CN1121805C - Method and apparatus for directional radio communication - Google Patents

Abstract

The present invention relates to a method for carrying out directional wireless communication between a first station (BTS4) and a second station (MS), which comprises the following steps: the first station (BTS4) receives a plurality of continuous signals sent by the second station (MS); all of the signals can be received in at least one direction of a plurality of different directions; at least one parameter value is determined for each signal of a plurality of serial signals received by the first station (BTS4) from the second station (MS); at least one parameter value is selected for signals needed to be sent from the first station (BTS4) to the second station (MS); the selection of at least one parameter value of the signals needed to be sent by the first station (BTS4) depends on at least one parameter determined value of a plurality of serial signals.

Description

Method and device for directional wireless communication

The present invention relates to a method and apparatus for directional wireless communication in which signals between a first station and a second station can be transmitted only in a specific direction. In particular, the invention can be applied to cellular communication networks employing space division multiple access, but the invention is not specific to such systems.

Currently implemented cellular communication networks are configured with a base station (BTS) which transmits signals through the cell or cell sector it serves to a designated mobile station (MS), which may be a mobile phone. However, a Space Division Multiple Access (SDMA) system is now proposed. In a space division multiple access system, instead of transmitting a signal to a given mobile station through a cell or cell sector, a base station transmits a signal only in the direction of the beam from which the mobile station signal was received. The SDMA system also enables the base station to determine the direction from which the mobile station signal was received.

SDMA systems have many advantages over existing systems. In particular, since the beam transmitted by the BTS can only be transmitted in a certain direction and thus can be relatively narrow, the power of the transceiver can be concentrated on the narrow beam. This is believed to result in a better signal-to-noise ratio for both the signal transmitted from the base station and the signal received by the base station. Furthermore, because of the directivity of the base station, the signal-to-interference ratio of the signal received by the base station can be improved. In addition, in the direction of transmission, the directivity of the BTS allows energy to be concentrated on a narrow beam, so that the signal transmitted by the BTS reaches a distant mobile station with a lower power level than a conventional BTS. This also allows the mobile station to effectively operate in areas farther from the base station, which in turn means that the size of each cell or cell sector of the cellular network can be increased. Because the cell size increases, the number of required base stations can be reduced, thereby reducing network costs. SDMA systems typically require multiple antenna elements to achieve the desired number of different beam directions in which signals can be transmitted and received. The provision of multiple antenna elements increases the sensitivity of the BTS to received signals. This means that a larger cell size will not negatively affect the BTS' reception of signals from mobile stations.

The SDMA system also increases the capacity of the system, that is, the number of mobile stations that the system can support simultaneously increases. This is because the directional nature of the communication causes the BTS to experience less interference from mobile stations in other cells using the same frequency. When communicating with a designated MS in the associated cell, the BTS causes less interference to other mobile stations using the same frequency in other cells.

Finally, it is believed that an SDMA system can simultaneously use the same frequency to transmit information to two, or even multiple, different mobile stations located at different locations in the same cell. This greatly increases the traffic that the cellular network can carry.

SDMA systems can be implemented in both analog and digital cellular networks and can incorporate various existing standards such as GSM, DCS 1800, TACS, AMPS and NMT. SDMA systems can also work with other existing multiple access techniques, such as Time Division Multiple Access (TDMA), Code Division Multiple Access (CDMA) and Frequency Division Multiple Access (FDMA) techniques.

One problem with SDMA systems is the need to determine the direction in which to send the signal to the mobile station. In certain circumstances, relatively narrow beams may be used to transmit signals from base stations to mobile stations. Therefore, a fairly accurate estimate of the direction of the mobile station is required. It is well known that signals from a mobile station typically follow multiple paths to a BTS. These paths are generally referred to as multipaths. Thus, a given signal transmitted by a mobile station may be received by the base station from more than one direction due to multipath effects.

In general, the beam direction that the BTS uses to transmit signals to the mobile station is determined based on information corresponding to bursts of data that the BTS has previously received from a given MS. Since this decision is based on received information corresponding to only one subframe, problems arise if, for example, a data subframe transmitted by a mobile station is superimposed with strong interference.

Another problem is to determine the direction in which the BTS sends signals to the mobile station based on the uplink signals the BTS receives from the mobile station. However, the frequency of the downlink signal sent from the mobile station to the BTS is different from the frequency used by the BTS to send the signal to the mobile station. The difference in frequencies used for uplink and downlink signals means that the behavior of the channel in the uplink direction may differ from the behavior of the channel in the downlink direction. Thus, the optimal direction determined for uplink signals is not always the optimal direction for downlink signals.

A method for transmitting pilot signals in a CDMA cellular radio system is disclosed in WO 96/37969. The method involves that the first station receives multiple signals from the second station, determines a parameter value for each received signal, and selects the parameter value for the signal to be sent according to the value of the parameter of the received signal. The method continuously searches for the best signal, using multiple phasing devices to determine the properties of the wireless environment. A directional wireless communication method based on a similar principle is disclosed in US 5,515,378.

It is therefore an aim of certain embodiments of the present invention to address these difficulties.

According to a first aspect of the present invention, there is provided a method for directional wireless communication between a first station and a second station, the method comprising the following steps:

said first station receives a plurality of consecutive signals from said second station, each said signal being receivable in at least one of a plurality of different directions;

determining a value for at least one parameter for each of the plurality of serial signals received by the first station from the second station in succession; and

selecting a value of at least one parameter for a signal to be transmitted from said first station to said second station, the selection of said value of at least one parameter of a signal to be transmitted by the first station being dependent on said plurality of serial signals said determined value of said at least one parameter, wherein said selecting step comprises applying a weighting pattern to said plurality of serial signals.

By basing the parameters of the signal to be sent from the second station to the first station on the parameters of a number of signals previously received from the first station, problems caused by eg strong interference in recently received signals can be reduced.

Preferably said step of determining a value for at least one parameter comprises determining the or each direction of each of the plurality of serial signals, said step of selecting comprising selecting at least one direction in which the signal is sent from the first station to the second station , the selection of the at least one direction depends on the determined directions of the plurality of serial signals. By basing the direction or each direction in which a signal is sent from a first station to a second station on a plurality of signals received from the second station, the probability that a signal sent by a first station is received by a second station can be increased.

Said step of determining at least one parameter value optionally and/or additionally comprises determining the strength of each of said plurality of serial signals, said selecting step comprising selecting a signal to be transmitted to said second station strength, the selection of the signal strength being dependent on the strengths determined for the plurality of serial signals. By basing the strength of the signal that needs to be transmitted from the first station to the second station on a number of signals received from the second station, the probability that the signal strength is the correct value can be increased. If the signal strength is too low, the second station may not be able to receive the signal, while if the signal strength is too high, the possibility of interference is unnecessarily increased. It should be understood that in embodiments of the present invention where signals are sent in multiple different directions, the strengths of the signals in these different directions may be different.

Preferably, the selecting step includes applying a weighting pattern to said plurality of serial signals. The term weighting scheme also includes algorithms providing weighting functions. The weighting pattern may be a uniform weighting pattern giving each signal of the plurality of serial signals the same weight. The weighting mode may also be that a newer received signal among the plurality of serial signals is given a greater weight than an earlier received signal among the plurality of serial signals. The weighting pattern can be an exponential or a linearly increasing weighting pattern. These are just two of the possible modes. Any other suitable mode can be used. Alternatively, the weighting pattern can be defined by an algorithm. It should be appreciated that in some embodiments, a weighting pattern is applied to the values determined for the parameters.

The selection means preferably selects one of the plurality of weighting patterns according to the radio environment. The weighting scheme can be as described above. For example, in a static or slowly changing wireless environment, a consistent weighting pattern may be used because it is expected that the determined direction or each direction or strength of the plurality of serial signals will generally remain constant across the serial signals. Alternatively, if the wireless environment is a rapidly changing wireless environment, a linear growth or exponential weighting pattern may be used. Using these two weighting modes, the previously determined directions of the plurality of received serial signals or the influence of each direction or strength on the selected beam direction is negligible.

Preferably, the method further comprises the step of: the first station defines a plurality of beam directions for transmitting the radiation beam, wherein each of said beam directions is individually selectable.

According to a second aspect of the present invention, there is provided a first station for directional wireless communication with a second station, the first station comprising:

receiver means for receiving a plurality of consecutive signals from said second station, each said signal being receivable in at least one of a plurality of different directions;

means for determining the value of at least one parameter for each of the plurality of serial signals received by the first station from the second station;

a transmitter device for transmitting a signal from the first station to the second station; and

control means for controlling said transmitter means, said control means selecting the value of at least one parameter for a signal to be transmitted from said transmitter means, said selection of at least one parameter value being dependent on said plurality of serial signals A determined value of said at least one parameter, wherein said control means applies a weighting pattern to said plurality of serial signals.

Preferably said determining means determines the direction or each direction of each of the plurality of serial signals, and the control means selects at least one direction in which the transmitter means transmits the signal, said at least one direction being selected depending on said plurality of serial signals. The determined direction or each direction of the line signal. Said determining means optionally and/or additionally determine the strength of each of said plurality of serial signals, the control means selects the strength of the signal transmitted by the transmitter means, said signal strength being selected depending on the signal strength for said plurality of serial signals The strength determined by a serial signal.

The control means may apply a weighting pattern to the plurality of serial signals.

Storage means may be provided for storing said determined parameters for each of the plurality of serial signals.

The receiver means and the transmitter means may comprise antenna arrays for providing a plurality of signal beams in a plurality of different directions. The antenna array may comprise a phased antenna array, or may comprise a plurality of different antenna elements, each element providing a beam in a prescribed direction. Two different arrays can be provided, one for receiving signals and one for transmitting signals. Signals can also be sent and received simultaneously by a single array.

The transmitter device can provide the radiation beam in a plurality of beam directions, wherein each beam direction is individually selectable. The intensity of each beam direction is also preferably individually selectable.

The invention is particularly applicable to cellular communication networks. In such a network, the first station and the second station may be a base station and a mobile station, respectively. However, it should be understood that the embodiments of the present invention can be applied to any other types of wireless communication systems, where both the first station and the second station can be static or mobile.

For a better understanding of the invention, and how it can be practiced, please refer now to the example in the accompanying drawing, in which:

Figure 1 shows an overview of a base station (BTS) and its associated cell sectors;

Figure 2 shows a simplified representation of a base station antenna array;

Figure 3 shows the fixed beam pattern provided by the antenna array of Figure 2;

Figure 4 shows an overview of the components of the beam selection section of the digital signal processor;

Figure 5 shows 4 different weighting modes;

Fig. 6 shows the general diagram of the digital signal processor of Fig. 2;

Figure 7 illustrates the channel impulse responses for 4 of the 8 channels; and

Figures 8a to 8c respectively illustrate the pattern selection data stored in the memory 1, the weighting pattern stored by the spatio-temporal weighting pattern unit, and the data calculated by this unit.

Referring first to Figure 1, there is shown three cell sectors 2 defined in a cell 3 of a cellular mobile telephone network. The 3 cell sectors 2 are served by different base stations (BTS) 4 . Three different base stations 4 are configured at the same location. Each BTS 4 has different transceivers, which send and receive signals from/to one of the 3 cell sectors 2 respectively. In this way, each cell sector 2 is equipped with a dedicated base station. In this way, the BTS 4 is able to communicate with mobile stations (MS), such as mobile phones, located in each cell sector 2.

This embodiment is described for a GSM (Global System for Mobile Communications) network. In the GSM system, the frequency division/time division multiple access F/TDMA system is used. Data is transmitted between BTS 4 and MS in the form of subframes. A data subframe includes a training sequence, which is a known data sequence. The purpose of the training sequence will be described later. Each subframe of data is sent in a given time slot of a frequency band within a given frequency band. The use of directional antenna arrays enables space division multiple access as well. Therefore, in an embodiment of the present invention, each data subframe is sent in a given time slot, within a given frequency band, and in a given direction. An associated channel may be determined for a given subframe of data sent in a given direction in a given time slot on a given frequency. As will be discussed in detail later, in some embodiments of the present invention, the same data subframe may be sent in two different directions in the same frequency band and in the same time slot.

Figure 2 shows an overview of an antenna array 6 of a BTS 4 acting as a transceiver. It should be understood that the array 6 shown in FIG. 2 serves only one of the three cell sectors shown in FIG. 1 . The other two antenna arrays 6 serve the other two cell sectors 2 . The antenna array 6 has 8 antenna elements a _{1 .} . . a ₈ . The spacing between each antenna element a ₁ ...a ₈ is half a wavelength, and these elements a ₁ ...a ₈ are arranged in a horizontal straight line. Each antenna element _a1 ... _a8 transmits and receives signals and may have any suitable structure. Each antenna element a ₁ ...a ₈ may be a dipole antenna, a patch antenna or any other suitable antenna. The eight antenna elements a _{1 .} . . a ₈ jointly form the phased array antenna 6 .

As is well known, to each antenna element _a1 ... _a8 of the phased array antenna 6 is transmitted the same signal that needs to be sent to the mobile station MS. However, the phases of the signals transmitted to the individual antenna elements a ₁ ...a ₈ are shifted. The difference in the phase relationship between the signals delivered to the individual antenna elements a ₁ ... a ₈ produces a directional radiation pattern. Thus, signals from the BTS 4 may only be sent in the specific direction of the cell sector 2 associated with the array 6 . The directional radiation pattern achieved by the array 6 is the result of constructive and destructive interference between signals transmitted by each antenna element a ₁ . . . a ₈ which are phase shifted from each other. In this regard, reference is made to FIG. 3, which illustrates the directional radiation pattern achieved by the antenna array 6. As shown in FIG. The antenna array 6 can be controlled to provide the beams b ₁ . . . b ₈ in any of the eight directions shown in FIG. 3 . For example, the antenna array 6 can be controlled to transmit signals to the mobile station MS only in the direction of beam _b5 , or only in the direction of beam _b6 . As will be discussed in further detail below, it is also possible to control the antenna array 6 to transmit signals in more than one beam direction simultaneously. For example, it is possible to transmit in both directions specified by beam _b5 and beam _b6 . FIG. 3 is only a schematic representation of the eight possible beam directions that can be achieved by the antenna array 6 . In practice, however, there will be an overlap between adjacent beams to ensure that all cell sectors 2 are served by the antenna array 6 .

The relative phase of the signals provided by each antenna element _a1 ... _a8 is controlled by the Butler matrix circuit 8 so that the signals are transmitted in the desired beam direction. Thus, the Butler matrix circuit 8 provides a phase shifting function. The Butler matrix circuit 8 has 8 inputs 10a-h from the BTS 4 and 8 outputs, each to an antenna element a ₁ ... a ₈ . The signal received by each input 10a-h includes the data bursts to be transmitted. Each of the eight inputs 10a-h represents the direction of the beam from which a given subframe of data is transmitted. For example, if the Butler matrix circuit 8 receives the signal at the first input 10a, the Butler matrix circuit 8 transmits the signal provided on the input 10a to each antenna element a ₁ ...a ₈ with the desired phase difference, resulting in the generation of beam b ₁ , so that the data subframe is sent in the direction of beam b ₁ . Likewise, a signal provided on input 10b results in the generation of a beam in the direction of beam _b2 , and so on.

As previously discussed, the antenna elements _a1 ... _a8 of the antenna array 6 receive signals from the MS and transmit signals to the MS. The signal transmitted by the MS is generally received by one of the 8 antenna elements a ₁ . . . a ₈ . However, there is a phase difference between each signal received by the antenna elements a ₁ ...a ₈ . Thus, the Butler matrix circuit 8 is able to determine, from the relative phases of the signals received by the individual antenna elements a ₁ . . . a ₈ , the beam direction in which the signal was received. Therefore, the Butler matrix circuit 8 has 8 inputs, one for each antenna element a _{1 .} . . a ₈ , for transmitting the signal received by each antenna element. The Butler matrix circuit 8 also has eight outputs 14a-h. Each output 14a to 14h corresponds to a particular beam direction in which a given subframe of data is received. For example, if antenna array 6 receives a signal from a MS from the direction of beam _b1 , Butler matrix circuit 8 will output the received signal on output 14a. A signal received from the direction of beam _b2 will cause the Butler matrix circuit 8 to output the received signal on output 14b and so on. In summary, the Butler matrix circuit 8 will receive 8 versions of the same signal through the antenna elements a ₁ . . . a ₈ , which are phase shifted with respect to each other. The Butler matrix circuit 8 determines the direction in which the received signal was received based on the relative phase shift and outputs a signal at a given output 14a-h based on the direction in which the received signal was received.

It should be understood that in some circumstances a single signal or data subframe from the MS may appear on more than one beam direction. Butler matrix circuit 8 will provide a signal on each output 14a-h corresponding to each beam direction in which a given signal or data subframe occurs. Thus, the same burst of data may be provided on more than one output 14a-h of the Butler matrix circuit 8 . However, the signals on the respective outputs 14a-h may be timed differently from each other.

Each output 14a-h of the Butler matrix circuit 8 is connected to the input of a respective amplifier 16 which amplifies the received signal. An amplifier 16 is provided for each output 14a-h of the Butler matrix circuit 8 . The amplified signal is then processed by processors 18 which operate on the amplified signal to reduce the frequency of the received signal to baseband frequency so that the signal can be processed by the BTS 4. To achieve this, processor 18 removes the carrier frequency component from the input signal. Likewise, a processor 18 is provided for each output 14a-h of the Butler matrix circuit 8 . Then, an analog-to-digital (A/D) converter 20 converts the received signal in analog form into a digital signal. Eight A/D converters 20 are provided, one for each output 14a-h of the Butler matrix circuit 8 . The digital signals are then input to a digital signal processor 21 via respective inputs 19a-h for further processing.

The digital signal processor 21 also has eight outputs 22a-h, each outputting a digital signal representing a signal to be sent to a given MS. Selected outputs 22a-h indicate the beam direction of the transmitted signal. The digital signal is converted into an analog signal by a digital-to-analog (D/A) converter 23 . A digital-to-analog converter 23 is provided for each output 22a-h of the digital signal processor 21 . The analog signal is then processed by the processor 24, which is a modulator, which modulates the analog signal to be transmitted to a carrier frequency. Prior to processing the analog signal by processor 24, the signal is at baseband frequency. The resulting signals are then amplified by the amplifier 26 and passed to the respective inputs 14a-h of the Butler matrix circuit 8 . A processor 24 and amplifier 26 are provided for each output 22a-h of the digital signal processor 21 .

Referring now to FIG. 4, there is shown a schematic diagram of beam selection means 101 forming part of the digital signal processor 21 for selecting the beam direction in which the BTS 4 transmits data subframes to a given mobile station. The beam selection device 101 comprises a beam preselection component 100 . The beam preselection component 100 determines for each data subframe received from a given mobile station the or each direction in which the given data subframe is received. The beam preselection component 101 may also determine the intensity of the data subframes received in each of the different directions. Beam pre-selection component 101 may select only a limited number of beam directions as directions for receiving a given sub-frame of data. For example, beam preselection component 100 may select a single beam direction for each subframe of data. A single beam direction corresponds to the direction in which the strongest signal is received, or to the direction in which the delay in receiving the subframe of data is minimal. It is of course possible to select two or more beam directions as the directions in which to receive a given sub-frame of data. Any suitable criteria may be used in beam selection. In some embodiments of the invention, the beam selection criteria may be varied, eg to take into account changes in the wireless environment or the distance between the mobile station and the base station.

The beam direction determined by beam preselection unit 100 is output to memory 102 via output 103 . Furthermore, the intensity information of the data subframes received in the respective beam directions can also be output to the memory 102 . Memory 102 may store strength information received by the BTS from the desired MS for n previous subframes. Memory 102 may take the form of FIFO registers. Thus, if the beam direction and signal strength information for the next data subframe received from the desired MS is determined, this information is stored in memory 102, where the previous n+1th data subframe is eliminated. information. In other words, the oldest information in the registers is evicted to make way for the newest information. Referring to Figure 8a, an example of information stored in memory 102 is illustrated. For each beam, a 1 or a 0 is stored for the given data subframe. A 1 indicates that the beam preselection component 100 has selected the beam for the given data subframe. In the example shown in Figure 8a, the beam preselection component 100 has selected 3 beams for a given data subframe. For example, beams 3, 4 and 5 are selected for the ith data subframe, while beams 1, 2, 6, 7 and 8 are not selected.

The memory 102 has a plurality of outputs 104 which are connected to a spatio-temporal weighted pattern component 106 . In one embodiment of the invention, an output 104 is provided. Thus, an output 104 provides information stored in memory 102 for an antenna element. Spatio-temporal weighting pattern component 106 applies a weighting pattern to the stored K data burst information previously received from a given MS.

Referring to Figure 8b, an example of a weighting pattern stored by component 106 is illustrated. As can be seen from this figure, the weighting pattern can be viewed as 8 different patterns, one for each beam. Furthermore, it can be seen that the weighting patterns are different for different beam directions. For example, the weighting pattern of the first beam is (0,0,0,0) and that of the second beam is (1,1,1,1). In this example, the weighting pattern is only applied to the first 4 data subframes. Therefore, K=4. In this example, although the weighting pattern is different for each beam, the same weights are applied to each of the 4 previously received data subframes.

Referring now to FIG. 5, an example of four weighting patterns is shown. However, it should be understood that these 4 modes are for illustration only. Virtually any suitable weighting scheme or algorithm may be used.

In the first weighting scheme shown in Figure 5a, all K previous beam selections of previously received data subframes are equally important in determining the beam direction for the BTS 4 to send the next data subframe to the desired MS. In the second weighting mode shown in Figure 5b, the closer the previously received data subframe is to the currently received data subframe, the more it affects the selection of the beam direction for sending the next data subframe to the desired MS bigger. In other words, the beam selection made by the Kth preceding subframe has less influence on the direction in which the BTS decides to send the next subframe to the desired MS than the most recently received data subframe. The weighting pattern shown in Figure 5b suggests a linear growth pattern.

The weighting pattern shown in Fig. 5c is similar to that shown in Fig. 5b, and the previous K-th beam selection made by the BTS has less influence on determining the direction in which the BTS sends the next data subframe to the desired MS. Thus, the exponentially growing weighting scheme shown in Fig. 5c only takes into account the information determined by the beam preselection component 100 only for the most recently received data subframe.

In the limited case shown in Figure 5d, the selection of the beam direction in which the BTS transmits a data burst to the desired MS is based solely on the most recent data burst received by the BTS from the desired MS. This selection occurs when the wireless environment changes so rapidly that previous beam selections have very little effect on the current beam selection.

The spatio-temporal weighting pattern component 106 receives input from the weighting pattern set component 108 . This part stores, for example, 4 modes as shown in FIG. 5 . The weighting pattern set component also stores any other suitable weighting patterns or weighting pattern algorithms. In the latter case, the weighting pattern need not follow a specific function. One or more weighting pattern algorithms may be provided to replace the stored weighting patterns. Any suitable number of weighting patterns may be stored. For example, in one embodiment of the invention, only a single weighting scheme or the algorithm specifying the weighting scheme is stored.

The weighting mode group component 108 can also select the most suitable weighting mode for the current wireless environment, and output the weighting mode to the spatiotemporal weighting mode component 106 . If the wireless environment is static or changes slowly, the mode shown in Figure 5a can be selected. In this way, all K previous subframes have equal importance when determining the direction of the BTS to send the next data subframe. In this way, considering the average influence of the K previous data subframes, the influence of data subframes with eg a lot of interference will be greatly reduced.

In rapidly changing wireless environments, the patterns of Figure 5b or Figure 5c may be used. Rapidly changing wireless environments can occur when a mobile station passes rapidly through an urban environment. In this environment, newer data bursts received are more realistic than data bursts received earlier. Therefore, the information of a newer received data burst should have a greater influence on the direction or each direction in which the BTS 4 sends the next data burst.

In order to determine which weighting pattern is most appropriate, the radio environment information is input into the weighting pattern set component 108 . In an embodiment of the present invention, the information of the wireless environment can be independently determined. However, in a preferred embodiment of the present invention, the received data subframes are used to provide information about the radio environment. For example, weighting pattern group component 108 can take into account the direction in which the BTS received L previous subframes from the desired MS. L may be the same as K or different. If the direction in which consecutive bursts of data are received varies somewhat, it can be assumed that the radio environment is static, or only changes slowly. The weighting scheme shown in Figure 5a can be selected. On the contrary, if it is determined that the previous L data subframes are received from quite a number of different directions, it can be assumed that the mobile station is in a rapidly changing wireless environment, and the mode shown in Figure 5b or c can be selected. The wireless environment information can also be stored in advance. In another aspect, statistical information stored in memory 102 may be used to determine the type of wireless environment.

The spatiotemporal weighting pattern component 106 applies the selected weighting pattern to the information provided by the beam preselection component 101 for the K previous data subframes received by the BTS from the desired MS. The spatio-temporal weighted mode component 106 calculates a beam score for each of the 8 possible beam directions using the following formula:

s(i,k)=W _ST (i,k) ^T b _ST (i,k) k=1,2....8

s(i,k) represents the score of the kth beam in deciding the beam direction chosen by the BTS 4 for transmitting the i'th (or next) subframe to the desired mobile station.

W _ST (i, k) is a Kx1 vector containing the K spatiotemporal real weights of the K th beam in the i th subframe. It takes into account the weighting pattern applied to the receive beam selection information of the K preceding subframes.

b _ST (i,k) is a Kx1 binary vector containing the statistics of the previous K selections of the Kth beam. If b _sT (i,k)=1, where r=(K−1), . . . , i, this means that the beam preselection part 100 selects the kth beam during the rth subframe. If _bST (i,k) = 0, this means that the kth beam is not selected during the rth subframe. Thus, the score s(i, k) is 0 if the beam preselection component 100 did not select the kth beam for the K previous subframes of data received from the mobile station.

Referring now to FIG. 8c, there is shown the score vector computed by spatiotemporal weighting pattern component 106 for each of the eight beams using the selected weighting pattern. For example, using the weighting mode shown in Figure 8b, the vector score of the third beam is as follows: $\mathrm{the\; s}(i,3)=W_{\mathrm{ST}}{(i,3)}^T{b}_{\mathrm{st}}(i,3)=\left[\mathrm{2,2,2,2}\right]\left[\begin{array}{c}1\\ 1\\ 1\\ 1\end{array}\right]=2.1+2.1+2.1+2.1=8$

The spatio-temporal weighted pattern block has 8 outputs 110 which are inputs to the final beam selection block 112 . Each of the outputs 110 corresponds to one of eight possible beam directions. In one embodiment of the invention, the score vector computed by the spatio-temporal weighting pattern component 106 is input to the final beam selection component 112 via eight outputs 110 . The final beam selection component 112 then selects these beam directions for the BTS to transmit the next data subframe to the given MS. If only one beam direction is used, the beam direction with the highest score is chosen from the 8 possible calculated scores as the transmitted beam direction. In the example shown in Figure 8c, if only one beam is selected, the third beam with the highest score is selected. If 3 beams are selected, beams 3, 4 and 5 are selected. However, it should be understood that any suitable number of beam directions may be selected. However, the number of selected beam directions is generally much smaller than the number of possible beam directions that can be selected.

In a development of the invention, the number of beams selected can vary. For example, a threshold can be set such that only beams with a score of eg 6 or higher can be selected. This has the advantage that more beams are selected when the mobile station is relatively close to the base station, and fewer beams are selected when the mobile station is relatively far away from the base station. This is preferred because when the mobile station is located at a considerable distance from the BTS, farther than a critical distance, only a relatively small number of beam directions can be selected, each beam having a relatively high energy. However, if the distance between the BTS and the mobile station is less than the critical distance, it is preferable to select a relatively large number of beam directions, each beam having a relatively low energy. The critical distance depends on the circumstances of each cell and may be about 0.5 to 1 km. The weighting pattern or weighting algorithm can make the number of selected beams variable, with more beams being selected when the mobile station is relatively close to the base station.

Any suitable method may be used to determine whether the distance between the MS and the BTS is greater than a critical distance. In one embodiment, the resulting channel impulse responses for each possible direction are compared. If the majority of the received energy is distributed in 3 or fewer beam directions, it can be assumed that the distance between the BTS and the MS is greater than the critical distance. Optionally, if most of the received energy comes from 4 or more beam directions, it is assumed that the distance between the MS and the BTS is less than the critical distance.

The comparison component may also use the timing advance information to determine whether the distance between the MS and the BTS is greater or less than a threshold distance. This method is preferred in some embodiments of the invention because it gives more accurate results than previously described.

Once the distance between the mobile station and the BTS is determined, this information can be used to select an appropriate weighting pattern or algorithm. For example, if the BTS and the mobile station are in close proximity, the weighting pattern shown in Figure 5c or d is selected.

In another development of the invention, individual weighting patterns can be stored in the weighting pattern set, a fixed pattern being used in all cases.

It should be emphasized that any suitable weighting scheme may be used, not just the ones shown in Figures 5a to d. The weighting scheme shown can be replaced by a suitable algorithm to calculate the appropriate weights. This algorithm takes into account different factors, such as the wireless environment and the distance between the mobile station and the base station, when calculating the appropriate weights. The weighting pattern can vary for different beam directions, or be the same for all beam directions.

In another modification of the invention, not only the beam direction and signal strength information is determined or stored by the beam preselection means and memory means 100 and 102, but the weighting pattern means applies a weighting pattern to the beam intensity information to The strength of data bursts sent in the or each beam direction is determined. The power of the data subframes transmitted in the selected beam direction is then determined based on the data subframe strengths of the K previously received data subframes. When more than one beam direction is selected, the determined intensities can be different in different beam directions.

It should be appreciated that the weighting pattern applied by the spatio-temporal weighting pattern component may vary from subframe to subframe.

Referring now to Figure 6, the digital signal processor 21 is schematically illustrated in more detail. It should be understood that the different components shown in Figure 6 do not necessarily correspond to the different elements of an actual digital signal processor 21 embodying the invention. Specifically, different components shown in FIG. 6 correspond to different functions performed by the digital signal processor 21 . In one embodiment of the present invention, the digital signal processor 21 is at least partially implemented in an integrated circuit, and multiple functions can be performed by the same component.

Each signal received by the digital signal processor 21 on a respective input 19a-h is input to a respective channel impulse response (CIR) estimator block 30 . The CIR estimator block 30 includes memory capacity to store the estimated channel impulse response. The CIR estimator block also includes memory capacity for temporarily storing received signals. The channel impulse response estimator block 30 estimates the channel impulse response of the channel of each input 19a-h. As discussed above, the associated channel may be determined for a given data subframe transmitted in the direction of the beam receiving the signal in the allocated time slot on the selected frequency band. The beam direction of the received signal is determined by the Butler matrix circuit 8, so that the signal received at the input 19a of the digital signal processor is mainly the signal received from the direction of the beam _b1 . It should be understood that a signal received on a given input may also include side lobes of signals received on eg adjacent inputs.

Each data burst sent from the mobile station MS to the base station BTS 4 comprises a training sequence TS. However, the training sequence TS _RX received by the BTS 4 is affected by noise and multipath effects which cause interference between adjacent bits of the training sequence. The latter interference is called intersymbol interference. TS _RX is also interfered by other mobile stations, for example, mobile stations located in other cells or cell sectors using the same frequency will cause co-channel interference. It should be understood that a given signal from an MS may follow more than one path to reach the BTS, with antenna array 6 detecting more than one version of a given signal from a given direction. The CIR estimator part 30 cross-correlates the training sequence TS _RX received from the input 19 a with the reference training sequence TS _REF stored in the data memory 32 . The reference training sequence TS _REF is the same as the training sequence originally sent by the mobile station. In practice, the received training sequence TS _RX is a signal modulated onto a carrier frequency, while the reference training sequence TS _REF is stored in the data memory 32 in the form of a bit sequence. Therefore, a similar modulation is performed on the stored reference training sequence prior to cross-correlation. In other words, the distorted training sequence received by the BTS 4 is associated with the undistorted version of the training sequence. In an optional embodiment of the invention, the received training sequence is demodulated before being associated with the reference training sequence. In this case, the form of the reference training sequence is also the same as the received training sequence. In other words, the reference training sequence is not modulated.

The lengths of both the reference training sequence TS _REF and the received training sequence TS _RX are L, equal to L data bits, for example, 26 bits. The exact position of the received training sequence TS _RX in the allocated time slot may be indeterminate. This is because the distance between the mobile station MS and the BTS 4 affects the position of the data subframes transmitted by the MS in the allocated time slots. For example, if the mobile station MS is farther away from the BTS 4, the training sequence may occur later in the allocated time slot than if the mobile station MS is close to the BTS 4.

In order to take into account the uncertainty of the position of the received training sequence TS _RX in the allocated time slot, the received training sequence TS _RX is associated n times with the reference training sequence TS _REF . n is generally eg 7 or 9 times. n is preferably an odd number. N associations are generally on either side of the largest association obtained. Between each successive association, the received training sequence TS _RX is shifted by one position relative to the reference training sequence TS _REF . Each position is equivalent to a bit in the training sequence, representing a time delay segment. Each association of the received training sequence TS _RX with the reference training sequence TS _REF results in a branch representing the channel impulse response for that association. N different associations result in a sequence of branches with n values.

Referring now to FIG. 7, channel impulse responses for 4 of 8 possible channels corresponding to 8 spatial directions are shown. In other words, Figure 5 shows the channel impulse responses of 4 channels corresponding to a given data subframe received from a mobile station in 4 of 8 beam directions, the data subframe being located in a given frequency band, and at a given time slot. The x-axis of each graph is a latency measurement, while the y-axis is a relative power measurement. Each row (or branch) labeled in the figure represents the received multipath signal for a given associated delay. Each graph has n rows or branches, one branch corresponds to one association.

From the estimated channel impulse response, the position of the training sequence in the allocated time slot can be determined. The maximum branch value can be obtained when the received training sequence TS _RX and the reference training sequence TS _REF are optimally correlated.

The CIR estimator block 30 also determines for each channel the 5 (or any other suitable number) consecutive branches that give the maximum energy. Calculate the maximum energy for a given channel as follows: $\mathrm{E.}=\underset{j=1}{\overset{5}{\mathrm{\Σ}}}{\left(h_j\right)}^2---\left(I\right)$

Where h represents the branch amplitude obtained from the cross-correlation between the reference training sequence TS _REF and the received training sequence TS _RX . The CIR estimator block 30 estimates the maximum energy for a given channel using a sliding window technique. In other words, the CIR estimator component 30 considers each of the 5 neighboring values and calculates the energy from these 5 values. The 5 adjacent values giving the maximum energy are chosen as channel impulse response representatives.

Energy can be seen as a measure of the energy of the desired signal received by the BTS 4 from a given MS in a given direction. This process is done for each of the 8 channels representing 8 different directions for receiving the same subframe of data. The path along which the received signal with the greatest energy passes has the least attenuation of the signal.

Analysis block 34 stores the maximum energy calculated by CIR estimator block 30 for each channel it selects as five adjacent values of the channel impulse response. Analysis component 34 may also analyze the channel impulse response determined by CIR component 30 to determine the minimum delay. The time delay is a measure of the position of the received training sequence TS _RX in the allocated time slot, and is therefore a relative measure of the distance traveled by the signal between the mobile station and BTS4. The signal of the channel with the smallest delay travels the shortest distance. In certain cases, this shortest distance can be expressed as a line-of-sight path between the mobile station MS and the BTS 4 .

The analysis component 34 determines the starting position of a window defining the 5 values providing the maximum energy. Then, the time delay is determined based on the time between the reference point and the window start point. This reference point may be a common moment at which all received training sequences in each branch start to correlate, which corresponds to the earliest window bound of all branches or similar common point. In order to accurately compare the different time delays of the different channels, a common timing scale is used, which depends on the synchronization signal provided by the BTS 4 to control the TDMA mode of operation. In other words, the position of the received training sequence TS _RX in the allocated time slot is a measure of delay. It should be understood that in known GSM systems, in order to provide timing advance information, it is necessary to calculate the time delay for a given channel. The timing advance information is used to ensure that the signal sent by the mobile station to the BTS is in its assigned time slot. The timing advance information may be determined according to the calculated relative delay and the current timing advance information. If the mobile station MS is far away from the base station, the time when the BTS notifies the MS to send its data subframe is earlier than when the mobile station MS is close to the base station.

The results of the analysis performed by each of the analysis parts 34 are input to the beam selection part 101, which has been described with respect to FIG. 4 .

The beam preselection part 100 of the beam selection part 101 performs beam preselection using the estimated channel impulse response. This can be achieved in many different ways. If, for example, the beam preselection component 100 determines a single beam direction for a given subframe, the beam preselection component 100 can determine which channel, and thus which beam direction, has the data for a given subframe in a given time slot in a given frequency band maximum energy required. This means that the beam direction for a given subframe of data that received the strongest version can be determined. This direction can be used to select the beam direction. Beam preselection component 100 may also determine which channel has the smallest delay. In other words, the channel, and thus the beam direction, with the data subframe along the shortest path can be determined and used as the selected beam direction for a given data subframe.

It should be understood that in embodiments of the present invention, beam preselection component 100 may select more than one beam direction for a given data subframe. For example, the two directions in which the strongest versions of a given data signal are received may be selected as given beam directions. Likewise, the two beam directions that provide the signal with the smallest time delay can be selected as beam directions. Of course, the beam preselection component 100 can determine the direction receiving the strongest signal and the direction having the smallest time delay, and take these two directions as the selected directions.

The beam preselection component 100 also receives the associated energy calculated by each analysis component 34 for a given selected beam direction.

The beam selection component 100 provides an output to the generation component 38 which indicates which beam direction to use to transmit the signal from the BTS 4 to the MS, and the appropriate power value to use for each of these beam directions.

The generating part 38 is responsible for generating the signal output from the digital signal processor 21 . The input 40 of the generating means 38 represents the speech and/or information to be sent to the mobile station MS. The generating part 38 is responsible for encoding the speech and/or information to be sent to the mobile station MS, and includes the training sequence and the synchronization sequence in the signal. The generation component 38 is also responsible for generating the modulated signal. The generating means 38 provide signals at respective outputs 22a-h of the digital signal processor 21 based on the generated modulated signal and determined beam direction. The generation component 38 also provides an output 50 for controlling the amplification provided by the amplifier 24 to ensure that the signal transmitted in the principal one or more beam directions has the desired power level.

The output of the channel impulse response section 30 is also used to equalize and match the signal received from the mobile station MS. Specifically, the matched filter (MF) and equalization unit 42 can eliminate or mitigate the impact of intersymbol interference caused by multipath propagation on the received signal. It should be appreciated that the matched filter (MF) and equalization block 42 has an input (not shown) for receiving the signal received from the MS. The output of each component 42 is received by a recovery component 44, which is responsible for recovering the speech and/or information sent by the MS. The steps performed by the recovery means include demodulating and decoding the signal. The recovered speech or information is output on output 48 .

It should be understood that although the above-described embodiments are implemented in a GSM cellular communication network, the present invention can also be used in other digital cellular communication networks, as well as in analog cellular networks. The above embodiment uses a directional array with 8 elements. The array can of course have any number of elements. The phased array can also be replaced by discrete directional antennas, each emitting a beam in a given direction. Where a Butler matrix circuit is desired, the Butler matrix circuit may be replaced by any other suitable phase shifting circuit. Butler matrix circuits are analog beamformers. Of course a digital beamformer DBF or any other suitable type of analog beamformer could be used. Even if only 8 elements are provided, the array can be controlled to generate more than 8 beams based on the signals delivered to these elements.

Multiple phased arrays may also be provided. Phased arrays can provide different numbers of beams. If a wider angular range is required, use a lower element count array, and if a narrower beam is required, use a higher element count array.

It will be appreciated that the above described embodiment provides 8 outputs through the Butler matrix circuit. It should be understood that in practice each output on the Butler matrix can simultaneously output multiple different channels. These channels can have different frequency bands. Channels of different time slots are also provided on each output. Although a single amplifier, processor, analog-to-digital converter and digital-to-analog converter have been shown, in practice these could all be provided in a single element with multiple inputs and outputs.

It should be understood that the embodiments of the present invention are not only applicable to cellular communication networks. For example, embodiments of the present invention may be used in any environment where directional wireless communication is required. For example, this technique can be used in PMR (Private Radio Network) or similar networks.

Claims (15)

1. A method for directional wireless communication between a first station and a second station, said method comprising the following steps: said first station receives a plurality of consecutive signals from said second station, each said signal being receivable in at least one of a plurality of different directions; determining a value for at least one parameter for each of the plurality of serial signals received by the first station from the second station in succession; and selecting a value of at least one parameter for a signal to be transmitted from said first station to said second station, the selection of said value of at least one parameter of a signal to be transmitted by the first station being dependent on said plurality of serial signals said determined value of said at least one parameter, wherein said selecting step comprises applying a weighting pattern to said plurality of serial signals. 2. A method according to claim 1, wherein said step of determining at least one parameter value comprises determining, for each of a plurality of serial signals, the or each direction in which said serial signal is received, said selecting step comprising At least one direction for sending a signal from the first station to the second station is selected, the selection of the at least one direction being dependent on the determined direction of the plurality of serial signals. 3. A method according to claim 1 or 2, wherein said step of determining at least one parameter value comprises determining the strength of each of said plurality of serial signals, said selecting step comprising selecting the signal to be sent to said second A strength of a signal of a station, said signal strength being selected depending on said strength determined for said plurality of serial signals. 4. A method according to claim 1 or 2, wherein said weighting pattern is a uniform weighting pattern whereby each of the plurality of serial signals is given the same weight. 5. The method according to claim 1 or 2, wherein said weighting pattern is to give a weight of a signal received more recently in said plurality of serial signals than a weight of a signal received next in said plurality of serial signals value. 6. The method of claim 5, wherein said weighting pattern is an exponentially or linearly increasing weighting pattern. 7. A method according to claim 1 or 2, comprising the step of determining the wireless environment. 8. The method according to claim 7, wherein said selecting means selects a weighting pattern from a plurality of weighting patterns according to a wireless environment. 9. A method according to claim 1 or 2, comprising the step of the first station defining a plurality of beam directions for transmitting a radiation beam, wherein each of said beam directions is individually selectable. 10. A method according to claim 1 or 2, wherein said first station is a base station of a cellular network. 11. A method according to claim 1 or 2, wherein said second station is a mobile station. 12. A first station performing directional wireless communication with a second station, the first station comprising: receiver means for receiving a plurality of consecutive signals from said second station, each said signal being receivable in at least one of a plurality of different directions; means for determining the value of at least one parameter for each of the plurality of serial signals received by the first station from the second station; a transmitter device for transmitting a signal from the first station to the second station; and control means for controlling said transmitter means, said control means selecting the value of at least one parameter for a signal to be transmitted from said transmitter means, the selection of said at least one parameter value being dependent on said plurality of serial signals The determined value of said at least one parameter, wherein said control means applies a weighting pattern to said plurality of serial signals. 13. A first station according to claim 12, wherein said determining means determines the direction or each direction of each signal in a plurality of serial signals, the control means selects at least one direction in which the transmitter means transmits the signal, said at least one The choice of direction depends on the determined direction of the plurality of serial signals. 14. The first station according to claim 12 or 13, wherein said determination means determines the strength of each signal in said plurality of serial signals, and the control means selects the strength of the signal transmitted by the transmitter means, the selection of the signal strength being determined by based on the determined strengths for the plurality of serial signals. 15. A first station according to claim 12 or 13, wherein storage means are provided for storing said determined parameters for each of the plurality of serial signals.

Images