 4</figref> is shown. The peak of the characteristic curve is typically around a peak temperature of approximately 25 ° C. The oscillation frequency decreases with increasing and with falling temperature. If a heat cost allocator 3 sends out a data packet every approx. 15 minutes, for example, within a transmission period, ie the period between a transmission time 14 and the subsequent transmission time 14, the quartz frequency with a temperature increase from 25 ° C to 65 ° C, as can be done in the case of heating a radiator, by 50 ppm. Conversely, cooling the quartz from 65 ° C to 25 ° C will take at least one hour, so that within the transmission period of 15 minutes a maximum increase in the quartz frequency of 20 ppm can be expected. Thus, in the event of a temperature rise, a delay in the transmission of up to 45 ms per transmission period must be expected, in the case of a temperature drop, however, only a transmission which is premature by 18 ms compared to the case of a constant temperature.

In this example, the reception time window 15 is therefore selected such that it begins 18 ms before the (nominal) target transmission time expected without a temperature change and ends 45 ms after the nominal transmission time if there is no reception. Due to the generally valid consideration of the typical ambient temperature, in the case of a heat cost allocator 3, ie the slow cooling of a radiator 2, the reception time window 15 can be shortened compared to a conventional view and thus the average energy consumption of the receiver 12 can be considerably reduced.

According to another variant of the present invention, the temperatures of the time-determining elements 10 (watch crystals) in the transmitters 7 and / or the receivers 12 can be measured regularly. For this purpose, as in the case of the heat cost allocator 3, temperature sensors 11 used in any case can be used, which, due to their arrangement and typical empirical values, for example by weighted averaging, give a good estimate of the temperature of the time control 9 with the timer 10 (clock quartz). In the case of a data collector 6, which is typically not equipped with a temperature sensor 11, a temperature sensor 11 can be arranged directly on the watch quartz 10.

Various corrections are possible with this information. On the one hand, the time controller 9 of the transmitter 7 can, based on the known temperature, transmit the time 14 according to the known temperature line (<figref idref="f0002">Fig. 4</figref>) correct the time-determining components 10. This enables the receiver 12 to choose shorter reception time windows 15 and thus to save energy.

However, an alternative method embodiment is preferred and much more precise, in which the recorded temperatures of the time control 9 of the transmitters 7 are also transmitted to the receiver 12 and its correction device 13. They then serve as the basis for the calculations of the length and temporal shift of the next reception time window 15 for the respective transmitter 7.

If in the transmitters 7 watch crystals 10 with the in <figref idref="f0002">Fig. 4</figref> shown parabolic temperature characteristic are used and, for example, a quartz temperature of 25 ° was transmitted in the last received data packet of a transmitter 7, it is clear that the time base of the transmitter can not accelerate, but can only slow down. Therefore, in the reception time window 15 for the transmitter 7 of this example, no advance, but only a follow-up need be provided compared to the nominal transmission time, as shown in the lower illustration of<figref idref="f0002">Fig. 3</figref> has already been explained. With knowledge of the type of temperature characteristic of the time-determining components 10 in the transmitters 7, knowledge of the quartz temperature ϑ₀ at the time of the last transmission received and knowledge of the maximum possible temperature change since the last transmission received (ϑ_Max -ϑ₀, ϑ₀ -ϑ_min) the receivers can calculate the resulting relative quartz frequency deviations (Δ<i>f</i>_min to Δ<i>f</i>_Max) specifically calculate the optimal length and the temporal position of the reception time window 15. The corresponding values ​​and ranges are for illustration in<figref idref="f0002">Fig. 4</figref> drawn.

It is possible for each transmitter 7 to individually measure the frequency response of the time-determining clock quartz 10 above temperature, for example during production, for example by measuring the frequency at 25 ° C. and 60 ° C. This gives the exact position of the individual temperature characteristic of each Transmitter 7 is known and can be described in the case of a watch quartz 10 by individual parameters, for example the peak temperature and the slope. The transmitter-specific parameters determined in this way are either stored in the respective transmitter 7 and offset when the transmission time is corrected, or preferably stored in the receiver 12, which takes them into account when calculating the length and temporal position of the reception time window 15. The storage and calculation can take place in the correction device 13. It is also possible to store the characteristic curve parameters in the transmitter 7 and to transmit them once or regularly in data packets of the transmitter 7.

In a further variant of the proposed method, the length of the reception time window 15 is determined from statistical information about the deviations of the actual transmission times 14 from the nominal transmission times 14. It is advisable to normalize this time deviation to the time interval from the last received data packet. The standardized time deviation Δ<i>t</i> is then <maths id="math0001" num=""><math display="block"><mi mathvariant="normal">Δ</mi><mo>⁢</mo><mi mathvariant="italic">t</mi><mo>=</mo><mfrac><mrow><msub><mi>T</mi><mi mathvariant="italic">is</mi></msub><mo>-</mo><msub><mi>T</mi><mi mathvariant="italic">should</mi></msub></mrow><mrow><msub><mi>T</mi><mi mathvariant="italic">is</mi></msub><mo>-</mo><msub><mi>T</mi><mrow><mi mathvariant="italic">is</mi><mo>-</mo><mn>1</mn></mrow></msub></mrow></mfrac></math><img file="EP2150084A2_D0001.tif" /></maths>With <i>T_should</i> as a predicted target reception time, <i>T_is</i> as the actual time of reception and <i>T</i>_<i>is</i>-1 as the penultimate actual time of reception.

From a large population of determined values ​​for Δ<i>t</i> then a compression density function Φ (Δ<i>t</i>) determines how they are in <figref idref="f0003">Fig. 5</figref> as a function of the time deviation Δ<i>t</i> is shown. The reception time window 15 is now selected so that it has a normalized time interval of Δ<i>t</i>₁ to Δ<i>t</i>₂ corresponds. The probability<i>p</i>₁is then that a data packet is actually sent in this reception time window 15 <maths id="math0002" num=""><math display="block"><msub><mi>p</mi><mn>1</mn></msub><mo>=</mo><munderover><mo mathvariant="normal">∫</mo><msub><mrow><mi mathvariant="normal">Δ</mi><mo>⁢</mo><mi mathvariant="italic">t</mi></mrow><mn>1</mn></msub><msub><mrow><mi mathvariant="normal">Δ</mi><mo>⁢</mo><mi mathvariant="italic">t</mi></mrow><mn>2</mn></msub></munderover><mi mathvariant="normal">ϕ</mi><mfenced separators=""><mi mathvariant="normal">Δ</mi><mo>⁢</mo><mi mathvariant="italic">t</mi></mfenced><mo>ⅆ</mo><mi mathvariant="normal">Δ</mi><mo>⁢</mo><mi mathvariant="italic">t</mi><mn>.</mn></math><img file="EP2150084A2_D0002.tif" /></maths>

For the probability <i>p</i>₁ a nominal value, for example 80%, is then specified and the standardized time interval of Δ<i>t</i>₁ to Δ<i>t</i>₂ chosen so that it is minimal. So you have one for a reception probability<i>p</i>₁ optimally short reception time window of 80% 15.

Conversely, there is a probability of <i>q</i>₁ = 1 - <i>p</i>₁, in this example 20%, that the data packet is not sent in the reception time window 15 and therefore cannot be received either. The actual transmission time 14 was therefore either before or after the selected reception time window 15, provided there was no other reason for the non-reception, for example a disturbance on the transmission channel.

After a predetermined number of failed reception attempts, the reception time window 15 for the next data packet is divided into two sub-reception time windows 15a, 15b, as in FIG <figref idref="f0003">Fig. 6</figref> shown. The first sub-reception time window 15a represents an attempt “2a”, which assumes that the data packet missed in the normal reception time window 15 was sent before this reception time window 15. In a further attempt “2b”, the second sub-reception time window 15b assumes that the data packet missed in the original reception time window 15 was sent after this reception time window 15. Based on the probability distribution Φ (Δ<i>t</i>) a sub-reception time window 15a is now defined for the experiment "2a", which is based on the specification of a reception probability <i>p</i>_2<i>a</i> of, for example, 55% is optimized according to the method described above in such a way that the duration of the sub-reception time window 15a is minimal. The same procedure is followed for dimensioning the sub-reception time window 15b for the experiment “2b”. Of course, this attempt "2b" only has to be carried out if attempt "2a" has been unsuccessful. By calculating with the standardized time deviation Δ<i>t</i> it is automatically taken into account that from the last received data packet until the execution of the tests "2a" and "2b", further transmission periods may have elapsed in which the clock rate deviations may have been integrated.

In the numerical example above, the residual probability is that the data packet is not "1" in the reception time window 15 in any of the attempts (cf. <figref idref="f0003">Fig. 5</figref>) and "2a" or "2b" in the reception time windows 15a or 15b (cf. <figref idref="f0003">Fig. 6</figref>) was received, only about 4%. In this case, in order to maintain the synchronicity between transmitter 7 and receiver 12, a further reception attempt can be started, in which the sub-reception time windows are correspondingly expanded or arranged further away from the original reception time window 15.

The above-described division of the reception time window 15 into sub-reception time windows 15a, 15b is preferably not already carried out during a second reception attempt by one of the assigned transmitters 7, but only when several reception attempts have been unsuccessful. This procedure is particularly useful for radio transmission systems in an environment in which many other system-compliant and / or non-system transmitters use the same working frequency. This can result in data packets not being received due to collisions. A temporal drift of the transmission time 14 with respect to the desired transmission time 14 calculated by the receiver 12, which corresponds to a desired reception time, due to clock cycle deviations can only be considered here as a reasonably certain cause for the non-reception, if sufficiently often also in an undivided reception time window 15 Attempts to receive were made. With an assumed collision or interference probability of 1% in the radio channel after two unsuccessful attempts to receive, the probability that in the past two attempts a disturbance led to non-reception and therefore the next transmission time 14 does not deviate significantly from the desired transmission time 14 is 100 ppm. This low probability makes it seem sufficiently certain that a division of the reception time window 15 into sub-reception time window 15a, 15b is appropriate for the third reception attempt.

The statistical information on the probability distribution Φ (Δ<i>t</i>) the time deviation Δ<i>t</i> can also be determined in advance or fixed and stored permanently in the receiver 12 or its correction device 13. It is also possible to collect this statistical information during ongoing operation and to update it cyclically. At the beginning, a probability distribution fest (Δ<i>t</i>) are used, which is then replaced by the statistics determined in the data transmission system during operation. In a mixture of the two variants, it is possible, for example, to work with longer reception time windows 15 during the statistical collection phase and later on the determined time deviation statistics Φ (Δ<i>t</i>) can be switched with reception time windows 15 shortened according to the invention.

The statistical information on the probability distribution Φ (Δ<i>t</i>) of the time deviation can be determined and applied separately according to the type of the respective transmitter 7. For example, separate statistics for heat cost allocators 3, water meters 5, gas meters etc. are used in a system for recording consumption. It is also possible to collect and use these statistics for individual stations.

The invention described in detail above thus optimizes the average receiver energy consumption compared to known systems in that the prediction of the transmission times 14 is carried out more precisely with fewer tolerances. Due to the small tolerance, shorter reception time windows 15 are made possible with the same synchronicity security. The higher precision of the prediction is achieved by the collection and use of further parameters and / or statistical information.

The economic benefit lies in the fact that the data transmission system described enables the receiver 12 to operate on the battery for a considerably longer time, and that any battery changes that may be required therefore have to be carried out less frequently. Alternatively, smaller or fewer batteries can be used for a given period of use.

A further advantage lies in the fact that for a given receiver operating time and battery capacity, more transmitters 7 can be assigned to each receiver 12 of the system, since the reception time windows 15 are shorter for each transmitter 7. The latter is particularly economically advantageous if the number of receivers has not so far been determined by the range of the data transmission, but rather by the battery charge per assigned transmitter 7 or the duration of the reception time window 12 was determined. Possibly. the use of the method proposed according to the invention and the device provided therefor enables economically sensible battery operation of the receiver 12 in the first place.

Reference symbol list:

<dl id="dl0002" compact="compact"><dt>1</dt><dd>House</dd><dt>2.1 to 2.5</dt><dd>radiator</dd><dt>3.1 to 3.5</dt><dd>Heat cost allocator</dd><dt>4</dt><dd>Water connection</dd><dt>5</dt><dd>water meter</dd><dt>6</dt><dd>Data collector</dd><dt>7</dt><dd>Channel</dd><dt>8</dt><dd>antenna</dd><dt>9</dt><dd>Time control</dd><dt>10</dt><dd>Watch quartz, time-determining element</dd><dt>11</dt><dd>Temperature sensor</dd><dt>12</dt><dd>receiver</dd><dt>13</dt><dd>Correction device</dd><dt>14</dt><dd>Time of transmission</dd><dt>15</dt><dd>Receive time window</dd><dt>15a, 15b</dt><dd>Under reception time window</dd><dt>16</dt><dd>Fluctuation in the transmission time</dd></dl><dl id="dl0003" compact="compact"><dt>Φ (Δ<i>t</i>)</dt><dd>Distribution density function of the time deviations Δ<i>t</i></dd></dl>