CN103297087B - A kind of method of estimation time of advent of ultra-wideband positioning system - Google Patents

Abstract

The method of estimation time of advent for ultra-wideband positioning system, the method has three large steps: step one: the energy sampled of Received signal strength; Step 2: resolving of decision threshold; Step 3: the time of advent of signal is estimated.Thresholding involved in the present invention resolves model because based on constant false alarm rate constraint, can obtain estimated accuracy good time of advent in different ultra-wideband channel patterns; Involved computational process has carried out pre-sorting to energy sequence, has decreased the amount of calculation arriving time Estimate; Involved iteration threshold selection algorithm has independence, does not rely on the acquisition of channel prior information, can be applied among actual ultra-wideband positioning system.It has good practical value and wide application prospect in signal detection and estimation technical field.

Description

A kind of method of estimation time of advent of ultra-wideband positioning system

Technical field

The present invention relates to a kind of method of estimation time of advent of impulse ultra-wideband signal, particularly relate to a kind of method of estimation time of advent of ultra-wideband positioning system, belong to signal detection and estimation technical field, it be utilize interative computation to detect the head of multipath signal reaches path, thus estimate the time of advent of signal.The present invention is applicable to based on estimating the time of advent of the navigation system ranging process of impulse ultra-wideband signal.

Background technology

In recent years, along with the high speed development of mobile communication and technology of Internet of things, the demand of people to hi-Fix business increases day by day.Pulse ultra-broad band technology can obtain very high temporal resolution and good anti-multipath performance, and this makes it show one's talent from numerous wireless location technology, becomes and realizes pinpoint preferred option, extensive in the application prospect in the field such as military, medical, industrial.

In a wireless location system, be that signal by propagating between evaluating objects node and some reference nodes obtains to the estimation of the positional information of destination node, usual whole estimation procedure comprises parameter extraction and location compute two step.First, system extracts the parameter information-related with destination node location from the signal propagated, and then adopts corresponding location algorithm from relevant parameter, calculate the positional information of destination node.According to the difference of the parameter obtained, location technology can be divided into based on angle of arrival, based on received signal strength with based on these three kinds of estimation modes time of advent.The invention belongs to last a kind of mode, by estimating to arrive the distance that time delay calculates transmitting-receiving two-end, taking full advantage of the temporal resolution that ultra-broadband signal is higher, the advantage of ultra broadband hi-Fix can be embodied.

Summary of the invention

1, object: the method for estimation time of advent that the object of this invention is to provide a kind of ultra-wideband positioning system, this method provides a kind of head for impulse ultra-wideband signal and reaches path detection method.One is related to for calculating the iterative process of optimum decision thresholding in described method, it adopts horizontal false alarm rate to retrain as resolving model, use greatest hope value-based algorithm to complete iteration, under concrete channelling mode, the time of advent of ultra-broadband signal when prior information the unknown, can be estimated thus.

2, technical scheme: Fig. 1 is the system flow chart that the present invention relates to.Wherein mainly contain two parts: the energy sampled of signal and the time of advent of signal are estimated.In the pretreatment stage of signal, when completing frame level synchronization to the received signal, first square law detection is carried out to signal, then the energy signal exported is sampled, obtain the energy sampled sequence of Received signal strength.In the estimation stages time of advent, utilize iteration threshold selection algorithm to calculate decision threshold from energy sampled sequence, the head of signal reaches path to use this thresholding to detect from energy sampled sequence, and the time of advent of settling signal is estimated.

In sum, the method for estimation time of advent of a kind of ultra-wideband positioning system of the present invention, the method concrete steps are as follows:

Step one: the energy sampled of Received signal strength

Square law detection is carried out to the ultra-broadband signal after frame synchronization, obtains the energy signal of Received signal strength; Again this energy signal is sampled, obtain the energy sampled sequence of Received signal strength.

During the jumping received, impulse ultra-wideband signal can be expressed as

$r\left(t\right)=\underset{j=-\∞}{\overset{\∞}{\mathrm{\Σ}}}{d}_j{w}_{\mathrm{mp}}(t-{\mathrm{jT}}_f-c_j{T}_c-{\mathrm{\τ}}_{\mathrm{toa}})+n\left(t\right)$

Wherein, T _fand T _cbe respectively frame length and chip lengths; d _jfor the polarity of jth frame signal, τ _toafor the time of advent of signal; N (t) is white Gaussian noise, and average is zero, and variance is δ ², bilateral power spectral density is N ₀/ 2; c _jbeing the time-hopping code in order to prevent the signal conflict between different user from distributing, which determining pulse chip position in a frame, the time-hopping sequence that a kth user is assigned to meets ; w _mpt () is the multi-path pulse waveform received, can be expressed as

$w_{\mathrm{mp}}\left(t\right)=\sqrt{E}\underset{l=1}{\overset{L}{\mathrm{\Σ}}}{a}_{l}w(t-{\mathrm{\τ}}_l)$

The individual pulse waveform that w (t) is energy normalized, the duration is T _p; L is multipath number; a _land τ _lbe respectively attenuation coefficient and the retardation coefficient of channel; E is pulse energy.The signal received by after energy integral, with sampling interval T _benergy signal is sampled.Make N _ffor the number of the frame in each symbol, the energy sequence of each symbol is

$z\left[n\right]=\underset{j=1}{\overset{N_f}{\mathrm{\Σ}}}{\∫}_{(j-1)T_f+(c_j+n-1)T_b}^{(j-1)T_f+(c_j+n)T_b}{\left|r\left(t\right)\right|}^2\mathrm{dt}$

Step 2: resolving of decision threshold

The energy sampled sequence of signal is sorted, under the constraint of certain constant false alarm rate, calculates an energy threshold by iteration optimization, reach path for head to received signal and adjudicate.

After signal completes frame synchronization, the time of advent in first footpath is uniformly distributed in a frame, considers the crosstalk of interframe, will observe the frame length being set to 1.5 times.Make T _obfor observing interval, then sequence z [n] comprises individual sampled energy block.Comprising the stack power block of pure noise energy block and signal and noise.

In K energy sampled block of energy sequence, comprise pure noise energy block and noise signal stack power block two kinds.Wherein, the value of pure noise energy block obeys the distribution of center card side, and average is M δ ², variance is 2M δ ⁴, the degree of freedom is M=2BT _b+ 1, B is the bandwidth of signal; Noise signal stack power block then obeys the distribution of non-central card side, and average is M δ ²+ E _n, variance is 2M δ ⁴+ 4 δ ²e _n, E _nfor the signal energy of this sampling block.Iteration threshold algorithm adopts Niemann-Pearson came hypothesis testing, and under constant false alarm rate constraint, iteration goes out threshold value, for the noise energy block of center card side distribution, and false alarm rate P _faas follows with the relation of threshold value ξ:

$P_{\mathrm{fa}}=P\left(z\right[n]>\mathrm{\ξ})=\exp (-\frac{\mathrm{\ξ}}{{2\mathrm{\δ}}^2})\underset{k=0}{\overset{M/2-1}{\mathrm{\Σ}}}\frac{1}{k!}{\left(\frac{\mathrm{\ξ}}{{2\mathrm{\δ}}^2}\right)}^k$

Fig. 2 is the flow chart of the iteration threshold selection algorithm in the present invention.This algorithm key be that noise energy block is being got rid of successively, complete the renewal of threshold value in each iteration, thus detect a part of noise energy block, when all noise energy blocks are all detected, the performance of thresholding reaches the estimated accuracy time of advent of algorithm optimum.

Step 3: the time of advent of signal is estimated

Utilize the decision threshold obtained in step 2 from energy sampled sequence, detect that the head of signal reaches path, the head of signal reaches the estimated value being time of arrival (toa) the time of advent in path.Under the condition of low signal-to-noise ratio, the time of advent completed under thresholding failure conditions by predetermined compensation policy is estimated.

After calculating optimum gate limit value, the time of advent that head reaches place, path energy block can be expressed as

${\mathrm{\τ}}_{\mathrm{ITS}}=\{\underset{K}{\min }\left(k\right|z\left[k\right]{>\mathrm{\ξ}}_{\mathrm{opt}})-0.5\}\×T_b$

Under the condition of low signal-to-noise ratio, the difference of conspicuousness is not had between the sampled value of pure noise energy block and noise signal stack power interblock, this may cause the threshold value of iteration cannot complete the situation of detection, energy sampled values all in the threshold value great-than search sequence that namely iteration goes out.When thresholding lost efficacy, this method have employed two kinds of compensation policies.The first is with the time of advent signal time of advent of the most intensity values in observation sequence, and the second is with the intermediate value of observation window signal time of advent.

Wherein, the basic procedure of the iteration threshold selection algorithm described in step 2 is as follows

1) observation sequence z [n] is sorted by ascending order, suppose that top n energy block is all noise block;

2) value and the false alarm probability P of this N number of energy block is utilized _facalculate corresponding threshold value ξ, the energy block being less than this threshold value so is all considered to noise block;

3) the noise block utilizing previous step to detect and false alarm probability P _farecalculate corresponding threshold value, the threshold value of renewal can detect again one group of new noise block;

4) constantly repeat this process until iterations reaches preset value, the threshold value now obtained is designated as ξ _opt.

3, advantage and effect: the present invention can detect that the head in the multipath signal received reaches path, this time of advent, method of estimation mainly possessed following advantage:

1) thresholding involved in the present invention resolves model because based on constant false alarm rate constraint, can obtain estimated accuracy good time of advent in different ultra-wideband channel patterns.

2) computational process involved in the present invention has carried out pre-sorting to energy sequence, has decreased the amount of calculation arriving time Estimate.

3) iteration threshold selection algorithm involved in the present invention has independence, does not rely on the acquisition of channel prior information, can be applied among actual ultra-wideband positioning system.

Accompanying drawing explanation

Fig. 1 is the system flow chart that the inventive method relates to

Fig. 2 is the iteration threshold selection algorithm flow chart in the present invention

Embodiment

See Fig. 1, the method for estimation time of advent of a kind of ultra-wideband positioning system of the present invention, the method concrete steps are as follows:

Step one: the energy sampled of Received signal strength

During the jumping received, impulse ultra-wideband signal can be expressed as

$r\left(t\right)=\underset{j=-\∞}{\overset{\∞}{\mathrm{\Σ}}}{d}_j{w}_{\mathrm{mp}}(t-{\mathrm{jT}}_f-c_j{T}_c-{\mathrm{\τ}}_{\mathrm{toa}})+n\left(t\right)$

Wherein, T _fand T _cbe respectively frame length and chip lengths; d _jfor the polarity of jth frame signal, τ _toafor the time of advent of signal; N (t) is white Gaussian noise, and average is zero, and variance is δ ², bilateral power spectral density is N ₀/ 2; c _jbeing the time-hopping code in order to prevent the signal conflict between different user from distributing, which determining pulse chip position in a frame, the time-hopping sequence that a kth user is assigned to meets ; w _mpt () is the multi-path pulse waveform received, can be expressed as

$w_{\mathrm{mp}}\left(t\right)=\sqrt{E}\underset{l=1}{\overset{L}{\mathrm{\Σ}}}{a}_{l}w(t-{\mathrm{\τ}}_l)$

The individual pulse waveform that w (t) is energy normalized, the duration is T _p; L is multipath number; a _land τ _lbe respectively attenuation coefficient and the retardation coefficient of channel; E is pulse energy.The signal received by after energy integral, with sampling interval T _benergy signal is sampled.Make N _ffor the number of the frame in each symbol, the energy sequence of each symbol is

$z\left[n\right]=\underset{j=1}{\overset{N_f}{\mathrm{\Σ}}}{\∫}_{(j-1)T_f+(c_j+n-1)T_b}^{(j-1)T_f+(c_j+n)T_b}{\left|r\left(t\right)\right|}^2\mathrm{dt}$

Step 2: resolving of decision threshold

After signal completes frame synchronization, the time of advent in first footpath is uniformly distributed in a frame, considers the crosstalk of interframe, will observe the frame length being set to 1.5 times.Make T _obfor observing interval, then sequence z [n] comprises individual sampled energy block.Comprising the stack power block of pure noise energy block and signal and noise.

In K energy sampled block of energy sequence, comprise pure noise energy block and noise signal stack power block two kinds.Wherein, the value of pure noise energy block obeys the distribution of center card side, and average is M δ ², variance is 2M δ ⁴, the degree of freedom is M=2BT _b+ 1, B is the bandwidth of signal; Noise signal stack power block then obeys the distribution of non-central card side, and average is M δ ²+ E _n, variance is 2M δ ⁴+ 4 δ ²e _n, E _nfor the signal energy of this sampling block.Iteration threshold algorithm adopts Niemann-Pearson came hypothesis testing, and under constant false alarm rate constraint, iteration goes out threshold value, for the noise energy block of center card side distribution, and false alarm rate P _faunder entering with the relation of threshold value ξ

$P_{\mathrm{fa}}=P\left(z\right[n]>\mathrm{\ξ})=\exp (-\frac{\mathrm{\ξ}}{{2\mathrm{\δ}}^2})\underset{k=0}{\overset{M/2-1}{\mathrm{\Σ}}}\frac{1}{k!}{\left(\frac{\mathrm{\ξ}}{{2\mathrm{\δ}}^2}\right)}^k$

Fig. 2 is the flow chart of the iteration threshold selection algorithm in the present invention.This algorithm key be that noise energy block is being got rid of successively, complete the renewal of threshold value in each iteration, thus detect a part of noise energy block, when all noise energy blocks are all detected, the performance of thresholding reaches the estimated accuracy time of advent of algorithm optimum.The basic procedure of this algorithm is as follows

1) observation sequence z [n] is sorted by ascending order, suppose that top n energy block is all noise block;

2) value and the false alarm probability P of this N number of energy block is utilized _facalculate corresponding threshold value ξ, the energy block being less than this threshold value so is all considered to noise block;

3) the noise block utilizing previous step to detect and false alarm probability P _farecalculate corresponding threshold value, the threshold value of renewal can detect again one group of new noise block;

4) constantly repeat this process until iterations reaches preset value, the threshold value now obtained is designated as ξ _opt.

Step 3: the time of advent of signal is estimated

After calculating optimum gate limit value, the time of advent that head reaches place, path energy block can be expressed as

${\mathrm{\τ}}_{\mathrm{ITS}}=\{\underset{K}{\min }\left(k\right|z\left[k\right]{>\mathrm{\ξ}}_{\mathrm{opt}})-0.5\}\×T_b$

Under the condition of low signal-to-noise ratio, the difference of conspicuousness is not had between the sampled value of pure noise energy block and noise signal stack power interblock, this may cause the threshold value of iteration cannot complete the situation of detection, energy sampled values all in the threshold value great-than search sequence that namely iteration goes out.When thresholding lost efficacy, this method have employed two kinds of compensation policies.The first is with the time of advent signal time of advent of the most intensity values in observation sequence, and the second is with the intermediate value of observation window signal time of advent.

Claims (2)

1. the method for estimation time of advent for ultra-wideband positioning system, is characterized in that: the method concrete steps are as follows:

Step one: the energy sampled of Received signal strength

Square law detection is carried out to the ultra-broadband signal after frame synchronization, obtains the energy signal of Received signal strength; Again this energy signal is sampled, obtain the energy sampled sequence of Received signal strength;

During the jumping received, impulse ultra-wideband signal is expressed as

$r\left(t\right)=\underset{j=-\∞}{\overset{\∞}{\mathrm{\Σ}}}{d}_j{w}_{\mathrm{mp}}(t-j{T}_f-c_j{T}_c-{\mathrm{\τ}}_{\mathrm{toa}})+n\left(t\right)$

Wherein, T _fand T _cbe respectively frame length and chip lengths; d _jfor the polarity of jth frame signal, τ _toafor the time of advent of signal; N (t) is white Gaussian noise, and average is zero, and variance is δ ², bilateral power spectral density is N ₀/ 2; One-sided power spectrum density is N ₀; c _jbeing the time-hopping code in order to prevent the signal conflict between different user from distributing, which determining pulse chip position in a frame, the time-hopping sequence that a kth user is assigned to meets n _cfor the number of chip in every frame signal, N _c=T _f/ T _cround; w _mpt () is the multi-path pulse waveform received, be expressed as

$w_{\mathrm{mp}}\left(t\right)=\sqrt{E}\underset{l=1}{\overset{L}{\mathrm{\Σ}}}{a}_{l}w(t-{\mathrm{\τ}}_l)$

The individual pulse waveform that w (t) is energy normalized, the duration is T _p; L is multipath number; a _land τ _lbe respectively attenuation coefficient and the retardation coefficient of channel; E is pulse energy, the signal received by after energy integral, with sampling interval T _benergy signal is sampled, makes N _ffor the number of the frame in each symbol, the energy sequence of each symbol is

$z\left[n\right]=\underset{j=1}{\overset{N_f}{\mathrm{\Σ}}}{\∫}_{(j-1)T_f+(c_j+n-1)T_b}^{(j-1)T_f+(c_j+n)T_b}{\left|r\left(t\right)\right|}^2\mathrm{dt}$

Step 2: resolving of decision threshold

The energy sampled sequence of signal is sorted, under constant false alarm rate constraint, calculates an energy threshold by iteration optimization, reach path for head to received signal and adjudicate;

After signal completes frame synchronization, the time of advent in first footpath is uniformly distributed in a frame, considers the crosstalk of interframe, will observe the frame length being set to 1.5 times; Make T _obfor observing interval, then sequence z [n] comprises individual sampled energy block; Comprising the stack power block of pure noise energy block and signal and noise;

In K energy sampled block of energy sequence, comprise pure noise energy block and noise signal stack power block two kinds; Wherein, the value of pure noise energy block obeys the distribution of center card side, and average is M δ ², variance is 2M δ ⁴, the degree of freedom is M=2BT _b+ 1, B is the bandwidth of signal; Noise signal stack power block then obeys the distribution of non-central card side, and average is M δ ²+ E _n, variance is 2M δ ⁴+ 4 δ ²e _n, E _nfor the signal energy of this sampling block; Iteration threshold algorithm adopts Niemann-Pearson came hypothesis testing, and under constant false alarm rate constraint, iteration goes out threshold value, for the noise energy block of center card side distribution, and false alarm rate P _faas follows with the relation of threshold value ξ:

$P_{\mathrm{fa}}=P\left(z\right[n]>\mathrm{\ξ})=\exp (-\frac{\mathrm{\ξ}}{{2\mathrm{\δ}}^2})\underset{k=0}{\overset{M/2-1}{\mathrm{\Σ}}}\frac{1}{k!}{\left(\frac{\mathrm{\ξ}}{{2\mathrm{\δ}}^2}\right)}^k$

This iteration threshold selection algorithm key be that noise energy block is being got rid of successively, complete the renewal of threshold value in each iteration, thus detect a part of noise energy block, when all noise energy blocks are all detected, the performance of thresholding reaches the estimated accuracy time of advent of algorithm optimum;

Step 3: the time of advent of signal is estimated

Utilize the decision threshold obtained in step 2 from energy sampled sequence, detect that the head of signal reaches path, the head of signal reaches the estimated value being time of arrival (toa) the time of advent in path, under the condition of low signal-to-noise ratio, the time of advent completed under thresholding failure conditions by predetermined compensation policy is estimated;

After calculating optimum gate limit value, the time of advent that head reaches place, path energy block is expressed as

${\mathrm{\τ}}_{\mathrm{ITS}}=\{\underset{K}{\min }\left(k\right|z\left[k\right]>{\mathrm{\ξ}}_{\mathrm{opt}})-0.5\}\×T_b$

Under the condition of low signal-to-noise ratio, the difference of conspicuousness is not had between the sampled value of pure noise energy block and noise signal stack power interblock, this can cause the threshold value of iteration cannot complete the situation of detection, energy sampled values all in the threshold value great-than search sequence that namely iteration goes out; When thresholding lost efficacy, adopt two kinds of compensation policies, the first is with the time of advent signal time of advent of the most intensity values in observation sequence, and the second is with the intermediate value of observation window signal time of advent.

2. the method for estimation time of advent of a kind of ultra-wideband positioning system according to claim 1, is characterized in that: the basic procedure of the iteration threshold selection algorithm described in step 2 is as follows:

1) observation sequence z [n] is sorted by ascending order, suppose that top n energy block is all noise block;

2) value and the false alarm probability P of this N number of energy block is utilized _facalculate corresponding threshold value ξ, the energy block being less than this threshold value so is all considered to noise block;

3) the noise block utilizing previous step to detect and false alarm probability P _farecalculate corresponding threshold value, the threshold value of renewal can detect again one group of new noise block;

4) constantly repeat this process until iterations reaches preset value, the threshold value now obtained is designated as ξ _opt.