CN103297087A

CN103297087A - Arrival time estimation method for ultra-wideband positioning system

Info

Publication number: CN103297087A
Application number: CN2013101748513A
Authority: CN
Inventors: 王翔; 尹勃; 徐斌; 赵泽西; 卢颖; 张溢
Original assignee: Beihang University
Current assignee: Beihang University
Priority date: 2013-05-13
Filing date: 2013-05-13
Publication date: 2013-09-11
Anticipated expiration: 2033-05-13
Also published as: CN103297087B

Abstract

A time-of-arrival estimation method for an ultra-wideband positioning system, the method has three steps: step one: energy sampling of a received signal; step two: solution of a decision threshold; step three: time-of-arrival estimation of a signal. Because the threshold solution model involved in the present invention is based on the constant false alarm rate constraint, it can achieve good arrival time estimation accuracy in different ultra-wideband channel modes; the involved calculation process pre-sorts the energy sequence, The calculation amount for completing the time-of-arrival estimation is reduced; the iterative threshold selection algorithm involved is independent, does not depend on the acquisition of channel prior information, and can be applied to the actual ultra-wideband positioning system. It has good practical value and broad application prospect in the field of signal detection and estimation technology.

Description

A time-of-arrival estimation method for ultra-wideband positioning system

技术领域technical field

本发明涉及一种脉冲超宽带信号的到达时间估计方法，尤其涉及一种超宽带定位系统的到达时间估计方法，属于信号检测与估计技术领域，它是利用迭代运算检测出多径信号的首达路径，从而估计出信号的到达时间。本发明适用于基于脉冲超宽带信号的定位系统测距过程的到达时间估计。The present invention relates to a time-of-arrival estimation method of a pulse ultra-wideband signal, in particular to a time-of-arrival estimation method of an ultra-wideband positioning system, which belongs to the technical field of signal detection and estimation, and is the first method to detect multipath signals by iterative operation path to estimate the arrival time of the signal. The invention is applicable to the time-of-arrival estimation of the ranging process of the positioning system based on the pulse ultra-wideband signal.

背景技术Background technique

近年来，随着移动通信和物联网技术的高速发展，人们对高精度定位业务的需求日益增加。脉冲超宽带技术可获得很高的时间分辨率和良好的抗多径性能，这使其从众多无线定位技术中脱颖而出，成为实现精确定位的首选方案，在军事、医疗、工业等领域的应用前景广泛。In recent years, with the rapid development of mobile communication and Internet of Things technology, people's demand for high-precision positioning services is increasing. Pulse ultra-wideband technology can obtain high time resolution and good anti-multipath performance, which makes it stand out from many wireless positioning technologies and become the first choice for precise positioning. It has application prospects in military, medical, industrial and other fields widely.

在一个无线定位系统中，对目标节点的位置信息的估计是通过分析目标节点和若干参考节点间传播的信号得到的，通常整个估计过程包括参数提取和位置解算两步。首先，系统从传播的信号中提取出与目标节点位置信息有关的参数，然后采用相应的定位算法从相关参数中解算出目标节点的位置信息。根据获得的参数的不同，定位技术可以分为基于到达角度、基于接收信号强度和基于到达时间这三种估计方式。本发明属于最后一种方式，通过估计到达时延来计算收发两端的距离，充分利用了超宽带信号较高的时间分辨率，能体现出超宽带高精度定位的优势。In a wireless positioning system, the estimation of the position information of the target node is obtained by analyzing the signals propagated between the target node and several reference nodes. Usually, the whole estimation process includes two steps of parameter extraction and position calculation. First, the system extracts the parameters related to the position information of the target node from the propagated signal, and then uses the corresponding positioning algorithm to solve the position information of the target node from the relevant parameters. According to the different parameters obtained, positioning technology can be divided into three estimation methods based on angle of arrival, based on received signal strength and based on time of arrival. The present invention belongs to the last method, and calculates the distance between the transmitting and receiving ends by estimating the arrival time delay, fully utilizes the relatively high time resolution of the ultra-wideband signal, and can reflect the advantages of ultra-wideband high-precision positioning.

发明内容Contents of the invention

1、目的：本发明的目的是提供一种超宽带定位系统的到达时间估计方法，该方法提供了一种针对脉冲超宽带信号的首达路径检测方法。所述方法中涉及一个用于解算出最优判决门限的迭代过程，它采用横虚警率约束作为解算模型，使用最大期望值算法完成迭代，由此可以在具体信道模式下先验信息未知的情况下估计出超宽带信号的到达时间。1. Purpose: The purpose of the present invention is to provide a time-of-arrival estimation method for an ultra-wideband positioning system, which provides a first-arrival path detection method for pulsed ultra-wideband signals. The method involves an iterative process for solving the optimal decision threshold, which uses the horizontal false alarm rate constraint as the solution model, and uses the maximum expected value algorithm to complete the iteration, so that the specific channel mode with unknown prior information can Estimate the time of arrival of UWB signals in this case.

2、技术方案：图1为本发明涉及的系统流程图。其中主要有两个部分：信号的能量采样和信号的到达时间估计。在信号的前处理阶段，在对接收到的信号完成帧一级同步的情况下，先对信号进行平方率检波，再对输出的能量信号进行采样，得到接收信号的能量采样序列。在到达时间估计阶段，利用迭代门限选择算法从能量采样序列中解算出判决门限，使用该门限从能量采样序列中检测出信号的首达路径，完成信号的到达时间估计。2. Technical solution: Fig. 1 is a flow chart of the system involved in the present invention. There are two main parts: the energy sampling of the signal and the time-of-arrival estimation of the signal. In the pre-processing stage of the signal, when the frame-level synchronization of the received signal is completed, the square rate detection is performed on the signal first, and then the output energy signal is sampled to obtain the energy sampling sequence of the received signal. In the arrival time estimation stage, the decision threshold is calculated from the energy sampling sequence by using the iterative threshold selection algorithm, and the first arrival path of the signal is detected from the energy sampling sequence by using the threshold, and the arrival time estimation of the signal is completed.

综上所述，本发明一种超宽带定位系统的到达时间估计方法，该方法具体步骤如下：In summary, the present invention provides a time-of-arrival estimation method for an ultra-wideband positioning system. The specific steps of the method are as follows:

步骤一：接收信号的能量采样Step 1: Energy Sampling of Received Signal

对帧同步后的超宽带信号进行平方率检波，得到接收信号的能量信号；再对此能量信号进行采样，得到接收信号的能量采样序列。The square rate detection is performed on the UWB signal after frame synchronization to obtain the energy signal of the received signal; and then the energy signal is sampled to obtain the energy sampling sequence of the received signal.

接收到的跳时脉冲超宽带信号可以表示为The received time-hopping pulse UWB signal can be expressed as

$r r ((t t)) = = {Σ Σ}_{j j = = - - \infty \infty}^{\infty \infty} {d d}_{j j} {w w}_{mp mp} ((t t - - {jT J}_{f f} - - {c c}_{j j} {T T}_{c c} - - {τ τ}_{toa toa})) + + n no ((t t))$

其中，T_f和T_c分别为帧长和码片长度；d_j为第j帧信号的极性，τ_toa为信号的到达时间；n(t)为高斯白噪声，均值为零，方差为δ²，双边功率谱密度为N₀/2；c_j是为了防止不同用户之间的信号冲突而分配的跳时码，它决定了脉冲码片在一帧中的位置，第k个用户分配到的跳时序列满足

；w_mp(t)为接收到的多径脉冲波形，可以表示为Among them, T _f and T _c are the frame length and the chip length respectively; d _j is the polarity of the jth frame signal, τ _toa is the arrival time of the signal; n(t) is Gaussian white noise, the mean value is zero, and the variance is δ ² , the bilateral power spectral density is N ₀ /2; c _j is the time-hopping code allocated to prevent signal conflicts between different users, which determines the position of the pulse chip in a frame, and the kth user allocates The time-hopping sequence obtained satisfies

; w _mp (t) is the received multipath pulse waveform, which can be expressed as

${w w}_{mp mp} ((t t)) = = \sqrt{E E.} {Σ Σ}_{l l = = 11}^{L L} {a a}_{l l} w w ((t t - - {τ τ}_{l l}))$

w(t)为能量归一化的单个脉冲波形，持续时间为T_p；L为多径数量；a_l和τ_l分别为信道的衰减系数和延迟系数；E为脉冲能量。接收到的信号通过能量积分后，以采样间隔T_b对能量信号进行采样。令N_f为每个符号中的帧的数目，每个符号的能量序列为w(t) is a single pulse waveform normalized by energy, and the duration is T _p ; L is the number of multipaths; a _l and τ _l are the attenuation coefficient and delay coefficient of the channel, respectively; E is the pulse energy. After the received signal is integrated by energy, the energy signal is sampled at a sampling interval T _b . Let N _f be the number of frames in each symbol, the energy sequence of each symbol is

$z z [[n no]] = = {Σ Σ}_{j j = = 11}^{{N N}_{f f}} {&Integral; &Integral;}_{((j j - - 11)) {T T}_{f f} + + (({c c}_{j j} + + n no - - 11)) {T T}_{b b}}^{((j j - - 11)) {T T}_{f f} + + (({c c}_{j j} + + n no)) {T T}_{b b}} {| | r r ((t t)) | |}^{22} dt dt$

步骤二：判决门限的解算Step 2: Calculation of the decision threshold

对信号的能量采样序列进行排序，在一定的恒虚警率约束下通过迭代优化解算出一个能量门限值，用于对接收信号的首达路径进行判决。The energy sampling sequence of the signal is sorted, and an energy threshold value is calculated through an iterative optimization solution under a certain constant false alarm rate constraint, which is used to judge the first arrival path of the received signal.

在信号已经完成帧同步之后，首径的到达时间在一帧中均匀分布，考虑到帧间的串扰，将观察间隔设置为1.5倍的帧长。令T_ob为观察间隔，则序列z[n]包括个

采样能量块。其中包括纯噪声能量块和信号与噪声的叠加能量块。After the signal has completed the frame synchronization, the arrival time of the first path is evenly distributed in a frame. Considering the crosstalk between frames, the observation interval is set to 1.5 times the frame length. Let T _ob be the observation interval, then the sequence z[n] includes

Sample energy blocks. These include pure noise energy blocks and superimposed signal and noise energy blocks.

在能量序列的K个能量采样块中，包括纯噪声能量块和噪声信号叠加能量块两种。其中，纯噪声能量块的值服从中心卡方分布，均值为Mδ²，方差为2Mδ⁴，自由度为M=2BT_b+1，B为信号的带宽；噪声信号叠加能量块则服从非中心卡方分布，均值为Mδ²+E_n，方差为2Mδ⁴+4δ²E_n，E_n为该采样块的信号能量。迭代门限算法采用尼曼-皮尔逊假设检验，在恒虚警率约束下迭代出门限值，对于中心卡方分布的噪声能量块，虚警率P_fa和门限值ξ的关系如下：The K energy sampling blocks of the energy sequence include pure noise energy blocks and noise signal superimposed energy blocks. Among them, the value of the pure noise energy block obeys the central chi-square distribution, the mean is Mδ ² , the variance is 2Mδ ⁴ , the degree of freedom is M=2BT _b +1, and B is the bandwidth of the signal; the noise signal superposition energy block obeys the non-central card square distribution, the mean is Mδ ² +E _n , the variance is 2Mδ ⁴ +4δ ² E _n , and E _n is the signal energy of the sampling block. The iterative threshold algorithm adopts the Niemann-Pearson hypothesis test, and iterates out the threshold value under the constraint of constant false alarm rate. For the noise energy block with central chi-square distribution, the relationship between the false alarm rate P _fa and the threshold value ξ is as follows:

${P P}_{fa fa} = = P P ((z z [[n no]] > > ξ ξ)) = = exp exp ((- - \frac{ξ ξ}{{22 δ δ}^{22}})) {Σ Σ}_{k k = = 00}^{M m / / 22 - - 11} \frac{11}{k k!!} {((\frac{ξ ξ}{{22 δ δ}^{22}}))}^{k k}$

图2是本发明中的迭代门限选择算法的流程图。该算法的的关键是在将噪声能量块依次排除，在每次迭代中完成门限值的更新，从而检测出一部分噪声能量块，当所有噪声能量块都被检测出来时，门限的性能达到算法最优的到达时间估计精度。Fig. 2 is a flowchart of the iterative threshold selection algorithm in the present invention. The key of the algorithm is to exclude the noise energy blocks in turn, and update the threshold value in each iteration, so as to detect a part of the noise energy blocks. When all the noise energy blocks are detected, the performance of the threshold reaches the algorithm Optimal time-of-arrival estimation accuracy.

步骤三：信号的到达时间估计Step 3: Estimate the time of arrival of the signal

利用步骤二中得到的判决门限从能量采样序列中检测出信号的首达路径，信号的首达路径的到达时间即为信号到达时间的估计值。在低信噪比的条件下，通过预定的补偿策略完成门限失效情况下的到达时间估计。The first-arrival path of the signal is detected from the energy sampling sequence by using the decision threshold obtained in step two, and the arrival time of the first-arrival path of the signal is the estimated value of the signal arrival time. Under the condition of low signal-to-noise ratio, the time-of-arrival estimation in the case of threshold failure is accomplished through a predetermined compensation strategy.

在解算出最优门限值后，首达路径所在能量块的到达时间可表示为After calculating the optimal threshold value, the arrival time of the energy block where the first arrival path is located can be expressed as

${τ τ}_{ITS ITS} = = {{\underset{K K}{min min} ((k k | | z z [[k k]] {> > ξ ξ}_{opt opt})) - - 0.5 0.5}} \times \times {T T}_{b b}$

在低信噪比的条件下，纯噪声能量块和噪声信号叠加能量块间的采样值之间没有显著性的差异，这可能会导致迭代的门限值无法完成检测的情况，即迭代出的门限值大于搜索序列中所有的能量采样值。在门限失效的情况下，本方法采用了两种补偿策略。第一种是以观察序列中的最强值的到达时间作为信号的到达时间，第二种是以观察窗的中值作为信号的到达时间。Under the condition of low signal-to-noise ratio, there is no significant difference between the sampling values between the pure noise energy block and the noise signal superimposed energy block, which may lead to the situation that the iterative threshold value cannot complete the detection, that is, the iterated The threshold is greater than all energy samples in the search sequence. In the case of threshold failure, this method adopts two compensation strategies. The first one takes the arrival time of the strongest value in the observation sequence as the arrival time of the signal, and the second one takes the median value of the observation window as the arrival time of the signal.

其中，步骤二所述的迭代门限选择算法的基本流程如下所示Among them, the basic flow of the iterative threshold selection algorithm described in step 2 is as follows

1)将观察序列z[n]按升序进行排序，假设前N个能量块都为噪声块；1) Sort the observation sequence z[n] in ascending order, assuming that the first N energy blocks are all noise blocks;

2)利用这N个能量块的值和虚警概率P_fa计算出对应的门限值ξ，这样小于此门限值的能量块都被认为是噪声块；2) Utilize the values of these N energy blocks and the false alarm probability P _fa to calculate the corresponding threshold value ξ, so that the energy blocks smaller than this threshold value are all considered as noise blocks;

3)利用上一步检测出来的噪声块和虚警概率P_fa重新计算对应的门限值，更新的门限值又会检测出一组新的噪声块；3) Utilize the noise blocks detected in the previous step and the false alarm probability P _fa to recalculate the corresponding threshold value, and the updated threshold value will detect a group of new noise blocks;

4）不断重复这个过程直到迭代次数达到预设值，此时得到的门限值记为ξ_opt。4) Repeat this process until the number of iterations reaches the preset value, and the threshold obtained at this time is denoted as ξ _opt .

3、优点及功效：本发明可以检测出接收到的多径信号中的首达路径，这种到达时间估计方法主要具备以下几个优点：3. Advantages and effects: The present invention can detect the first arrival path in the received multipath signal, and this time-of-arrival estimation method mainly possesses the following advantages:

1)本发明所涉及的门限解算模型因为是基于恒虚警率约束的，在不同的超宽带信道模式都能取得很好的到达时间估计精度。1) Because the threshold solution model involved in the present invention is based on the constant false alarm rate constraint, it can achieve good arrival time estimation accuracy in different ultra-wideband channel modes.

2)本发明所涉及的计算过程对能量序列进行了预先排序，减少了完成到达时间估计的计算量。2) The calculation process involved in the present invention pre-sorts the energy sequence, which reduces the calculation amount for completing the arrival time estimation.

3)本发明所涉及的迭代门限选择算法具有独立性，不依赖于信道先验信息的获取，能够应用于实际的超宽带定位系统之中。3) The iterative threshold selection algorithm involved in the present invention is independent, does not depend on the acquisition of channel prior information, and can be applied to an actual ultra-wideband positioning system.

附图说明Description of drawings

图1是本发明方法涉及的系统流程图Fig. 1 is the system flowchart that the inventive method relates to

图2是本发明中的迭代门限选择算法流程图Fig. 2 is the iterative threshold selection algorithm flowchart in the present invention

具体实施方式Detailed ways

见图1，本发明一种超宽带定位系统的到达时间估计方法，该方法具体步骤如下：See Fig. 1, the time of arrival estimation method of a kind of UWB positioning system of the present invention, the concrete steps of this method are as follows:

; w _mp (t) is the received multipath pulse waveform, which can be expressed as

步骤二：判决门限的解算Step 2: Calculation of the decision threshold

在能量序列的K个能量采样块中，包括纯噪声能量块和噪声信号叠加能量块两种。其中，纯噪声能量块的值服从中心卡方分布，均值为Mδ²，方差为2Mδ⁴，自由度为M=2BT_b+1，B为信号的带宽；噪声信号叠加能量块则服从非中心卡方分布，均值为Mδ²+E_n，方差为2Mδ⁴+4δ²E_n，E_n为该采样块的信号能量。迭代门限算法采用尼曼-皮尔逊假设检验，在恒虚警率约束下迭代出门限值，对于中心卡方分布的噪声能量块，虚警率P_fa和门限值ξ的关系入下The K energy sampling blocks of the energy sequence include pure noise energy blocks and noise signal superimposed energy blocks. Among them, the value of the pure noise energy block obeys the central chi-square distribution, the mean is Mδ ² , the variance is 2Mδ ⁴ , the degree of freedom is M=2BT _b +1, and B is the bandwidth of the signal; the noise signal superposition energy block obeys the non-central card square distribution, the mean is Mδ ² +E _n , the variance is 2Mδ ⁴ +4δ ² E _n , and E _n is the signal energy of the sampling block. The iterative threshold algorithm adopts the Niemann-Pearson hypothesis test, and iterates out the threshold value under the constant false alarm rate constraint. For the noise energy block with the central chi-square distribution, the relationship between the false alarm rate P _fa and the threshold value ξ is as follows

图2是本发明中的迭代门限选择算法的流程图。该算法的的关键是在将噪声能量块依次排除，在每次迭代中完成门限值的更新，从而检测出一部分噪声能量块，当所有噪声能量块都被检测出来时，门限的性能达到算法最优的到达时间估计精度。该算法的基本流程如下所示Fig. 2 is a flowchart of the iterative threshold selection algorithm in the present invention. The key of the algorithm is to exclude the noise energy blocks in turn, and update the threshold value in each iteration, so as to detect a part of the noise energy blocks. When all the noise energy blocks are detected, the performance of the threshold reaches the algorithm Optimal time-of-arrival estimation accuracy. The basic flow of the algorithm is as follows

1）将观察序列z[n]按升序进行排序，假设前N个能量块都为噪声块；1) Sort the observation sequence z[n] in ascending order, assuming that the first N energy blocks are all noise blocks;

2）利用这N个能量块的值和虚警概率P_fa计算出对应的门限值ξ，这样小于此门限值的能量块都被认为是噪声块；2) Use the value of these N energy blocks and the false alarm probability _Pfa to calculate the corresponding threshold value ξ, so that the energy blocks smaller than this threshold value are considered as noise blocks;

3）利用上一步检测出来的噪声块和虚警概率P_fa重新计算对应的门限值，更新的门限值又会检测出一组新的噪声块；3) Use the noise block detected in the previous step and the false alarm probability P _fa to recalculate the corresponding threshold value, and the updated threshold value will detect a new set of noise blocks;

Claims

1. A time-of-arrival estimation method of an ultra-wideband positioning system, characterized in that: the method concrete steps are as follows:

Step 1: Energy Sampling of Received Signal

Perform square rate detection on the UWB signal after frame synchronization to obtain the energy signal of the received signal; then sample the energy signal to obtain the energy sampling sequence of the received signal;

The received time-hopping pulse UWB signal is expressed as

r r ((t t)) = = {Σ Σ}_{j j = = - - \infty \infty}^{\infty \infty} {d d}_{j j} {w w}_{mp mp} ((t t - - {jT J}_{f f} - - {c c}_{j j} {T T}_{c c} - - {τ τ}_{toa toa})) + + n no ((t t))

Among them, T _f and T _c are the frame length and the chip length respectively; d _j is the polarity of the jth frame signal, τ _toa is the arrival time of the signal; n(t) is Gaussian white noise, the mean value is zero, and the variance is δ ² , the bilateral power spectral density is N ₀ /2; c _j is the time-hopping code allocated to prevent signal conflicts between different users, which determines the position of the pulse chip in a frame, and the kth user allocates The time-hopping sequence obtained satisfies

; w _mp (t) is the received multipath pulse waveform, expressed as

{w w}_{mp mp} ((t t)) = = \sqrt{E E.} {Σ Σ}_{l l = = 11}^{L L} {a a}_{l l} w w ((t t - - {τ τ}_{l l}))

w(t) is a single pulse waveform normalized by energy, and the duration is T _p ; L is the number of multipaths; a _l and τ _l are the attenuation coefficient and delay coefficient of the channel respectively; E is the pulse energy, and the received signal After energy integration, the energy signal is sampled at a sampling interval _Tb , let _Nf be the number of frames in each symbol, and the energy sequence of each symbol is

z z [[n no]] = = {Σ Σ}_{j j = = 11}^{{N N}_{f f}} {&Integral; &Integral;}_{((j j - - 11)) {T T}_{f f} + + (({c c}_{j j} + + n no - - 11)) {T T}_{b b}}^{((j j - - 11)) {T T}_{f f} + + (({c c}_{j j} + + n no)) {T T}_{b b}} {| | r r ((t t)) | |}^{22} dt dt

Step 2: Calculation of the decision threshold

Sort the energy sampling sequence of the signal, and calculate an energy threshold value through iterative optimization solution under the constant false alarm rate constraint, which is used to judge the first arrival path of the received signal;

After the signal has completed the frame synchronization, the arrival time of the first path is evenly distributed in one frame. Considering the crosstalk between frames, the observation interval is set to 1.5 times the frame length; let T _ob be the observation interval, then the sequence z[n ] includes a

Sampling energy blocks; including pure noise energy blocks and superposition energy blocks of signal and noise;

In the K energy sampling blocks of the energy sequence, there are two types of pure noise energy blocks and noise signal superimposed energy blocks; among them, the value of the pure noise energy block obeys the central chi-square distribution, the mean value is Mδ ² , the variance is 2Mδ ⁴ , and the free The degree is M=2BT _b +1, B is the bandwidth of the signal; the noise signal superimposed energy block obeys the non-central chi-square distribution, the mean is Mδ ² +E _n , the variance is 2Mδ ⁴ +4δ ² E _n , and E _n is the The signal energy of the sampling block; the iterative threshold algorithm adopts the Niemann-Pearson hypothesis test, and iterates out the threshold value under the constraint of the constant false alarm rate. For the noise energy block with the central chi-square distribution, the false alarm rate P _fa and the threshold value ξ The relationship is as follows:

{P P}_{fa fa} = = P P ((z z [[n no]] > > ξ ξ)) = = exp exp ((- - \frac{ξ ξ}{22 {δ δ}^{22}})) {Σ Σ}_{k k = = 00}^{M m / / 22 - - 11} \frac{11}{k k!!} {((\frac{ξ ξ}{{22 δ δ}^{22}}))}^{k k}

The key of the iterative threshold selection algorithm is to exclude the noise energy blocks in turn, and update the threshold value in each iteration, so as to detect a part of the noise energy blocks. When all the noise energy blocks are detected, the threshold The performance reaches the optimal arrival time estimation accuracy of the algorithm;

Step 3: Estimate the time of arrival of the signal

Use the decision threshold obtained in step 2 to detect the first-arrival path of the signal from the energy sampling sequence, and the arrival time of the first-arrival path of the signal is the estimated value of the signal arrival time. Under the condition of low SNR, through the predetermined Time-of-arrival estimation when the compensation strategy completes the threshold failure;

After calculating the optimal threshold value, the arrival time of the energy block where the first arrival path is located is expressed as

{τ τ}_{ITS ITS} = = {{\underset{K K}{min min} ((k k | | z z [[k k]] > > {ξ ξ}_{opt opt})) - - 0.5 0.5}} {\times \times T T}_{b b}

Under the condition of low signal-to-noise ratio, there is no significant difference between the sampling values between the pure noise energy block and the noise signal superimposed energy block, which will lead to the situation that the iterative threshold value cannot complete the detection, that is, the iterated gate The limit value is greater than all the energy sampling values in the search sequence; in the case of threshold failure, two compensation strategies are adopted. The first one uses the arrival time of the strongest value in the observation sequence as the arrival time of the signal, and the second one is The median of the observation window is used as the arrival time of the signal.

2. the time-of-arrival estimation method of a kind of UWB positioning system according to claim 1, is characterized in that: the basic process of the iterative threshold selection algorithm described in step 2 is as follows:

1) Sort the observation sequence z[n] in ascending order, assuming that the first N energy blocks are all noise blocks;

2) Utilize the values of these N energy blocks and the false alarm probability P _fa to calculate the corresponding threshold value ξ, so that the energy blocks smaller than this threshold value are all considered as noise blocks;

3) Utilize the noise blocks detected in the previous step and the false alarm probability P _fa to recalculate the corresponding threshold value, and the updated threshold value will detect a group of new noise blocks;

4) Repeat this process until the number of iterations reaches the preset value, and the threshold obtained at this time is denoted as ξ _opt .