JP2025122325A - wireless communication system - Google Patents

Abstract

A wireless communication system is provided that performs tap coefficient acquisition processing of a decision feedback adaptive equalizer at high speed.
[Solution] A data frame DF_B is transmitted from station B at a timing delayed by an offset number M of symbols, which is the same number N of symbols as the synchronization word SW_A, relative to a data frame DF_A transmitted from station A, and the following processes are executed for a combined data frame DF_A+DF_B formed by combining data frames DF_A and DF_B: a first process for pulling in tap coefficients; a second process for performing an equalization process based on an equalizer input and tap coefficients to obtain an approximation value of the symbol of block data contained in the combined block data as an equalizer output; and a third process for calculating the error between the hard decision value of the equalizer output and the equalizer output, and updating the tap coefficient to the latest value using the calculated error, tap coefficients, and new equalizer input; and the second and third processes are repeated until all approximations of the symbol of the block data are output.
[Selected Figure] Figure 7

Description

The present invention relates to a wireless communication system that uses same-wave multi-station transmission technology.

In wireless communication systems that cover a wide area with narrowband signals (e.g., metropolitan disaster prevention radio systems, prefectural disaster prevention radio systems, etc.), it is known that frequency utilization efficiency can be improved by using same-wave multi-station transmission technology, in which multiple transmitting base stations transmit the same signal at the same frequency to mobile receiving stations. For example, the wireless communication system shown in Figure 12 (A) is composed of multiple transmitting base stations BS1, BS2, etc., arranged so that their communication areas partially overlap, and multiple mobile receiving stations MS2-MS4, such as a mobile receiving station MS1 mounted on a moving object such as an automobile, and carried by pedestrians. The multiple transmitting base stations BS1 and BS2 are arranged so that the communication areas reached by their transmission waves partially overlap each other, and they broadcast transmission waves Tx1 and Tx2 containing the same data using the same frequency. This allows mobile receiving stations MS1-MS4 to receive transmission waves Tx1 or Tx2 at the same carrier frequency regardless of their location within the communication areas of transmitting base stations BS1 and BS2.

In such a wireless communication system, as shown in Figure 12(B), in areas where the transmitted wave Tx2 is out of phase with the transmitted wave Tx1 and the signal level ratio of Tx1/Tx2 is 0 dB, identical wave interference (hereinafter referred to as beat interference) occurs in the composite wave Tx1+Tx2 of the transmitted waves Tx1 and Tx2. Such beat interference occurs in multiple stripes, as depicted by vertical lines in Figures 12(A) and 12(B). Mobile receiving stations MS2 and MS3, located outside the beat interference occurrence point, can communicate with transmitting base station BS1 or BS2. In contrast, mobile receiving station MS4, located at the beat interference occurrence point, cannot communicate with transmitting base station BS1 or BS2. Furthermore, as the moving speed of mobile receiving station MS1 increases, the effects of Doppler shift become more pronounced, and temporal fluctuations in the propagation path cause Rayleigh fading and Ricean fading. In addition, since each incoming wave from multiple transmitting base stations experiences independent Rayleigh fading or Rician fading, demodulation performance deteriorates due to the independent propagation path fluctuations of each incoming wave.

In order to reduce the effects of beat interference, a known technique is to use different mappings for modulating communication data between adjacent base stations, thereby constantly shifting the location where beat stripes occur and suppressing beat interference (see, for example, Patent Document 1). Furthermore, known techniques for compensating for fading include the use of decision feedback equalization (DFE) and maximum likelihood sequence estimation (MLSE) (see, for example, Patent Document 2).

JP 2018-166293 A JP 2012-070348 A

However, with the technology described in Patent Document 1, the communication areas of transmitting base stations using the same mapping must not overlap, which places restrictions on the placement of transmitting base stations, etc. Furthermore, the MLSE described in Patent Document 2 achieves a low bit error rate, but the amount of calculation required increases dramatically depending on the number of states. On the other hand, the DFE described in Patent Document 2 requires less calculation than MLSE, but in same-wave multi-station transmission, the tap coefficients of the forward filter and feedback filter are trained using a synchronization word (SW) multiplexed with transmission waves from multiple base stations. When the receiving station moves at high speed, the propagation paths between each base station and the receiving station become independent Rayleigh fading environments, making it impossible to adequately perform tap acquisition using the multiplexed synchronization word.

The present invention aims to provide a wireless communication system that can quickly perform tap coefficient acquisition processing in a decision feedback adaptive equalizer.

In order to solve the above problem, the invention described in claim 1 is a wireless communication system comprising a plurality of transmitting base stations equipped with wireless transmitting devices that transmit the same modulated data frames at the same frequency, and a mobile receiving station equipped with wireless receiving devices that receive and demodulate the data frames, wherein the plurality of wireless transmitting devices respectively transmit the data frames, each comprising known header data, a plurality of block data having the same number of symbols as the header data, and a no-data section arranged between the header data and the first block data and having the same number of symbols as the header data, with a time difference corresponding to the number of symbols of the header data as an offset symbol number, and the wireless receiving device comprises receiving means for receiving a combined data frame formed by combining the plurality of data frames in space and outputting a received signal for each symbol, and a decision feedback adaptive equalizer to which the received signal is input, a first process for adjusting tap coefficients used in equalization processing based on received signals of the plurality of header data included in the combined data frame and a pre-stored reference signal; a second process for using received signals of combined block data constituting the combined data frame as equalizer inputs and performing equalization processing based on the equalizer inputs and current tap coefficients to obtain approximation values of the symbols of the block data included in the combined block data as equalizer outputs; and a third process for calculating an error between a hard decision value of the equalizer output and the equalizer output, and updating the tap coefficients to their latest values using the calculated error, the current tap coefficients, and a new equalizer input; and the second process and the third process are repeated until all approximation values of the symbols of the block data are output from the combined data frame.

According to the invention described in claim 1, when the same data frame is transmitted from multiple transmitting base stations, data frames having no-data intervals with the same number of symbols as the header data are transmitted with a time difference equivalent to the number of symbols in the header data, using an offset symbol number. This allows the header data used in the tap coefficient acquisition process in the decision feedback adaptive equalizer to be received independently without overlapping with other data (in other words, without being combined), eliminating the need for means to identify which of the multiple transmitting base stations the header data comes from (for example, encoding the transmission signal from each station), and enabling high-speed tap coefficient acquisition.

1 is a diagram showing a schematic configuration of a wireless communication system according to a first embodiment of the present invention; 1A shows the structure of a data frame transmitted from a base station, and FIG. 1B shows a composite data frame received by a mobile station. 2 is a functional block diagram showing a schematic configuration of a wireless communication device of the mobile station shown in FIG. 1. FIG. 4 is a block diagram showing a schematic configuration of the decision feedback adaptive equalizer shown in FIG. 3. FIG. 5 is a block diagram showing the configuration of a delay device shown in FIG. 4 . FIG. 10 is a block diagram showing a schematic configuration of a decision feedback adaptive equalizer when there are two base stations. FIG. 10 is an explanatory diagram showing a process of outputting approximate values of symbols of block data from a combined data frame by a decision feedback adaptive equalizer. FIG. 10 is an explanatory diagram showing a process of outputting an approximate value of a symbol of block data from a specific composite block data by a decision feedback adaptive equalizer. 10A shows the configuration of data frames transmitted from three base stations in the second embodiment, and FIG. 10B shows a composite data frame received by a mobile station. FIG. 10 is a block diagram showing a schematic configuration of a decision feedback adaptive equalizer when there are three base stations. 10 is an explanatory diagram showing a process of outputting an approximate value of a symbol of block data from a combined data frame of three stations by a decision feedback adaptive equalizer. FIG. FIG. 1 is a schematic diagram showing a state in which beat interference occurs in a conventional wireless communication system.

The present invention will be described below based on the illustrated embodiment. Note that the following describes the characteristic configuration of the present invention, and omits a description of the conventional mechanisms for wireless communication.

(Embodiment 1)
1 is a diagram showing a schematic configuration of a wireless communication system 1 according to an embodiment of the present invention. The wireless communication system 1 is used to communicate disaster prevention information relating to, for example, natural disasters, and includes a plurality of transmitting base stations, e.g., base station A and base station B (hereinafter also referred to as stations A and B), installed at arbitrary locations, and a mobile receiving station, e.g., mobile station D, which can move to arbitrary locations. Base station A and base station B each include a wireless communication device (wireless transmitting device) 2, and mobile station D includes a wireless communication device (wireless receiving device) 3. These wireless communication devices 2 and 3 are connected to each other by a wireless line 4 via antennas 2a and 3a.

In this embodiment, the wireless communication system 1 performs high-speed tap coefficient acquisition processing in a decision feedback adaptive equalizer when a mobile station D receives and demodulates the same data frame at the same frequency that has been modulated and transmitted from multiple base stations A and B.

In order to suppress the beat interference described above, the wireless communication devices 2 of base stations A and B according to this embodiment transmit data frames each including a known synchronization word (header data), multiple block data pieces each having the same number of symbols (data length) as the synchronization word, and a no-data section between the synchronization word and the first block data piece, each having the same number of symbols as the synchronization word, with a time difference equivalent to the number of symbols in the synchronization word, as the offset symbol number.

Figure 2 (A) shows an example of a data frame transmitted from base station A and base station B. Base station A's data frame DF_A is, for example, PSK modulated and comprises, from the beginning on the left, a synchronization word SW_A (SW: Sync Word), a no-data section ND_A, and multiple, for example, seven, blocks of data BD_A. The blocks of data BD_A are assigned block numbers A1, A2, A3, A4, A5, A6, and A7, respectively.

The synchronization word SW_A is known data comprising a predetermined number of symbols N (for example, 32 symbols in this embodiment) and indicates the beginning of the data frame DF_A. The no-data section ND_A is provided so that when mobile station D simultaneously receives data frame DF_A from station A and data frame DF_B from station B, the synchronization word SW_B of data frame DF_B can be received independently without overlapping with other data (in other words, without being combined). The number of symbols in the no-data section ND_A is the same as the number of symbols N in the synchronization word SW_A.

Modulated transmission data is stored in A1, A2, A3, A4, A5, A6, and A7 of block data BD_A. The number of symbols in A1, A2, A3, A4, A5, A6, and A7 of block data BD_A is the same as the number of symbols N of synchronization word SW_A.

The data frame DF_B transmitted from base station B comprises, from the leftmost beginning, a synchronization word SW_B, a no-data section ND_B, and multiple, for example, seven, blocks of data BD_B. Data frame DF_B is identical to data frame DF_A from base station A, and the synchronization word SW_B, no-data section ND_B, and block data BD_B correspond to the synchronization word SW_A, no-data section ND_A, and block data BD_A, respectively.

Block numbers B1, B2, B3, B4, B5, B6, and B7 are assigned to block data BD_B, respectively. The transmission data stored in B1, B2, B3, B4, B5, B6, and B7 of block data BD_B is the same as the transmission data stored in A1, A2, A3, A4, A5, A6, and A7 of block data BD_A in data frame DF_A.

The no-data section ND_B is provided so that when mobile station D simultaneously receives data frame DF_A from station A and data frame DF_B from station B, the first block of data BD_A1 in data frame DF_A can be received independently without overlapping with other data (in other words, without being combined).

Base station A and base station B are connected, for example, to a central control unit (not shown) of a disaster prevention radio system. The wireless communication devices 2 of base station A and base station B modulate and convert transmission data received from this central control unit into data frames DF_A and DF_B, and transmit the data frames DF_A and DF_B in accordance with the instructions of the central control unit. In order to avoid beat interference at the time of reception, base station B transmits data frame DF_B with a time difference equivalent to the number of symbols N (32 symbols) of synchronization word SW_A, as offset symbol number M (shown as Msym in the figure), relative to the transmission of data frame DF_A. In other words, base station B delays the transmission of data frame DF_B by the number of symbols M relative to data frame DF_A. By delaying the transmission of data frame DF_B relative to data frame DF_A, when wireless equipment 3 of mobile station D simultaneously receives data frames DF_A and DF_B, it can use the no-data intervals ND_A and ND_B to receive synchronization word SW_A, synchronization word SW_B, and A1 of block data BD_A at the beginning of data frame DF_A independently without overlapping with (combining with) other data.

Data frames DF_A and DF_B transmitted as radio waves from antennas 2a of wireless communication devices 2 at base stations A and B are combined as they propagate through space, generating a combined data frame DF_A+DF_B, as shown in Figure 2(B).

The combined data frame DF_A+DF_B is created by transmitting the data frame DF_B of base station B offset by the number of symbols M relative to the data frame DF_A of base station A, so that the synchronization word SW_A of data frame DF_A, the synchronization word SW_B of data frame DF_B, the block data BD_A of data frame DF_A, and the block data BD_B of data frame DF_B are combined in transmission order. As shown in Figure 2(A), data frames DF_A and DF_B contain no-data intervals ND_A and ND_B, so the synchronization word SW_B is placed at the position of the no-data interval ND_A, and A1 of block data BD_A is placed at the position of the no-data interval ND_B.

Furthermore, among the multiple block data BD_A A1 to A7 and block data BD_B B1 to B7, block data transmitted at the same time are combined to generate combined block data.

Specifically, A1 of block data BD_A and the no-data section ND_B of block data BD_B are combined to generate composite block data A1. Similarly, A2 of block data BD_A and B1 of block data BD_B are combined to generate composite block data A2+B1. A3 of block data BD_A and B2 of block data BD_B are combined to generate combined block data A3+B2; A4 of block data BD_A and B3 of block data BD_B are combined to generate combined block data A4+B3; A5 of block data BD_A and B4 of block data BD_B are combined to generate combined block data A5+B4; A6 of block data BD_A and B5 of block data BD_B are combined to generate combined block data A6+B5; and A7 of block data BD_A and B6 of block data BD_B are combined to generate combined block data A7+B6. Note that B7 of block data BD_B is placed after combined block data A7+B6 because its reception time does not overlap with other data.

Data frames DF_A and DF_B transmitted as radio waves from wireless communication devices 2 of base stations A and B are received as a composite wave of combined data frame DF_A+DF_B by wireless communication device 3 of mobile station D. A received signal r _k of this composite wave at time k can be expressed by the following equation (1).

Here, in equation (1), L _A is the number of effective paths between base station A and mobile station D (multipath if two or more), L _B is the number of effective paths between base station B and mobile station D (multipath if two or more), h _A,i is the impulse response value between base station A and mobile station D, h _B,i is the impulse response value between base station B and mobile station D, x _k is the primary modulation symbol transmitted from base station A at time k, and n _k is noise (e.g., additive white Gaussian noise: AWGN) added by mobile station D at time k. As can be seen from equation (1), signal loss due to beat interference can be avoided by transmitting data frames offset by M symbols at base station A and base station B.

Figure 3 is a functional block diagram showing the general configuration of wireless communication device 3 of mobile station D. Note that wireless communication devices 2 of base station A and base station B have the same configuration as wireless communication device 3, so detailed explanations are omitted.

The wireless communication device 3 includes a modulation unit 31 and a transmission unit 32 used to transmit data frames, a reception unit (reception means) 33 used to receive data frames, a decision feedback adaptive equalizer 34, and a demodulation unit 35, and an interface unit 36.

The interface unit 36 mainly comprises a data circuit-terminating device 361 (including equipment known as data communications equipment or data circuit equipment). The interface unit 36 receives input of transmission data to be communicated and outputs the transmission data to the modulation unit 31 via the data circuit-terminating device 361.

The modulation unit 31 receives transmission data output from the interface unit 36, inserts a synchronization word into the transmission data to generate a data frame, and then superimposes a carrier signal of a predetermined frequency onto the data frame to digitally modulate and output it. Note that the modulation method used in the wireless communication system 1 is not limited to PSK, and QAM, etc., may also be used.

The transmitter 32 receives the digitally modulated data frame output from the modulator 31, performs digital-to-analog conversion on the data frame using a D/A converter, and then converts it into a high-frequency signal using a local oscillator and mixer. The transmitter 32 also passes the frequency-converted data frame through a transmission filter that only passes signals in a specified frequency band, and then amplifies it using a power amplifier before outputting it.

The data frame is then digitally modulated by the modulator 31 and frequency-converted by the transmitter 32. It is then guided from the transmitter 32 via the splitter 37 to the antenna 3a, and is then transmitted as radio waves from the antenna 3a via the wireless line 4 to the antennas 2a of the wireless communication devices 2 at base stations A and B.

Furthermore, when data frames DF_A and DF_B are transmitted as radio waves from antennas 2a of wireless communication devices 2 of base stations A and B, data frames DF_A and DF_B are combined as they propagate through space, generating a combined data frame DF_A+DF_B. When combined data frame DF_A+DF_B is received by antenna 3a of wireless communication device 3 of mobile station D, antenna 3a converts the received combined data frame DF_A+DF_B into an electrical signal (received signal) and outputs it.

The combined data frame DF_A+DF_B converted into an electrical signal and output from the antenna 3a is guided to the receiver 33 via the splitter 37.

The receiver 33 receives the combined data frame DF_A + DF_B, passes it through a receiving filter that only passes signals in a specified frequency band, amplifies it using a preamplifier, and then converts it into a lower frequency signal using a local oscillator and mixer.

The receiver 33 further amplifies the frequency-converted signal using a power amplifier and performs analog-to-digital conversion using an A/D converter, outputting a combined data frame DF_A+DF_B consisting of a digital signal.

The decision feedback adaptive equalizer (hereinafter referred to as equalizer or DFE) 34 is an adaptive equalizer that reshapes the received signal waveform distorted by the transmission path characteristics by feeding back and weighting the combined data frame DF_A+DF_B output from the receiver 33 using signals determined by a hard decision device, thereby eliminating the effects of inter-symbol interference caused by previously determined symbols. The DFE 34 also has the function of outputting an approximation of the symbol of the block data BD_A from the combined data frame DF_A+DF_B.

The demodulation unit 35 receives the block data BD_A output from the DFE 34, demodulates the block data BD_A to extract the transmission data, and outputs the extracted transmission data to the interface unit 36.

Figure 4 is a block diagram showing the general configuration of the DFE 34. The DFE 34 includes an equalization filter unit 341, a hard decision unit 342 connected to the output of the equalization filter unit 341, an error calculation unit 343 connected to the hard decision unit 342 and the equalization filter unit 341, and a tap update unit 344 connected to the error calculation unit 343 and the equalization filter unit 341.

The equalization filter unit 341 includes a feed-forward filter (hereinafter referred to as an FF filter) 3411, a feedback filter (hereinafter referred to as an FB filter) 3412, and an adder 3413 that adds the output of the FF filter 3411 and the output of the FB filter 3412. The equalization filter unit 341 reduces propagation path distortion and inter-symbol interference (ISI) by feeding back the output from the FB filter 3412 via the hard decision unit 342.

The FF filter 3411 has N _FF taps 3411a that receive the received signal r _k as input, and a plurality of delays 3411b. The FB filter 3412 has N _FB taps 3412a that receive the hard decision value y _d , _k of the output value y _k of the equalization filter unit 341 or the training signal x _k-(Tref-1) as input d _k , and a plurality of delays 3412b. The delays 3411b and 3412b provide a delay for M symbols, and as shown in FIG. 5, the filter is provided with M delays, each delaying by one symbol.

In such a DFE 34, the equalizer input at time k is N _FF received signals r _k , r _kM , ..., r _{kM(N FF-1)} over multiple times, and the filter output y _k at time k is obtained by the following equation (2). Here, a _i (i=1, ..., N _FF ) is an FF filter tap coefficient, b _i (i=1, ..., N _FB ) is an FB filter tap coefficient, and signal d _k is a hard decision value y _d ,k of the filter output y _k at time _k or a training signal x _{k-(T ref-1)} . When the equalizer input is in the synchronization word section (SW_A, SW_B) of the combined data frame DF_A+DF_B, the known signal x _{k-(T ref-1)} is used as the signal d _k , and when it is in the data section (A1, A2+B1 ... A7+B6, B7), the hard decision value y _d , _k of the filter output y _k is used. That is, the equalizer output y _d , _k can be obtained at time k.

The error calculation unit 343 calculates an error signal e _k which is the difference between the signal d _k and the filter output y _k . The tap update unit 344 uses an adaptive algorithm such as LS (Least Squares), LMS (Least Mean Squares), or RLS (Recursive Least Squares) to update the tap coefficients a _i and b _i based on the error signal e _k .

Here, to more simply explain the configuration of the DFE 34, consider the case where L _A =L _B =1 (when mobile station D is moving, even if there is only one wave arriving from base stations A and B, the propagation path impulse response value fluctuates over time, and the propagation paths of base stations A and B fluctuate independently). Also, consider that there is no AWGN. In this case, the received signal r _k can be expressed by the following equation (3).

Now, consider a case where desired symbol x _k-1 is demodulated from a received signal sequence by DFE 34 at time k. In this case, the tap coefficients of DFE 34 using the SW section may be pulled in using known signal x _k-(Tref-1) , in which case T _ref =M+1. The configuration of DFE 34 in this case is as shown in FIG. 6 . Specifically, it includes an FF filter 3411 having two taps 3411a and one delay unit 3411b to which received signal r _k is input, and an FB filter 3412 having one tap 3412a and one delay unit 3412b to which hard decision value y _d , _k of filter output value y _k or _training signal x _k-(Tref-1) is input, and the delay units 3411b and 3412b provide a delay of M symbols.

According to the characteristics of a general DFE described in the following reference (1), a diversity effect can be obtained by setting the number of taps of the FF filter to the number of delayed symbols contained in the delayed wave (in this embodiment, the wave arriving from base station B) plus 1 or more. In this embodiment, an M-symbol delay device is used, so "N _FF ≧(M/M)+1=2". Also, according to the following reference (1), the number of taps of the FB filter may be set to the number of delayed symbols contained in the delayed wave (in this embodiment, the wave arriving from base station B). In this embodiment, an M-symbol delay device is used, so "N _FB ≧(M/M)=1".
References (1): Masakazu Sanpei, Decision Feedback Adaptive Equalizer for Land Mobile Communications, Communications Research Laboratory Quarterly Report, 1991, https://www.nict.go.jp/publication/shuppan/kihou-journal/kihou-vol37no1/0501.pdf

Next, we will explain the process by which DFE34 outputs an approximate value of the symbol of block data BD_A from the combined data frame DF_A+DF_B.

The combined data frame DF_A+DF_B shown in Figure 7 represents the received signal sequence input to the DFE 34 as a frame, with the received signal for N symbol times expressed as a single block. Furthermore, the blocks further to the left in the figure represent older signals (signals received earlier).

DFE_m (m = 1, 2, ... 8) represents the same DFE 34 itself, in which the tap coefficient values of the FF filter 3411 and FB filter 3412 in each block m have been updated by the tap update unit 344. DFE_1 represents a DFE that has derived the tap coefficients at the frame start time using SW_A, SW_B, and a reference signal. The reference signal is pre-stored by the wireless communication device 3 of mobile station D.

The processing steps for outputting the approximate value of the symbol in block data BD_A from the combined data frame DF_A+DF_B are as follows [1] to [6].

[1] SW_A, SW_B section: FF filter tap coefficients and FB filter tap coefficients are derived based on SW_A, SW_B, and the reference signal. More specifically, the tap update unit 344 generates FF filter tap coefficients from SW_A and the reference signal using an adaptive algorithm such as LS, LMS, or RLS, and then generates FB filter tap coefficients using the generated FF filter tap coefficients, SW_A, SW_B, and the reference signal.

[2] Section A1: The input received signal is delayed by delay devices 3411b and 3412b.

[3] A2+B1 section: The received signal for composite block data A2+B1 and the received signal for section A1 held by delay devices 3411b and 3412b are input to the equalizer, and equalization processing is performed using the tap coefficients of DFE_1, obtaining an approximation of the transmitted symbol for block data A1 as the equalizer output.

[4] Calculate the error between the hard decision value of the approximation of the transmitted symbol in block data A1 obtained as the equalizer output and the equalizer output, and use this error, the current filter tap coefficients, and the equalizer input to update the FF filter tap coefficients and FB filter tap coefficients to the latest values to obtain DFE_2.

[5] A3+B2 section: The received signal in the A2+B1 section is delayed by delay devices 3411b and 3412b. The received signal in the A3+B2 section of the combined block data and the received signal in the A2+B1 section held by delay devices 3411b and 3412b are input to the equalizer, and equalization processing is performed using the tap coefficients of DFE_2, obtaining an approximation of the transmitted symbol in block data A2 as the equalizer output.

[6] Calculate the error between the hard decision value of the approximation of the transmitted symbol in block data A2 obtained as the equalizer output and the equalizer output, and use this error, the current filter tap coefficients, and the equalizer input to update the FF filter tap coefficients and FB filter tap coefficients to the latest values to obtain DFE_3. Thereafter, repeat the same processes as those shown in [5] and [6].

To explain the equalization process in more detail, the procedure for equalization in a certain block data section will be described. Fig. 8 shows the processing procedure of the DFE 34 in the mth combined block data section of the combined data frame DF_A+DF_B. In this processing procedure, the received signal at time k is r _k , and the DFE 34 at this time is DFE_m,k. The FF filter tap coefficients in DFE_m,k are respectively a _1,m,k , a _2,m,k , ..., a _NFF,m,k . Also, the FB filter tap coefficients in DFE_m,k are respectively b _1,m,k , b _2,m,k , ..., b _NFB,m,k .

The DFE 34 performs the following processes [1] to [5] in order at time k.
[1] The received signal r _k is input to the DFE 34 .
[2] The FF filter calculates the convolution value of the received signals r _k , r _kM , r _k-2M , ..., r _k-NFFM with the FF filter tap coefficients a _1,m,k , a _2,m,k , ..., a _NFF,m ,k . The FB filter calculates the convolution value of the hard decision values d _kM , d _k-2M , ..., d _k-NFBM of the equalizer output with the FB filter tap coefficients b _1,m,k , b _2,m,k , ..., b _NFB,m,k . Then, the filter output y _k shown in the following equation (4) is obtained.

[3] The filter output _yk is subjected to hard decision to obtain the equalizer output yd _,k . At this time, if the tap coefficients are pulled as desired, "yd _,k = xk- _(Tref-1) = _xkM ( _Tref = M + 1)" can be obtained.

Next, the tap coefficients are updated. Here, an update method based on the LMS algorithm will be described, but an algorithm such as RLS may also be used. For convenience, the filter tap coefficients in DFE_m,k are expressed as a column vector w _k = [a _1,m,k a _2,m,k ... a _NFF,m,k b _1,m,k b _2,m,k ... b _NFB,m,k ] ^T ,
The received signal held in the FF filter is defined as a column vector r _k =[r _k r _kM . . . r _k-NFFM ] ^T .

At time k, after the above demodulation process, the following tap coefficient update processes are performed in order.
[4] Calculate the error signal " _ek = _dk - _yk = yd _,k - _yk ".
[5] The filter tap coefficient w _k+1 to be used at the next time (k+1) is updated using the following equation (5), where μ is the step size and is set to a value smaller than 1. The superscript * represents a complex conjugate.

[Equation 5]
w _k+1 = w _k +μr _k e _k ^* ...(5)

Next, at time k+1, the DFE 34 performs the following processes [1] to [5] in order.
[1] The received signal r _k+1 is input to the DFE 34 .
[2] The FF filter calculates the convolution value of the received signals r _k+1 , r _k-M+1 , r _k-2M+1 , ... r _k-NFFM+1 with the FF filter tap coefficients a _1,m,k+1 , a _2,m,k +1 , ... a _NFF,m,k+1 . The FB filter calculates the convolution value of the hard decision values d _k-M+1 , d _k-2M+1 , ... d _k-NFBM+1 of the equalizer output with the FB filter tap coefficients b _1,m,k+1 , b _{2,m,k +1 , ... b NFB,m,k+1} . The filter output y _k+1 shown in the following equation (6) is then obtained.

[3] The filter output y _k+1 is subjected to hard decision to obtain the equalizer output y _d,k+1 . At this time, if the tap coefficients are pulled in as desired, "y _d,k+1 =x _k-M+1 " can be obtained.
[4] Calculate the error signal "e _k+1 =d _k+1 -y _k+1 =y _d,k+1 -y _k+1 ".
[5] The filter tap coefficient w _k+2 to be used at the next time (k+2) is updated using the following equation (7).

[Equation 7]
w _k+2 = w _k+1 +μr _k+1 e _k+1 ^* ...(7)

The same process is then repeated to obtain block data BD_A. The obtained block data BD_A is output from the DFE 34 to the demodulation unit 35.

As explained above, according to this embodiment, when the same data frame is transmitted from multiple transmitting base stations, data frames having a no-data section with the same number of symbols as the synchronization word are transmitted with a time difference equivalent to the number of symbols N of the synchronization word, with an offset symbol number M. This means that the synchronization word used in the tap coefficient acquisition process in the DFE can be received independently without overlapping with other data (in other words, without being combined), eliminating the need for a means to identify which of the multiple transmitting base stations the synchronization word belongs to (for example, by encoding the transmission signal from each station), and enabling high-speed tap coefficient acquisition.

(Embodiment 2)
Next, a second embodiment of the present invention will be described. In the first embodiment described above, the case where there are two transmitting base stations, base station A and base station B, has been described, but the present invention can also be applied to an environment where three or more base stations transmit simultaneously. Therefore, in the second embodiment, a case where the same data frame is transmitted from three transmitting base stations, base station A, base station B, and base station C, will be described. Note that detailed description of the same configuration as in the first embodiment will be omitted.

Figure 9 (A) shows an example of a data frame transmitted from base station A, base station B, and base station C. Data frame DF_A from base station A is, for example, PSK modulated and comprises, from the beginning on the left, a synchronization word SW_A, no-data intervals ND_A1 and ND_A2, and multiple, for example, seven, blocks of data BD_A. The block data BD_A are assigned block numbers A1, A2, A3, A4, A5, A6, and A7, respectively.

Synchronization word SW_A is known data comprising a predetermined number of symbols N (for example, 32 symbols in this embodiment) and indicates the beginning of data frame DF_A. No-data sections ND_A1 and ND_A2 are provided so that when mobile station D simultaneously receives data frame DF_A from base station A, data frame DF_B from base station B, and data frame DF_C from base station C, synchronization word SW_B of data frame DF_B and synchronization word SW_C of data frame DF_C can be received independently without overlapping with other data (in other words, without being combined). The number of symbols in no-data section ND_A is the same as the number of symbols N in synchronization word SW_A.

Modulated transmission data is stored in A1, A2, A3, A4, A5, A6, and A7 of block data BD_A. The number of symbols in A1, A2, A3, A4, A5, A6, and A7 of block data BD_A is the same as the number of symbols N of synchronization word SW_A.

Data frame DF_B transmitted from base station B comprises, from the leftmost beginning, a synchronization word SW_B, no-data intervals ND_B1 and ND_B2, and multiple, for example, seven, blocks of data BD_B. Data frame DF_B is identical to data frame DF_A from base station A, and the synchronization word SW_B, no-data intervals ND_B1 and ND_B2, and block data BD_B correspond to the synchronization word SW_A, no-data intervals ND_A1 and ND_A2, and block data BD_A, respectively.

Block numbers B1, B2, B3, B4, B5, B6, and B7 are assigned to block data BD_B, respectively. The transmission data stored in B1, B2, B3, B4, B5, B6, and B7 of block data BD_B is the same as the transmission data stored in A1, A2, A3, A4, A5, A6, and A7 of block data BD_A in data frame DF_A.

The no-data sections ND_B1 and ND_B2 are provided so that when mobile station D simultaneously receives data frame DF_A from station A, data frame DF_B from station B, and data frame DF_C from station C, the first block of data BD_A1 in data frame DF_A can be received independently without overlapping with other data (in other words, without being combined).

The data frame DF_C transmitted from base station C comprises, from the beginning on the left, a synchronization word SW_C, no-data intervals ND_C1 and ND_C2, and multiple, for example, seven, blocks of data BD_C. Data frame DF_C is identical to data frame DF_A from base station A, and the synchronization word SW_C, no-data intervals ND_C1 and ND_C2, and block data BD_C correspond to the synchronization word SW_A, no-data intervals ND_A1 and ND_A2, and block data BD_A, respectively.

Block numbers C1, C2, C3, C4, C5, C6, and C7 are assigned to block data BD_C, respectively. The transmission data stored in C1, C2, C3, C4, C5, C6, and C7 of block data BD_C is the same as the transmission data stored in A1, A2, A3, A4, A5, A6, and A7 of block data BD_A in data frame DF_A.

The no-data sections ND_C1 and ND_C2 are provided so that when mobile station D simultaneously receives data frame DF_A from station A, data frame DF_B from station B, and data frame DF_C from station C, the first block of data BD_A1 in data frame DF_A can be received independently without overlapping with other data (in other words, without being combined).

As in the first embodiment, base stations A, B, and C are connected to a central control unit (not shown) of the disaster prevention radio system. The wireless communication devices 2 of base stations A, B, and C modulate transmission data received from the central control unit and convert it into data frames DF_A, DF_B, and DF_C, and transmit the data frames DF_A, DF_B, and DF_C in accordance with instructions from the central control unit. In order to avoid beat interference at the time of reception, base station B transmits data frame DF_B with a time difference of M offset symbols (shown as Msym in the figure) from the data frame DF_A, which corresponds to the number N (32 symbols) of synchronization word SW_A. Similarly, in order to avoid beat interference at the time of reception, base station C transmits data frame DF_C with a time difference of 2M offset symbols from the data frame DF_A, which corresponds to the total number 2N (64 symbols) of synchronization word SW_A and synchronization word SW_B. That is, when transmitting a data frame from N _BS base stations, M is used as the offset time reference, and each station transmits the data frame by offsetting the time of M, 2M, . . . , N _BS M symbols.

Data frames DF_A, DF_B, and DF_C transmitted as radio waves from antennas 2a of wireless communication devices 2 at base stations A, B, and C are combined as they propagate through space, generating a combined data frame DF_A+DF_B+DF_C, as shown in Figure 9(B).

The combined data frame DF_A+DF_B+DF_C is obtained by offsetting data frame DF_B by M symbols relative to data frame DF_A, and offsetting data frame DF_C by 2M symbols relative to data frame DF_A, so that synchronization words SW_A, SW_B, and SW_C and block data BD_A, block data BD_B, and block data BD_C are combined in transmission order. As shown in FIG. 9(A), data frames DF_A, DF_B, and DF_C each have no-data intervals ND_A1 and ND_A2, no-data intervals ND_B1 and ND_B2, and no-data intervals ND_C1 and ND_C2. Therefore, synchronization word SW_B is placed in the no-data interval ND_A1, synchronization word SW_C is placed in the no-data intervals ND_A2 and ND_B1, and block data BD_A1 is placed in the no-data intervals ND_B2 and ND_C1.

Furthermore, among the multiple block data BD_A A1-A7, block data BD_B B1-B7, and block data BD_C C1-C7, block data transmitted at the same time are combined to generate composite block data. Specifically, composite block data A1, composite block data A2+B1, composite block data A3+B2+C1, composite block data A4+B3+C2, composite block data A5+B4+C3, composite block data A6+B5+C4, composite block data A7+B6+C5, composite block data B7+C6, and composite block data C7 are generated.

Data frames DF_A, DF_B, and DF_C transmitted as radio waves from wireless communication devices 2 of base stations A, B, and C are received as a composite wave of a composite data frame DF_A+DF_B+DF_C by wireless communication device 3 of mobile station D. A received signal r _k of this composite wave at time k can be expressed by the following equation (8).

Here, in equation (8), L _A is the number of effective paths between base station A and mobile station D (multipath if two or more), L _B is the number of effective paths between base station B and mobile station D (multipath if two or more), L _C is the number of effective paths between base station C and mobile station D (multipath if two or more), h _A,i is the impulse response value between base station A and mobile station D, h _B,i is the impulse response value between base station B and mobile station D, h _C,i is the impulse response value between base station C and mobile station D, x _k is the primary modulation symbol transmitted from base station A at time k, and n _k is the noise (AWGN) added at mobile station D at time k. As can be seen from equation (8), signal loss due to beat interference can be avoided by transmitting data frames with an offset from base station A, base station B, and base station C.

Here, to more simply explain the configuration of DFE 34, consider the case where L _A = L _B = L _C = 1 (when mobile station D is moving, even if there is only one incoming wave from base stations A, B, and C, the propagation path impulse response value fluctuates over time, and the propagation paths of base stations A, B, and C fluctuate independently). Also, consider that there is no AWGN. In this case, the received signal r _k can be expressed by the following equation (9).

Now, consider a case where desired symbol x _k-2M is demodulated from a received signal sequence by DFE 34 at time k. In this case, the tap coefficients of DFE 34 using the SW section may be pulled in using known signal x _k-(Tref-1) , in which case T _ref =2M+1. The configuration of DFE 34 in this case is as shown in FIG. 10 . Specifically, it includes an FF filter 3411 having three taps 3411a and two delays 3411b to which received signal r _k is input, and an FB filter 3412 having two taps 3412a and two delays 3412b to which hard decision value y _d , _k of filter output value y _k or _training signal x _k-(Tref-1) is input, where delays 3411b and 3412b provide a delay of M symbols.

According to the characteristics of a general DFE described in the above reference (1), a diversity effect can be obtained by setting the number of taps of the FF filter to the number of delayed symbols contained in the delayed waves (in this embodiment, the waves arriving from base stations B and C) plus one or more. In this embodiment, an M-symbol delay device is used, so "N _FF ≧(N _BS −1)+1=3". Also, according to reference (1), the number of taps of the FB filter may be set to the number of delayed symbols contained in the delayed waves (in this embodiment, the waves arriving from base stations B and C). In this embodiment, an M-symbol delay device is used, so "N _FB ≧N _BS −1=2".

In such a DFE 34, the equalizer input at time k is N _FF = three received signals r _k , r _kM , r _k-2M across multiple times, and the filter output y _k at time k is obtained by the following equation (10). Here, a ₁ , a ₂ , a ₃ are FF filter tap coefficients, b ₁ and b ₂ are FB filter tap coefficients, and signal d _k is the hard decision value y _d , _k of the filter output y _k at time k or the training signal x _k-(Tref-1) . When the equalizer input is in the synchronization word section (SW_A, SW_B, SW_C) of the combined data frame DF_A + DF_B + DF_C, the known signal x _{k- (Tref-1)} is used as the signal d k, and when it is in the data section (A1, A2 + B1 ... B7 + C6, C7), the hard decision value y _d , _k of the filter output y _k is used. That is, the equalizer output y _d , _k can be obtained at time k.

Fig. 11 explains the process of outputting an approximation of the symbol of block data BD_A from the combined data frame DF_A+DF_B+DF_C by the DFE 34 shown in Fig. 10. The process by the DFE 34 is the same as that in embodiment 1 except for the increase in the number of taps, so a detailed explanation will be omitted, but the input signal to the FF filter is as shown in the following equations (11) to (13). Furthermore, the filter output obtained when _dk = _xk-2M is as shown in the following equation (14). Here, if the relationship between the tap coefficients _a1 , _a2 , _a3 and _b1 , _b2 and the impulse response values hA _,0 , hB _,0 , hC _,0 between each base station A, B, C and mobile station D is expressed as equation (15) below, then equation (16) below is the optimal value of the tap coefficients _a1 , _a2 , _a3 and _b1 , _b2 such that _yk = xk _-2M , and these approximate values can be obtained by using an adaptive algorithm such as LMS or RLS.

As explained above, according to the second embodiment, as in the first embodiment, when the same data frame is transmitted from multiple transmitting base stations, data frames having a no-data section with the same number of symbols as the synchronization word are transmitted with a time difference equivalent to the number of symbols N of the synchronization word, with the number of offset symbols M. This means that the synchronization word used in the tap coefficient pull-in process in the DFE can be received independently without overlapping with other data (in other words, without being combined), eliminating the need for a means to identify which of the multiple transmitting base stations the synchronization word belongs to (for example, by encoding the transmission signal from each station), and enabling high-speed tap coefficient pull-in.

The above describes an embodiment of the present invention, but the specific configuration is not limited to the above embodiment, and design changes that do not deviate from the gist of the present invention are also included in the present invention.

1 Wireless communication system 2 Wireless communication device (wireless transmission device)
3. Wireless communication device (wireless receiving device)
33 Receiving unit (receiving means)
34 Decision feedback equalizer 341 Equalization filter section 342 Hard decision device 343 Error calculation section 344 Tap update section 3411 Feed-forward filter 3412 Feedback filter 3413 Adder A, B, C Base station D Mobile station DF_A, DF_B, DF_C Data frame DF_A + DF_B Combined data frame DF_A + DF_B + DF_C Combined data frame SW_A, SW_B Synchronization word BD_A, BD_B, BD_C Block data

Claims (1)

A wireless communication system comprising a plurality of transmitting base stations each having a wireless transmitting device for transmitting the same modulated data frame at the same frequency, and a mobile receiving station each having a wireless receiving device for receiving and demodulating the data frame,
the plurality of wireless transmission devices respectively transmit the data frame, the data frame including known header data, a plurality of block data each having the same number of symbols as the header data, and a no-data section disposed between the header data and a first block data and having the same number of symbols as the header data, with a time difference corresponding to the number of symbols of the header data as an offset symbol number;
the wireless receiving device comprises: receiving means for receiving a combined data frame obtained by spatially combining a plurality of the data frames and outputting a received signal of each symbol; and a decision feedback adaptive equalizer to which the received signal is input;
The decision feedback adaptive equalizer comprises:
a first process of performing tap coefficients used in equalization processing based on a received signal of the plurality of header data included in the combined data frame and a pre-stored reference signal;
a second process of performing equalization processing based on a received signal of composite block data constituting the composite data frame as an equalizer input and current tap coefficients to obtain an approximation value of a symbol of the block data included in the composite block data as an equalizer output;
a third process of calculating an error between the hard decision value of the equalizer output and the equalizer output, and updating the tap coefficients to the latest values using the calculated error, the current tap coefficients, and a new equalizer input;
repeating the second process and the third process until all approximate values of the symbols of the block data are output from the composite data frame;
A wireless communication system comprising: