CN105743540A - Signal processing system and method for ultra wide band system - Google Patents

Abstract

The invention provides a signal processing system and a signal processing method for an ultra wide band system. The system comprises a signal preprocessing unit, a multi-path channel and a signal combining unit, wherein the signal preprocessing unit performs branch preprocessing on a signal, and after passing through the multi-path channel, the processed signals are subjected to post-RAKE combination at a receiving end via the signal combining unit. In order to achieve the maximal output signal-to-noise ratio at the receiving end, a preprocessing diversity and post-RAKE parameter estimation method based on channel matrix feature vector is obtained through theoretical analysis.

Description

A kind of signal processing system for radio ultra wide band system and method

Technical field

The invention belongs to the communications field, particularly relate to a kind of signal processing system for radio ultra wide band system and method.

Background technology

Ultra broadband (UWB) technology is a kind of emerging wireless communication technology, different from traditional wireless communication technology, UWB system does not use carrier wave, but using the very low burst pulse of dutycycle as information carrier, has that traffic rate is high, low in energy consumption, realize the features such as simple, good confidentiality.

Radio ultra wide band system uses the nanosecond even pulse of picosecond pulsewidth to communicate, therefore there is very strong multi-path resolved ability, this feature makes all to use RAKE diversity receiver catch and merge multipath energy at receiving terminal in most transmission and reception scheme, improve receptivity, improve receiving efficiency.Traditional RAKE receiver for UWB system is all realize multi-path combing at receiving terminal, never realize at transmitting terminal, other communication systems such as TDD-CDMA system once had scheme by realizing separation and the transmission of multipath at transmitting terminal, because sending up channel in a tdd system is consistent with the channel model of down channel, the parameter of uplink channel estimation is used directly for down channel, the program does not directly receive doing multi-path combing at receiving terminal, directly maximum diameter is done sampling to adjudicate, in this, implementation is referred to as pre-Rake (Pre-Rake).Signal is made time delay merging treatment in advance at transmitting terminal according to channel condition information by Pre-Rake, time delay merges the coefficient conjugation equal to the channel fading factor, the premise of scheme application becomes channel when to be channel be slowly, up channel estimates of parameters can be used for down channel, Pre-Rake process is really a kind of diversity technique, can reaching to receive close performance with RAKE, advantage is reduction of the complexity of receiving terminal.

At present, other communication systems only consider transmitting terminal rake for the processing mode of Pre-Rake, and receiving terminal is directly adjudicated, it is achieved scheme is as shown in Figure 1.And in designing for UWB system, its up channel estimates of parameters can be used for down channel, therefore, feasible the Pre-Rake signal processing for UWB system is theoretically.

First, the current PRE-RAKE scheme for UWB does not have, and do not consider further the multipath energy increased to be merged at receiving terminal, receiving terminal how is utilized to merge multipath energy further, by utilizing the channel model of UWB to find the multi-path combing method for parameter estimation of a kind of the best at theory analysis, and don't the complexity of increase system, it is this problem to be addressed by invention.

Summary of the invention

(1) to solve the technical problem that

It is an object of the invention to, it is provided that a kind of signal processing system for radio ultra wide band system and method so that the more traditional RAKE of bit error rate performance receives and has greatly improved.

(2) technical scheme

The present invention provides a kind of signal processing system for radio ultra wide band system, and radio ultra wide band system includes transmitting terminal and receiving terminal, and signal processing system includes:

Signal Pretreatment unit, for carrying out branch process at transmitting terminal to sending signal, obtains the signal of multiple branch road；

Multipath channel, for transmitting the signal of multiple branch roads to receiving terminal from transmitting terminal；

Signal merges unit, for the signal of multiple branch roads being merged at receiving terminal, wherein, signal can carry out rear RAKE merging, therefore signal merges RAKE receiver after unit can adopt.

Wherein, the expression formula sending signal is:

$\begin{array}{c}{s}_i\left(t\right)=\sqrt{P}\underset{i=0}{\overset{\mathrm{\∞}}{\Σ}}{b}_{i}w(t-{\mathrm{iT}}_s),\\ w\left(t\right)=\underset{n=0}{\overset{N_s-1}{\Σ}}{c}_{n}g(t-{\mathrm{nT}}_c),\end{array}$

Wherein, i represents that i-th time point is numbered, and n represents nth in a symbol period, b_iRepresent the character signal being numbered i, c_nFor spreading code, P is for sending signal energy, and a symbolic information is by N_SIndividual pulse forms, and g (t) is for sending impulse waveform function, T_sRepresent a symbol period；

Described Signal Pretreatment unit is according to delay parameter T=[τ_K..., τ₁] transmission signal is carried out branch process, branch's time delay of kth bar branch road is T_s-τ_k, wherein, τ_kRepresent kth bar multidiameter delay.

Further, the sampled output value vector representation after multipath channel of the signal of multiple branch roads is:

Y=[y₁(i) ..., y_M(i)]^T=Ab_iHB+N,

Wherein, y_MI () represents the m column sampling output vector value that sample number is i, A represents the fixed gain of system, and H is the matrix of 2M × K, M=T_S/T_b,

$H^T=\left(\begin{array}{c}0,\mathrm{....0...},\mathrm{0...},{\mathrm{\α}}_1,\mathrm{...0...},{\mathrm{\α}}_2,\mathrm{...0...}{\mathrm{\α}}_L\mathrm{...},0\\ 0,\mathrm{....}{\mathrm{\α}}_1\mathrm{...0...},{\mathrm{\α}}_2,\mathrm{....0},\mathrm{........}{\mathrm{\α}}_L\mathrm{...........},0\\ \mathrm{......},\mathrm{..........}\\ 0,\mathrm{...}{\mathrm{\α}}_{K-1}\mathrm{..0...}{\mathrm{\α}}_K,\mathrm{......0},\mathrm{..}{\mathrm{\α}}_L\mathrm{...0..........},0\end{array}\right),$

B=[β_K..., β₁]^TRepresent and merge vector, β_KRepresent the k-th vector being numbered K, α_lRepresent the path fading parameter in l footpath, α in multipath channel_lFor real number,L represents a total L bar multipath, and N represents the quantized values of noise.

Further, signal merges unit and at receiving terminal, signal is merged according to branch merging weight W, and described branch merges weight vector representation W=[γ₁..., γ_M]^T, γ_MRepresenting the m-th vector weighted value being numbered M, the signal after merging is expressed as:

Z (i)=W^TY=Ab_mW^THB+W^TN。

Further, system also includes: parameter estimation unit, for calculation delay parameter T, merges vector B and branch merging weight W, and described delay parameter T expression formula is:

T=argmax{ | | D | | },

Wherein, D=HH^T, according to figure eigenvalue estimate and circles theorem, obtain:

1≤||D||≤argmax{d_{I, i}+R′_i(A) },

Wherein, d_{I, i}Represent the center of circle, R '_i(A) radius of circle is represented, and,

$\mathrm{argmax}\left\{d_{i,i}\right\}=\underset{i=1}{\overset{K}{\Σ}}{{\stackrel{\‾}{\mathrm{\α}}}_i}^2,$

Represent { α₁...., α_LIn K maximum absolute value and mutually different number；

It is D eigenvalue of maximum characteristic of correspondence vector that described branch merges weight W；

The expression formula of described merging vector B is:

B=(W^TH)^T。

The present invention also provides for a kind of signal processing method for radio ultra wide band system, including:

S1, carries out branch process at transmitting terminal to sending signal, obtains the signal of multiple branch road；

S2, by the signal of multiple branch roads from transmitting terminal transmission to receiving terminal；

S3, merges the signal of multiple branch roads at receiving terminal.

Wherein, the expression formula sending signal is:

$\begin{array}{c}{s}_i\left(t\right)=\sqrt{P}\underset{i=0}{\overset{\mathrm{\∞}}{\Σ}}{b}_{i}w(t-{\mathrm{iT}}_s),\\ w\left(t\right)=\underset{n=0}{\overset{N_s-1}{\Σ}}{c}_{n}g(t-{\mathrm{nT}}_c),\end{array}$

Wherein, i represents that i-th time point is numbered, and n represents nth in a symbol period, b_iRepresent the character signal being numbered i, c_nFor spreading code, P is for sending signal energy, and a symbolic information is by N_SIndividual pulse forms, and g (t) is for sending impulse waveform function, T_sRepresent a symbol period；

Described Signal Pretreatment unit is according to delay parameter T=[τ_K..., τ₁] transmission signal is carried out branch process, branch's time delay of kth bar branch road is T_s-τ_k, wherein, τ_kRepresent kth bar multidiameter delay.

Further, the sampled output value vector representation after multipath channel of the signal of multiple branch roads is:

Y=[y₁(i) ..., y_M(i)]^T=Ab_iHB+N,

y_MI () represents the m column sampling output vector value that sample number is i, A represents the fixed gain of system, and H is the matrix of 2M × K, M=T_S/T_b,

$H^T=\left(\begin{array}{c}0,\mathrm{....0...},\mathrm{0...},{\mathrm{\α}}_1,\mathrm{...0...},{\mathrm{\α}}_2,\mathrm{...0...}{\mathrm{\α}}_L\mathrm{...},0\\ 0,\mathrm{....}{\mathrm{\α}}_1\mathrm{...0...},{\mathrm{\α}}_2,\mathrm{....0},\mathrm{........}{\mathrm{\α}}_L\mathrm{...........},0\\ \mathrm{......},\mathrm{..........}\\ 0,\mathrm{...}{\mathrm{\α}}_{K-1}\mathrm{..0...}{\mathrm{\α}}_K,\mathrm{......0},\mathrm{..}{\mathrm{\α}}_L\mathrm{...0..........},0\end{array}\right),$

B=[β_K..., β₁]^TRepresent and merge vector, β_KRepresent the k-th vector being numbered K, α_lRepresent the path fading parameter in l footpath, α in multipath channel_lFor real number,L represents a total L bar multipath, and N represents the quantized values of noise.

Further, step S3 merges weight W according to branch and at receiving terminal, signal is merged, and described branch merges weight vector representation W=[γ₁..., γ_M]^T, γ_MRepresenting the m-th vector weighted value being numbered M, the signal after merging is expressed as:

Z (i)=W^TY=Ab_mW^THB+W^TN。

Further, method also includes: S0, and calculation delay parameter T, merging vector B and branch merge weight W, and described delay parameter T expression formula is:

T=argmax{ | | D | | },

Wherein, D=HH^T, according to figure eigenvalue estimate and circles theorem, obtain:

1≤||D||≤argmax{d_{I, i}+R′_i(A) },

Wherein, d_{I, i}Represent the center of circle, R '_i(A) radius of circle is represented, and,

$\mathrm{argmax}\left\{d_{i,i}\right\}=\underset{i=1}{\overset{K}{\Σ}}{{\stackrel{\‾}{\mathrm{\α}}}_i}^2,$

Represent { α₁...., α_LIn K maximum absolute value and mutually different number；

It is D eigenvalue of maximum characteristic of correspondence vector that described branch merges weight W；

The expression formula of described merging vector B is:

B=(W^TH)^T。

(3) beneficial effect

The present invention is from diversity and the consider carefully merged, it is proposed that the signal processing system of a kind of radio ultra wide band system and method, it is possible to utilize multipath, improves receptivity.Estimating it addition, the present invention is merged into line parameter based on channel eigenvalue of maximum characteristic vector, the signal to noise ratio obtained is higher than traditional maximum rate all the time and merges RAKE receiver performance.Finally, the present invention is based on the matrix eigenvalue of maximum method of estimation of circles theorem, thus drawing the delay time estimation method of the best, the method can also apply to other communications field.

Accompanying drawing explanation

Fig. 1 is PRE-RAKE diversity processing scheme of the prior art.

Fig. 2 is the schematic diagram of the signal processing system for radio ultra wide band system provided by the invention.

Fig. 3 is the structure chart of the signal processing system for radio ultra wide band system provided by the invention.

Fig. 4 is the matrix eigenvalue of maximum method of estimation schematic diagram based on circles theorem provided by the invention.

Detailed description of the invention

For making the object, technical solutions and advantages of the present invention clearly understand, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in more detail.

As shown in Figures 2 and 3, signal processing system provided by the invention includes a pretreatment unit at transmitting terminal and a rear RAKE receiver at receiving terminal.

Because in uwb systems, signal adopts DS-BPSK modulation system, and it sends signal and is represented by:

$s_i\left(t\right)=\sqrt{P}\underset{i=0}{\overset{\mathrm{\∞}}{\Σ}}{b}_{i}w(t-{\mathrm{iT}}_s)---\left(1\right)$

$w\left(t\right)=\underset{n=0}{\overset{N_s-1}{\Σ}}{c}_{n}g(t-{\mathrm{nT}}_c)---\left(2\right)$

Wherein, s_i{-1 ,+1} represents the binary information of modulation, c to ∈_nFor spreading code, P is for sending signal energy, and a symbolic information is by N_SIndividual pulse forms, and in order to meet the definition of ultra-broadband signal, g (t) is for sending impulse waveform function, and function uses 5 rank Gaussian functions:

$g\left(t\right)=K_2(-15\frac{t}{\mathrm{\σ}}+10\frac{t^3}{{\mathrm{\σ}}^3}-\frac{t^5}{n^5})e^{\frac{t^2}{2{\mathrm{\σ}}^2}}---\left(3\right)$

As it is shown on figure 3, send signal s_iT () first passes around pretreatment unit and carries out branch process, branch's time delay is T_s-τ_k, (τ_k≤T_s), if merging vector is B=[β_K..., β₁]^T, in order to keep normalization, it is desirable to | | B | |²=1.UWB signal faces serious multipath fading in indoor transmissions, UWB channel model is different from Rayleigh and Rice channel model, according to the up-to-date description about UWB channel model of the IEEE802.15.3TG3A working group, UWB indoor channel model can respond with the pulse shock of tapped delay line structure and describe:

$h\left(t\right)=\underset{l=1}{\overset{L}{\Σ}}{\mathrm{\α}}_l\mathrm{\δ}(t-{\mathrm{\τ}}_l)---\left(4\right)$

α_lRepresent the path fading parameter in l footpath, α_lFor real number,τ_lRepresenting signal time delay, L is the multipath number in multipath channel.Signal time delay by bunch and footpath distinguish, be respectively provided with each independent distribution function.

Signal is through multipath channel, and the reception signal at receiving terminal is:

$r_i\left(t\right)=\sqrt{P}\underset{l=1}{\overset{L}{\Σ}}\underset{k=1}{\overset{K}{\Σ}}{\mathrm{\β}}_k{\mathrm{\α}}_l{b}_{i}w(t+{\mathrm{\τ}}_k-{\mathrm{\τ}}_l+T_s)+n\left(t\right)---\left(5\right)$

Wherein n (t) is white Gaussian noise, by pulse matching wave filter, does relevant and despreading process to received signal, obtains output:

$y_m\left(i\right)={\∫}_{\left(i\right)T_s}^{(1+i)T_s}r(t-{\mathrm{\τ}}_m)*w\left(t\right)dt---\left(6\right)$

Signal after matched filtering is done time delay sampling, and sample rate is R=1/T_b, each branch time delay is T_s-τ_m, output includes M value y_m(i), y_mI () includes coding self-interference item and noise item, it is assumed that spreading code has good autocorrelation performance, ignores coding self-interference item, only considers white Gaussian noise item.The sampled output value vector representation then merging Qian Ge branch is: Y=[y₁(i) ..., y_M(i)]^T=Ab_iHB+N.Wherein, H is the matrix of 2M × K, M=T_S/T_b,

$H^T=\left(\begin{array}{c}0,\mathrm{....0...},\mathrm{0...},{\mathrm{\α}}_1,\mathrm{...0...},{\mathrm{\α}}_2,\mathrm{...0...}{\mathrm{\α}}_L\mathrm{...},0\\ 0,\mathrm{....}{\mathrm{\α}}_1\mathrm{...0...},{\mathrm{\α}}_2,\mathrm{....0},\mathrm{........}{\mathrm{\α}}_L\mathrm{...........},0\\ \mathrm{......},\mathrm{..........}\\ 0,\mathrm{...}{\mathrm{\α}}_{K-1}\mathrm{..0...}{\mathrm{\α}}_K,\mathrm{......0},\mathrm{..}{\mathrm{\α}}_L\mathrm{...0..........},0\end{array}\right)---\left(7\right)$

At receiving terminal signal being done and merge, branch merges weight vector representation W=[γ₁..., γ_M]^T, the signal after merging is expressed as:

Z (i)=W^TY=Ab_mW^THB+W^TN(8)

The parameter estimation algorithm of the best is obtained below by theory analysis.

By formula (7) it can be seen that output signal-to-noise ratioThe problem of optimization estimated value is just changed into: seek best parameter B=[β_K..., β₁]^T, W=[γ₁..., γ_M]^TWith T=[τ_K..., τ₁] so that output signal-to-noise ratio is maximum.In the estimation, first assume | W^T|²With T=[τ_K..., τ₁] i.e. H^TFixing, seek optimal vector B, according to Cauchy inequality:

$SNR=P_0\frac{|W^{T}HB{|}^2}{|W^T{|}^2}\≤P_0\frac{|W^T{\mathrm{HH}}^{T}W{|}^2}{|W^T{|}^2}---\left(9\right)$

As B=(W^TH)^TTime inequality take equal sign, now output signal-to-noise ratio is maximumAsk best W estimates of parameters, optimal parameter againDetermined by following formula:

$\tilde{W}=\mathrm{argmax}\left\{\frac{W^T{\mathrm{HH}}^{T}W}{W^{T}W}\right\}---\left(10\right)$

By formula (7) (10) it can be seen that matrix D=HH^TFor 2M rank positive definite E Mite matrix, and SNR/P₀It is actually the Rayleigh entropy of matrix D.According to Rayleigh entropy eigenvalue maximal value theorem,Wherein, λ₁≥λ₂≥…≥λ_mFor matrix D=HH^TEigenvalue, when W is matrix D eigenvalue of maximum characteristic of correspondence vector, SNR/P₀Maximum, it may be assumed that SNR_max=P₀λ₁.Therefore, this parameter estimation algorithm merges algorithm also referred to as eigenvalue of maximum.

Analyzing best T=[τ_K..., τ₁] before, analyze the signal to noise ratio under an extreme case first: as K=1, now there is no preprocessing branch, i.e. B=1, by (7) (9) (10) formula, obtain W=[α₁..., α_L]^T, this is exactly maximum rate and merges the best merging parameter, the now output signal-to-noise ratio of RAKE receiverWherein,σ²For noise variance.

And as B=[α_L..., α₁]^T, W=[0 ..., 1 ..., 0]^TTime, this mode is Pre-Rake diversity mode, the signal to noise ratio of output under Pre-Rake diversity modeIt is therefore seen that the performance of Pre-Rake diversity is equal to RAKE receptivity.

Analyze the situation of K > 1 again, (7) be apparent from, rank (HH^T)=K, because λ₁≥λ₂≥…≥λ_m, so:

${\mathrm{K\λ}}_1\≥\underset{i=1}{\overset{K}{\Σ}}{\mathrm{\λ}}_i=K\underset{i=1}{\overset{L}{\Σ}}|{\mathrm{\α}}_i{|}^2---\left(11\right)$

Can be seen that from inequality, as K > 1, merging the output signal-to-noise ratio that obtains of method always greater than traditional RAKE receiver performance based on maximal eigenvector, the maximum rate that namely lower bound of its output signal-to-noise ratio is traditional merges RAKE receiver output signal-to-noise ratio.Obviously, output signal-to-noise ratio and matrix D=HH^TEigenvalue of maximum | | D | | there is a upper bound, analyze the upper bound of | | D | | now by matrix exgenvalue estimation theory.

By (7) formula it can be seen that different the putting in order of matrix H brings D=HH^TEigenvalue of maximum | | D | | changes.So best delay parameter estimated value T=[τ_K..., τ₁] determined by following formula:

$\tilde{T}=\mathrm{argmax}\left\{\right|\left|D\right|\left|\right\}---\left(12\right)$

About solving T=[τ_K..., τ₁] best estimates of parameters can by arriving that complicated singular value decomposition algorithm is searched for.Shown below is a relatively simple method of estimation.It is sparse matrix according to UWB channel characteristic D, and | | D | | >=1.According to figure eigenvalue estimate and circles theorem: the value of | | D | | drops on dash area region as shown in Figure 4.

In Fig. 4, the radius of circled_{I, j}Value for matrix D the i-th row j row.It is readily obtained by (7), center of circle d_{I, j}It is consistently less than 1, and because eigenvalue of maximum is real number, and according to (11) formula, | | D | | is consistently greater than 1, so, | | D | | drops on following scope:

1≤||D||≤argmax{d_{I, i}+R′_i(A)}(13)

Wherein, according to UWB channel characteristic, D is sparse matrix, R '_i(A) it is held essentially constant, therefore best T=[τ_K..., τ₁] essential condition of parameter is:

$\mathrm{argmax}\left\{d_{i,i}\right\}=\underset{i=1}{\overset{K}{\Σ}}{{\stackrel{\‾}{\mathrm{\α}}}_i}^2---\left(14\right)$

WhereinRepresent { α₁..., α_LIn K maximum absolute value and mutually different number.Meet this condition, it is desirable in formula (7), matrix H is at certain string N < 2M, and column vector comprises this this K maximum absolute valueAnd nowThe time delay being precisely in channel K most powerful path, namely obtains a simple T=[τ_K..., τ₁] method for parameter estimation.

So far, best merging and delay parameter B=[β_K..., β₁]^T, W=[γ₁..., γ_M]^TWith T=[τ_K..., τ₁] method of estimation obtains, method can simply be summarized as follows several step:

1, receiving terminal obtains channel information by channel estimating, finds out the time delay of K bar most powerful path, parameter

2, according to formula (7) (10) to matrix D=HH^TDo singular value decomposition and obtain the estimates of parameters that eigenvalue of maximum characteristic of correspondence vector is namely best

3, according to B=(W^TH)^TCalculate B and go out the estimated value of the bestAnd parameter will be estimatedWithPass to transmitting terminal and do time delay merging pretreatment.

4, receiving terminal is according to delay parameterObtain delay parameter in conjunction with channel estimation parameter, pass through weight coefficientMerge and obtain final output.

Wherein, obtained by svd algorithm for Singular Value Decomposition Using, according to (8) formula, to D=HH^TDo SVD decomposition and obtain D=U Λ V^T, matrix Λ=diag [λ₁..., λ_K+L-1], wherein, λ₁For eigenvalue of maximum, the estimates of parameters of W is the first row vector value of matrix U.Because UWB channel matrix D is sparse matrix, it is possible to by the svd algorithm simplified to realize singular value decomposition, it is substantially reduced the complexity of computing.

Simultaneously it will be noted that estimate that in singular value decomposition, estimated value likely produces " upset effect " and causes erroneous judgement at actual parameter.Namely estimated value is obtained by SVD decomposition valueBut the estimates of parameters of reality should beAnd the transformation of symbol is random, if this problem is not processed, it may occur that mistake in judgment, produce symbol upset effect.The reason producing this problem is, when seeking output signal-to-noise ratio extreme value, formula (9) has been taken absolute value, thereforeIt is the solution of equation, and the solution only one of which of reality.The method solved is by estimated valueTake advantage of in the judgement factorMake to correct and obtain correct value.

Particular embodiments described above; the purpose of the present invention, technical scheme and beneficial effect have been further described; it is it should be understood that; the foregoing is only specific embodiments of the invention; it is not limited to the present invention; all within the spirit and principles in the present invention, any amendment of making, equivalent replacement, improvement etc., should be included within protection scope of the present invention.

Claims (10)

1., for a signal processing system for radio ultra wide band system, radio ultra wide band system includes transmitting terminal and receiving terminal, it is characterised in that signal processing system includes:

Signal Pretreatment unit, for carrying out branch process at transmitting terminal to sending signal, obtains the signal of multiple branch road；

Multipath channel, for transmitting the signal of the plurality of branch road to receiving terminal from transmitting terminal；

Signal merges unit, for the signal of the plurality of branch road being merged at receiving terminal.

2. the signal processing system for radio ultra wide band system according to claim 1, it is characterised in that the expression formula sending signal is:

$\begin{array}{c}{s}_i\left(t\right)=\sqrt{P}\underset{i=0}{\overset{\mathrm{\&infin;}}{\mathrm{\&Sigma;}}}{b}_{i}w\left(t-{\mathrm{iT}}_s\right),\\ w\left(t\right)=\underset{n=0}{\overset{N_s-1}{\&Sigma;}}{c}_{n}g(t-{\mathrm{nT}}_c),\end{array}$

Wherein, i represents that i-th time point is numbered, and n represents nth in a symbol period, b_iRepresent the character signal being numbered i, c_nFor spreading code, P is for sending signal energy, and a symbolic information is by N_sIndividual pulse forms, and g (t) is for sending impulse waveform function, T_sRepresent a symbol period；

Described Signal Pretreatment unit is according to delay parameter T=[τ_K..., τ₁] transmission signal is carried out branch process, branch's time delay of kth bar branch road is T_s-τ_k, wherein, τ_kRepresent kth bar multidiameter delay.

3. the signal processing system for radio ultra wide band system according to claim 2, it is characterised in that the sampled output value vector representation after multipath channel of the signal of multiple branch roads is:

Y=[y₁(i) ..., y_M(i)]^T=Ab_iHB+N,

Wherein, y_MI () represents the m column sampling output vector value that sample number is i, A represents the fixed gain of system, and H is the matrix of 2M × K, M=T_s/T_b,

$H^T=\left(\begin{array}{c}0,\mathrm{....0...},\mathrm{0...},{\mathrm{\&alpha;}}_1,\mathrm{...0...},{\mathrm{\&alpha;}}_2,\mathrm{...0...}{\mathrm{\&alpha;}}_L\mathrm{...},0\\ 0,\mathrm{....}{\mathrm{\&alpha;}}_1\mathrm{...0...},{\mathrm{\&alpha;}}_2,\mathrm{....0},\mathrm{........}{\mathrm{\&alpha;}}_L\mathrm{...........},0\\ \mathrm{......},\mathrm{..........}\\ 0,\mathrm{...}{\mathrm{\&alpha;}}_{K-1}\mathrm{..0...}{\mathrm{\&alpha;}}_K,\mathrm{......0},\mathrm{..}{\mathrm{\&alpha;}}_L\mathrm{...0..........},0\end{array}\right),$

B=[β_K..., β₁]^TRepresent and merge vector, β_KRepresent the k-th vector being numbered K, α_lRepresent the path fading parameter in l footpath, α in multipath channel_lFor real number,L represents a total L bar multipath, and N represents the quantized values of noise.

4. the signal processing system for radio ultra wide band system according to claim 3, it is characterised in that signal merges unit and at receiving terminal, signal merged according to branch merging weight W, and described branch merges weight vector representation W=[γ₁..., γ_M]^T, γ_MRepresenting the m-th vector weighted value being numbered M, the signal after merging is expressed as:

Z (i)=W^TY=Ab_mW^THB+W^TN。

5. the signal processing system for radio ultra wide band system according to claim 4, it is characterised in that also include:

Parameter estimation unit, for calculation delay parameter T, merges vector B and branch merging weight W, and described delay parameter T expression formula is:

T=argmax{ | | D | | },

Wherein, D=HH^T, according to figure eigenvalue estimate and circles theorem, obtain:

1≤||D||≤argmax{d_{I, i}+R′_i(A) },

Wherein, d_{I, i}Represent the center of circle, R '_i(A) radius of circle is represented, and,

$\arg \max \left\{d_{i,i}\right\}=\underset{i=1}{\overset{K}{\&Sigma;}}{{\stackrel{\&OverBar;}{\mathrm{\&alpha;}}}_i}^2,$

Represent { α₁..., α_LIn K maximum absolute value and mutually different number；

It is D eigenvalue of maximum characteristic of correspondence vector that described branch merges weight W；

The expression formula of described merging vector B is:

B=(W^TH)^T。

6. the signal processing method for radio ultra wide band system, it is characterised in that radio ultra wide band system includes transmitting terminal and receiving terminal, it is characterised in that signal processing method includes:

S1, carries out branch process at transmitting terminal to sending signal, obtains the signal of multiple branch road；

S2, by the signal of the plurality of branch road from transmitting terminal transmission to receiving terminal；

S3, merges the signal of the plurality of branch road at receiving terminal.

7. the signal processing method for radio ultra wide band system according to claim 6, it is characterised in that the expression formula sending signal is:

$\begin{array}{c}{s}_i\left(t\right)=\sqrt{P}\underset{i=0}{\overset{\mathrm{\&infin;}}{\mathrm{\&Sigma;}}}{b}_{i}w\left(t-{\mathrm{iT}}_s\right),\\ w\left(t\right)=\underset{n=0}{\overset{N_s-1}{\&Sigma;}}{c}_{n}g(t-{\mathrm{nT}}_c),\end{array}$

Wherein, i represents that i-th time point is numbered, and n represents nth in a symbol period, b_iRepresent the character signal being numbered i, c_nFor spreading code, P is for sending signal energy, and a symbolic information is by N_sIndividual pulse forms, and g (t) is for sending impulse waveform function, T_sRepresent a symbol period；

Described Signal Pretreatment unit is according to delay parameter T=[τ_K..., τ₁] transmission signal is carried out branch process, branch's time delay of kth bar branch road is T_s-τ_k, wherein, τ_kRepresent kth bar multidiameter delay.

8. the signal processing method for radio ultra wide band system according to claim 7, it is characterised in that the sampled output value vector representation after multipath channel of the signal of multiple branch roads is:

Y=[y₁(i) ..., y_M(i)]^T=Ab_iHB+N,

y_MI () represents the m column sampling output vector value that sample number is i, A represents the fixed gain of system, and H is the matrix of 2M × K, M=T_s/T_b,

$H^T=\left(\begin{array}{c}0,\mathrm{....0...},\mathrm{0...},{\mathrm{\&alpha;}}_1,\mathrm{...0...},{\mathrm{\&alpha;}}_2,\mathrm{...0...}{\mathrm{\&alpha;}}_L\mathrm{...},0\\ 0,\mathrm{....}{\mathrm{\&alpha;}}_1\mathrm{...0...},{\mathrm{\&alpha;}}_2,\mathrm{....0},\mathrm{........}{\mathrm{\&alpha;}}_L\mathrm{...........},0\\ \mathrm{......},\mathrm{..........}\\ 0,\mathrm{...}{\mathrm{\&alpha;}}_{K-1}\mathrm{..0...}{\mathrm{\&alpha;}}_K,\mathrm{......0},\mathrm{..}{\mathrm{\&alpha;}}_L\mathrm{...0..........},0\end{array}\right),$

B=[β_K..., β₁]^TRepresent and merge vector, β_KRepresent the k-th vector being numbered K, α_lRepresent the path fading parameter in l footpath, α in multipath channel_lFor real number,L represents a total L bar multipath, and N represents the quantized values of noise.

9. the signal processing method for radio ultra wide band system according to claim 8, it is characterised in that described step S3 merges weight W according to branch and at receiving terminal, signal merged, and described branch merges weight vector representation W=[γ₁..., γ_M]^T, γ_MRepresenting the m-th vector weighted value being numbered M, the signal after merging is expressed as:

Z (i)=W^TY=Ab_mW^THB+W^TN。

10. the signal processing method for radio ultra wide band system according to claim 9, it is characterised in that also include:

S0, calculation delay parameter T, merging vector B and branch merge weight W, and described delay parameter T expression formula is:

T=argmax{ | | D | | },

Wherein, D=HH^T, according to figure eigenvalue estimate and circles theorem, obtain:

1≤||D||≤argmax{d_{I, i}+R′_i(A) },

Wherein, d_{I, i}Represent the center of circle, R '_i(A) radius of circle is represented, and,

$\arg \max \left\{d_{i,i}\right\}=\underset{i=1}{\overset{K}{\&Sigma;}}{{\stackrel{\&OverBar;}{\mathrm{\&alpha;}}}_i}^2,$

Represent { α₁..., α_LIn K maximum absolute value and mutually different number；

It is D eigenvalue of maximum characteristic of correspondence vector that described branch merges weight W；

The expression formula of described merging vector B is:

B=(W^TH)^T。