Method and device for the optimisation of cellular wireless message networks

- Object of the invention -

p0001The invention relates to a method for optimizing the throughput of messages in cellular wireless communications networks, according to the preamble of claim 1, and to a device for implementing the method according to the preamble of the adjoining claim <i>13</i>.

- State of the art -

p0002With the introduction of cellular wireless communications networks, there has always been the problem of finding configurations that enable the best possible utilization of the available resources of hardware and frequency bandwidth during the setup, later expansion and operation of the networks. The aim is above all to maximize the throughput of messages with acceptable network quality and sufficient network availability or network coverage. The actual target size can be different. For example, the blocking rate can be minimized, as exemplified in FIG<patcit id="pcit0001" dnum="USPS5826218A"><text>U.S. Patent 5,826,218</text></patcit> As well. The<patcit id="pcit0002" dnum="WO9853621A"><text>WO 98/53621</text></patcit> In turn, is based on measured piloting performance values. In order to achieve the objectives of network optimization, different parameters of the network configuration are usually adapted locally. This is, for example, the adaptation of the antenna diagram or of the pilot power as in the<patcit id="pcit0003" dnum="USPS5276907A"><text>U.S. Patent 5,276,907</text></patcit> Or the receiver sensitivity as in the <patcit id="pcit0004" dnum="WO0137446A1"><text>WO 01/37446 A1</text></patcit>.

p0003In mobile radio networks, which generally only transmit information from fixed stations to mobile users, comparable to conventional broadcasting, methods are applied which optimize the network configuration mainly with regard to the network coverage to be achieved based on the radio propagation properties in the environment. Basically, radio propagation simulations and / or measurements of the average reception strength as well as the quality of service at certain locations in the network are used. The basic approach for such methods is described in<nplcit id="ncit0001" npl-type="b"><text>Norbert Geng, Werner Wiesbeck, "Planning Methods for Mobile Communication", Springer-Verlag Berlin, 1998, ISBN 3-540-64778-3</text></nplcit> Described.

p0004In telecommunications networks, such as the GSM Staudard, which use many frequency bands, the frequency bands are assigned to the individual radio cells of the network. A suitable distribution of the frequency bands on the radio cells plays a decisive role for interference reduction and thus throughput maximization. Such methods for frequency planning of the network are often combined under the term "Automatic Frequency Planning" (AFP).

p0005With the increasing importance of various mobile data services and the development to mobile networks of the third generation, eg according to the UMTS standard, which is developed within the framework of the international "3rd Generation Partnership Project" (3GPP), a network planning with regard to the expected Specific subscriber traffic. The various services offered have very different characteristics in terms of their burstiness, the circuit or packet-switched mode of transmission, the data rate, the average transmission power and the symmetry / asymmetry between downlink and uplink. In addition, the partial very different physical transfer characteristics of the individual services must be taken into account since they are generally used for different transport channels.

p0006<i>The</i><patcit id="pcit0005" dnum="US5561841A"><text>U.S. Patent 5,561,841</text></patcit><i>Describes an optimization method for a radio network. However, the optimization method uses only data generated from a network simulation to optimize the adjustable parameters. This leads to the fact that this optimization process works quite slowly. First, it is necessary to generate data which are then used in the optimization process. With the use of simulation data for the optimization, only unsatisfactory results are achieved during network optimization</i>.

p0007These special requirements for mobile radio networks of the third generation are not fulfilled by previously known and common methods and devices for the purpose of radio network optimization.

- Object of the invention -

p0008From the point of view of network operators, it seems important to provide all offered services with approximately equivalent subjective quality and equal coverage. It is therefore necessary to apply a network optimization process which provides the existing services in the communications network with sufficient quality throughout the supply area.

p0009It is an object of the invention to quickly and inexpensively and efficiently optimize the existing network configuration of a message network, taking into account the heterogeneous requirements of various services for message transmission, in that subsequently an at least equally good but usually improved network configuration with regard to throughput and network quality while maintaining the existing network coverage Available. Any modification of cell site varieties is thereby excluded as an optimization measure.

- Solution of the task -

p0010<i>This object is achieved by the method mentioned initially with the features of the characterizing part of claim 1. Furthermore, the object is achieved by means of a device of the type mentioned at the beginning with the characterizing features of claim 13.</i>

p0011Embodiments of the method are the subject of claims 2 to 12.

p0012Embodiments of the device are the subject of claims 14 to 18.

p0013The object is achieved according to the invention in particular by the fact that an improved network configuration is determined by means of single or multiple, in this case certainly iterative modification of certain settings of the network configuration, in particular the antenna alignment or the transmission power of a reference signal, eg a pilot channel, in cells selected by the method Which is compensated as much as possible by the radio network area-related subscriber traffic in the network area under consideration. In the case of an ideally balanced traffic distribution between the radio cells, the traffic at the given cell locations can be optimally dissipated. The method is supported by a graphical representation of the cell surfaces colored with colors which correspond to the traffic volume of the respective radio cell. This allows problem areas with high traffic volumes to be easily identified and the network configuration at these locations can be specifically adapted according to the invention. Another possibility of graphically supporting the identification of problem regions with regard to other sizes in a mobile radio network is described in the<patcit id="pcit0006" dnum="WO0028756A"><text>WO 00/28756</text></patcit> Suggested.

p0014The method according to the invention can in principle be used for any cellular wireless radio networks. In particular, it is suitable for third-generation networks which operate, for example, according to the UMTS standard, and which offer several different services for message transmission to the mobile subscribers. It is an advantage of the method that any number of offered services in the communications network can be considered with different requirements in the optimization of the network configuration.

p0015A further advantage of the method is the high speed against other methods with which results can be obtained. The reason for this is that the method is purely based on measurement data or planning data of the radio network or a combination of both, which are already present as a rule, since they are required in the planning process. No simulation of the funnel network behavior within the iterative optimization process is required, as is the case, for example, in the<patcit id="pcit0007" dnum="DE19619205A1"><text>DE 196 19 205 A1</text></patcit> the case is. The process can also be easily automated, which allows a computational implementation. The optional verification according to the method according to the invention adapted network configurations is independent of the actual optimization method, whereby the validation method is freely selectable and also interchangeable. In addition, validation gains much more meaning through this independence and is possible on the other hand without complex field measurements. The accuracy of the method can be increased arbitrarily by including additional data from the simulative validation of the adapted network configuration as well as further measurement data from the real radio network in the assessment of the achieved network quality.

p0016Advantageous embodiments and variants of the method as well as of the device are subject matter of the subclaims.

- Explanation -

p0017The invention relates to a method for optimizing the throughput of messages in wireless cellular network networks, as well as to a device for implementing the method. These are in particular cellular wireless message networks with mobile subscribers, which simultaneously offer a service or a mixture of several services for message transmission to one or more subscribers, as is the case, for example, in mobile radio networks according to the UMTS standard. An example configuration of such a message network with four base stations B110, B120, B130 and B140 is shown in FIG<figref idrefs="f0001">illustration 1</figref> Respectively.

p0018It is a crucial feature of the invention that no extensive simulations, examinations or measurements are necessary to carry out the method. Rather, the method is based on data that is generally available in the network planning and optimization process anyway. This ensures a fast implementation of the procedure, which is oriented to the requirements of network planning. The general procedure of the method is shown in the block diagram of<figref idrefs="f0002">Figure 2</figref> Respectively.

p0019According to step 210, <figref idrefs="f0002">Figure 2</figref> It is assumed that as input for the method location-related data are available for the message network or for the network section to be optimized with respect to the subscriber traffic and the radio network coverage. In this case, it is necessary for these data to be provided with sufficient spatial resolution in order to provide a larger number of values ​​of the variables mentioned for the coverage area of ​​any radio network of the message network. A suitable, but not mandatory, arrangement of these data is in the form of a rectangular matrix consisting of many relatively small rectangular surface pieces, each of which is assigned a value of the variables mentioned.

p0020The participant traffic is the traffic volume per service offered in the news network, which was determined in different places. Participant traffic data can be obtained either by measurements, computational or analytical methods or simulations, or by any combinations thereof.

p0021The network coverage required as input is formed by the cell-related reception power of a reference signal radiated with constant transmission power from each radio cell, which was determined at different locations. The reference signal can, for example, be a pilot channel, as is used in many mobile radio networks, inter alia, for the radio cell identification for the subscriber terminals. If no such reference signal exists in a communications network, a virtual reference signal, which is emitted by each radio cell of the message network with the same transmission power, can also be accepted. In the same way as in the case of subscriber traffic, the data acquisition of the reference signal reception performance can be effected both by measurements, by calculation or analytical methods or by simulations, as well as by any combinations thereof. Frequently, path-dependent electrical field strength data from each base station location, for example, in the form of a matrix, are provided for the region to be optimized, with the aid of which the location-related reference signal reception power is calculated by taking into consideration the transmission power of the reference signal, the antennas used and other radio propagation losses such as a body attenuation can be.

p0022At each location where reference signal reception power data is available, the radio cells with the best reception power are determined in a model of the communications network. In this way, an assignment of area pieces to radio cells arises. All the patches associated with a particular radio cell are considered in their entirety as a radio cell surface (hereinafter, also only a radio cell). Due to the physical propagation conditions in the radio channel, which apply to all services offered in the communications network, as well as by the cell exchange algorithm for ensuring subscriber mobility, which is usually based on the received power of the reference signal measured at the subscriber terminals, the radio cell areas determined in the communications network model can be used as a good approximation for The radio cells actually occurring in a communications network are used. In the exemplary section of a mobile phone network to<figref idrefs="f0001"><b>illustration 1</b></figref> The radio cells, which are arranged around the base stations B110, B120, B130 and B140 in the manner just described, are drawn in with solid lines.

p0023The aim of the method is to improve the throughput of messages in the communications network by providing the best possible distribution of the offered subscriber traffic between the radio cells. It has been shown that the subscriber traffic of all services can be dissipated with the best quality if there is a balanced distribution of traffic between the radio cells. To this extent, this method aims at establishing this balance of traffic between the radio cells. In order to achieve this goal, certain radio cell surfaces must therefore be displaced relative to one another in order to be able to exchange the otherwise unchanged local fixed traffic volume between the radio cells.

p0024Subscriber traffic per cell area is generally determined by accumulating the present traffic values ​​within the respective radio cells in the model of the communications network. Since, in the case of several services, the traffic volume per service can be very different at a certain location, the traffic values ​​of different services must be added in a suitably weighted manner. The weighting depends on the radio interference generated by the respective service and the resource utilization. Therefore, depending on the radio system, parameters such as the system bandwidth, the channel bandwidth, the ratio of signal power to interference power to be achieved, the spreading factor, the average or maximum data rate, the activity factor, the code limitation, etc., flow into the traffic weighting. In<figref idrefs="f0001"><b>illustration 1</b></figref> The summed traffic values ​​per cell are indicated by V110, V120, V130 and V140.

p0025After the cell-related values ​​of the subscriber traffic are ready, first of all, according to step 220 in FIG <figref idrefs="f0002"><b>Figure 2</b></figref> An assessment of the achieved approximation to the desired broadest possible balance of the subscriber traffic between the radio cells takes place. To this end, as a possible variant of the invention, a graphical representation of the examined mobile radio network, similar to FIG<figref idrefs="f0001"><b>illustration 1</b></figref><b>,</b> Can be used as an aid, the individual radio cell surfaces each being filled with a color assigned to the subscriber traffic carried in them <figref idrefs="f0002"><b>Figure 2</b></figref> A sufficient balance of the subscriber traffic between the radio cells has already been achieved, the method can be terminated and the found network configuration can be determined in the simplest case according to step 270 in FIG <figref idrefs="f0002"><b>Figure 2</b></figref> As an improved planning basis for network planning. If this state has not yet been reached, this representation is used in step 240 of the method<figref idrefs="f0002"><b>Figure 2</b></figref> The fast finding of problem zones, in which neighboring radio cells lead very different traffic. In another variant of the invention, the identification of the problem areas can also be carried out automatically by means of a computing technique in which differences in the subscriber traffic values ​​between adjacent radio cells are examined. This automated step is then integrated into the overall optimization method consisting of steps 220 to 250<figref idrefs="f0002"><b>Figure 2</b></figref>, For minimizing the differences in the subscriber traffic values ​​between the radio cells.

p0026A flow following step 250 in FIG <figref idrefs="f0002"><b>Figure 2</b></figref> Subsequent local shifting of radio cell surfaces can be achieved by the modification of certain network configuration parameters on individual network elements in the model of the communications network, in particular by a suitable adaptation of the antenna inclination angle or the antenna height of the particular antenna diagram used, or, for example, by an adaptation of the transmission power of the reference signal on selected radio cells . Since, as described above, all these parameters enter into the determination of the reference signal reception power, which is of decisive importance for the formation of the radio cell surfaces, the radio cell surfaces can thus be changed in this way. For example, in the network configuration of<figref idrefs="f0001"><b>illustration 1</b></figref> A locally limited high subscriber traffic volume V110 in the cell of base station B110 is ascertained, a suitable adaptation of network parameters should be made according to the invention, which results in a reduction of this cell area and thus a transfer of parts of the traffic from this cell into neighboring cells. After the network configuration has been adjusted, the displacements of the cell surfaces shown by dashed lines have been adjusted. As a result, the fixed partial interchanges Va112, V113 and V114 were subtracted from the traffic V110 of the cell at the base station B110. These were divided according to the figure to the cells of the base stations B120, B1330 and B140, which now have the traffic values ​​V120 + V112, V130 + V113 and V140 + V114, respectively. The achieved network configuration is better if, due to the change made, the subscriber traffic values ​​of the individual radio cells are more balanced than before.

p0027It is also possible to modify the radio cell areas iteratively in order to achieve the greatest possible balance of subscriber traffic between the radio cells so that the original network configuration can be approximated step by step to an optimized network configuration. In an iteration step, only one or a few parameters of the network configuration are changed. Thereafter, the approach to the desired balance of the subscriber traffic between the radio cells is repeated again according to steps 220/230 in FIG<figref idrefs="f0002"><b>Figure 2</b></figref> , The new radio cell areas being based on the changed network configuration. From the result of the check, the necessary modification of the network configuration is derived in the next iteration step. This procedure can also be automated by means of computational techniques. The step-by-step approach can be terminated if sufficient balancing of the subscriber traffic between the radio cells has been achieved. In particular, it is a criterion that the traffic in the radio cells with the originally highest share traffic was significantly reduced compared to the mean subscriber traffic in the wireless network model. Such a reduction in the traffic volume also reduces the local interference level, as a result of which the network quality can be improved.

p0028If a network configuration is achieved which has a considerably more balanced distribution of the subscriber traffic between the radio cells than the source configuration, as an advantageous variant of the invention, the result of the network adaptation according to step 260 can be determined <figref idrefs="f0002"><b>Figure 2</b></figref> Be validated. The validation can take place, for example, by a network simulation. In this case, it is expedient to strive, but not absolutely, to take into account the time behavior of the network in this simulation in order to be able to estimate the network quality. As a measure of the network quality, the number and the locations of blocked connection attempts or disconnected connections, the occurring error rate and delay of connections or the throughput per connection or radio cell can be used, among other things. Of course, sufficient network coverage still needs to be ensured after the network adaptation, ie, the different services must be available at the predetermined locations.

p0029Results of the radio network validation can subsequently be used as a further advantageous variant of the invention in order to use them in a renewed optimization of the radio network to increase the accuracy of the method. Such variables can be, in particular, the actual subscriber traffic that actually occurs in the network simulation, but also the quantities expressing radio-technical interference. The latter includes, among others, the so-called "noise rise", which characterizes the proportion of the radio interference caused by other subscribers locally caused by other subscribers at the cite-interference interference occurrence of the radio cell in the uplink. This size provides interesting indications for local network utilization, especially for CDMA networks. Information on the local occurrence of call rejections, connection breaks or failed cell-switching tests can also be used as boundary conditions for subsequent optimization of the radio network. In addition, quantities for assessing the cell separation, such as the ratio of the transmission power of the reference signal to total interference power per radio cell, are also suitable for this refinement of the method according to the invention.

p0030Likewise, in a further variant of the invention, such network simulations mentioned above can already be applied to the original network configuration beforehand in order to enable comparison with the optimized reconfiguration with regard to the network quality. As a result, the aforementioned additional variables, from the network simulation, can also flow into the method according to the invention. In addition to simulatively determined traffic values ​​as a direct input variable, further boundary conditions for the method can be set. This concerns, for example, the various reasons and the local limitations of connection breaks and rejections, the analysis of which provides interesting information for a fine control of the method according to the invention. As a variant to this, measured values ​​of further bursts from the real message network, in particular the aforementioned variables, can also be included in the initial evaluation of the original network configuration.

p0031In order to make use of the improved network configuration determined in the method according to the invention in the communications network, the method is integrated into the device according to the invention <figref idrefs="f0003"><b>Figure 3</b></figref> embedded. For a more detailed explanation, a distinction is made between the communications network in the radio access network 310 and the message switching network 330, which can communicate with one another via a bidirectional interface 320. The radio access network relates to the part of the communications network, which establishes the direct radio link to the mobile subscribers, that is to say it essentially contains the base stations of the radio network and possibly further components, provided these are necessary for direct communication with the subscribers. The message switching network forms the networking of the base stations in the radio access network, including switching stations and gateways to other networks or service providers. Via the interface 320, both the transmitted messages of the subscribers as well as various signaling and control information can be transmitted in both directions. A database 350, in which location data, configuration data of the individual network elements, measurement data of different sizes from the communications network, counter data from the network elements, callstististics, planning data, etc., are combined, is provided for the message network. The database symbol in<figref idrefs="f0003"><b>Figure 3</b></figref> Initially only a logical summary of all this data - depending on the implementation by a network operator, this data collection can also be realized in several databases which are linked as required. Data from the message network can be stored in the database at any time via the bidirectional interface 340, as well as new network configurations have been imported into the message network. In addition, measured variables and parameters from the message network can optionally be stored in the database with a mobile measuring device 399 (not necessarily motorized, as shown in the figure).

p0032For the implementation of the method according to the invention in the device, there is a computing device 370, which contains a computing-technical implementation of the radio access network in the form of a radio network model 380. This model reflects the initial reconfiguration, which was loaded from the database via the bidirectional interface 360. The inventive method 390 now operates in the computing device<figref idrefs="f0002"><b>Figure 2</b></figref> On the basis of the wireless network model to determine an improved network configuration. The changes to the network configuration are carried out in the computationally implemented radio network model 380, so that the improved configuration is then directly set there. If necessary, the optimization of the message network can be controlled and monitored by a radio network optimizer. However, this influence can be largely limited by the above-described automation of the method.

p0033The improved network configuration determined can now be transferred via the interface 360 ​​to the database and from there to the news network. By means of measured values ​​from the message network 310, 330 itself, or via a measuring device 399, the achieved improvement in the throughput and the network quality can finally be judged. Such measured variables can, in turn, also be played into the computing device in order to contribute to the validation of the improved network configuration.

p0034Since, in a message network, especially in the case of several offered services, the local subscriber traffic volume per service can also change quite rapidly, ie several times over a day, it is expedient to adjust the network configuration in a timely manner to the changed conditions with regard to the traffic volume. This can be done with a variant of the device according to the invention, whereby the high speed of the underlying method is advantageous. In this variant, the current state of the network configuration and of the traffic volume, which are stored in the database 350, is transferred to the computing device 370 and processed there immediately by the method according to the invention. The network configuration, which is better adapted to the current network conditions, is then transferred back into the database, from where it is transmitted directly to the message network and activated there. In this way, it is possible to react relatively quickly to a changing subscriber traffic volume in the message network.

p0035Although the present invention has been described in detail with all its advantages, it is to be understood that various other modifications and variations of the invention may be made within the scope of the foregoing remarks without departing from the spirit and scope of the invention as set forth in the appended claims Would be.

p0036The exemplary embodiment described in the following is illustrative of the method according to the invention for the optimization of wake-up networks by means of a selected section of a UMTS / FDD radio network, the general features of which are described, for example, in <nplcit id="ncit0002" npl-type="b"><text>H. Holma and A. Toskala (eds.), WCDMA for UMTS, 2nd ed., John Wiley &amp; Sons, Chichester, UK, 2002, ISBN 0-470-84467-1</text></nplcit> Are described. The starting point is about 10km² In which a total of 15 transmission stations with 3 cells each were arranged and configured according to today's customary and prior art planing methods. For each sending station, the path loss of the receiving locations (path loss matrix) arranged in a matrix with a spatial resolution of 10 m according to the current state of the art is calculated by means of radiation-optical propagation models with the aid of a detailed 3D environmental model. Furthermore, for a language service (<i>Speech</i>) And a WWW service (<i>WWW</i>) Is based on location-dependent traffic density in the form of traffic marrices with a spatial resolution of 25m.

p0037In the exemplary embodiment, the output network configuration is analyzed with regard to the capacity and quality of the message transmissions with the aid of a dynamic UMTS system simulator. The radio network is then optimized using the method according to the invention and the changed network configuration is subjected to a new dynamic analysis with respect to changes in capacity and quality.

p0038For each cell, the resulting equivalent traffic is calculated on the basis of the underlying traffic models and the present location-specific traffic density information. The local supply area of ​​each cell resulting from the wave propagation conditions and the network configuration can be characterized in color, the color code corresponding to the previously calculated equivalent traffic in Erlang. This is not the absolute value but the relative deviation of the values ​​of adjacent cells.

p0039For spatially adjacent cells, a very different traffic volume to be derived is to be expected. After analyzing the network configuration with the aid of a dynamic system simulator with respect to capacity and quality, according to the invention, by means of iterative variation of the antenna orientations and the transmission power of the reference signals, a wireless network configuration is determined based solely on the given location-related traffic and radio propagation data, which is a much anticipated traffic volume with regard to the output configuration Respectively.

p0040It can be shown that the network configuration determined with the inventive optimization method has a more balanced distribution of the expected traffic volume on the cells. This should make better use of the available global wireless network capacity.

p0041It can be seen that with the aid of the optimization method according to the invention, the original blocking and connection termination probability can be reduced in the example. Virtually all service requirements with optimized network configuration can be authorized and operated. In the exemplary embodiment, an increase in the connections served simultaneously by the radio network can be demonstrated under the described basic conditions of 11%.