CN106950528A - A kind of Wave arrival direction estimating method based on linear FM signal - Google Patents

Abstract

The invention discloses a kind of Wave arrival direction estimating method based on linear FM signal, the distance between two antennas of anchor node installation are d, and two antennas control to launch original frequency difference, chirp rate identical LFM signals simultaneously by same oscillator；The reception signal that label node is received to antenna carries out A/D samplings, carry out signal down-converted, NDFT conversion is carried out again, and the direct signal for reaching label node by two antennas under view distance environment can produce delay inequality, and the phase estimate AOA of signal is estimated after NDFT is converted；When there is multipath effect, there are different time delays in the direct signal for reaching label node by two antennas, obtain the direct signal delay inequality corresponding reception signal period, to carrying out the NDFT compared with multiple spot at Frequency point interested, find the frequency corresponding to peak-peak, estimation signal AOA.The present invention has the advantages such as anti-multipath effect, high dynamic, high accuracy, low hardware complexity.

Description

A kind of Wave arrival direction estimating method based on linear FM signal

Technical field

The invention belongs to wireless sensor technology and electronic communication system field, it is related to a kind of based on linear FM signal Wave arrival direction estimating method.

Background technology

Realize that the positioning to target is that RF alignment systems are often adopted by measuring the weighting vector (AOA) of RF (radio frequency) signal Working method.In recent years, with the popularization and the development of indoor positioning technologies of wireless network, location Based service (Location Based Service--LBS) is received more and more attention；For example in Emergency Assistance, health care, individual character Change the fields such as information transmission and show huge vigor.The purpose of wireless sensor network node positioning is exactly by receiving Radio wave carry out relevant parameter measurement, such as ripple up to the time (TOA), ripple up to the time difference (TDOA), direction of arrival (AOA) or letter Number field intensity (RSS) etc., the observation formula of these parameters is the function on target location, observation station by appropriate location algorithm come The position of signal source is estimated, the absolute or relative seat of each node in a plane or in space is provided further according to special algorithm Mark.

In the scheme of current relatively common positioning, the targeting scheme based on TOA/TDOA is commonplace, based on TOA/AOA's Targeting scheme also gradually causes the attention of people, and the targeting scheme based on AOA also has excessive quantifier elimination.Therefore for RF signals The technology of AOA estimations is also very important, with the development of RF technologies and continuing to bring out for novel measurement technique, angular surveying Technology also obtains continuous perfect with developing, and generally AOA estimation problems can also use modern using conventional method Method.Conventional method realizes that AOA estimates using mechanical type rotating directive antenna, is easily interfered, angular resolution ratio It is relatively low；Modernism is to use array antenna, and this method can overcome the shortcomings of conventional method, but need substantial amounts of antenna Therefore cost is costly.

In summary, there is following defect in existing AOA estimation techniques：Side based on mechanical type rotating directive antenna Method, resolution ratio is relatively low, and angle estimation precision is not high；Because multipath effect is very serious under complex environment so that the AOA of acquisition There is larger error, it is impossible to accurate to obtain the through through delay inequality of signal；Method based on array antenna is, it is necessary to more day Line, with higher cost and power consumption.

The content of the invention

There is estimation in order to overcome in existing AOA estimation techniques angle precision is high, multipath effect has larger error, day The higher defect of the more cost of line number amount, the invention discloses a kind of Wave arrival direction estimating method based on linear FM signal.

A kind of Wave arrival direction estimating method based on linear FM signal of the present invention, by installing two omni-directional antennas Anchor node and the label node composition of antenna is installed；

Specifically, the distance between two antennas that the anchor node is installed are d, anchor node has transmitting linear frequency modulation The function of LFM signals, and two antennas control transmitting original frequency difference, chirp rate phase simultaneously by same oscillator Same LFM signals.

A kind of Wave arrival direction estimating method based on linear FM signal proposed by the invention specifically includes following steps：

Step one：The anchor node is two antennas that mounting distance is d, and transmitting original frequency is different, chirp rate phase Same LFM signals：

Wherein：The time domain waveform of antenna A transmission signals is：

The instantaneous frequency of the signal is：

f_A(t)=f_at+βt, (2)

The time domain waveform of antenna B transmission signals is：

The instantaneous frequency of the signal is：

f_B(t)=f_b+βt, (4)

Wherein：f_a＞ f_b, and

f_aFor antenna A original frequency, f_bFor antenna B original frequency, t represents the time, and β represents chirp rate,Table Show antenna A initial phase,Represent antenna B initial phase；

Step 2：Because indoor range is smaller, it is believed that transmission signal is propagated in the case of near field, LOS (depending on Away from) under environment, the signal that label node is received to antenna carries out A/D samplings, while carrying out signal down-converted, sampling period For T_s, sample frequencyReceiving signal is：

s_RSS=| | r (t) | |²+w(t), (6)

Wherein：W (t) represents the white noise received, and r (t) carries out the later signal of down coversion for the signal that receives of antenna:

Wherein：P and Q representative antennas A and B multipath number,τ_Ai、τ_BiRepresentative antennas A and B is at i-th respectively Decay and time delay on path, wherein τ_Ai、τ_BiIt is antenna A and B LOS line of sights time delay (under LOS environment), indoor environment When in the presence of obvious multipath effect, step 5 is transferred to, step 3 is otherwise transferred to；

Step 3：Under LOS environment, there are different time delays, meeting in the direct signal for reaching label node by two antennas Produce delay inequality.Now, P=1, Q=1 in formula (7), now inside wushu (7) substitution formula (6), receive signal by arranging Obtain：

Wherein：

θ is that the direction of arrival AOA, d of signal are the spacing of A, B antenna, and c represents the light velocity；

Under LOS environment, it is to have (f to receive signal_a-f_b-β(τ_A1-τ_B1)) sine wave of frequency, step 8 is transferred to, it is right Receive signal and carry out NDFT conversion；Step 4 is transferred to again；

Step 4：The cycle that signal is obtained in NDFT changes later is carried out to signalObtain delay inequality τ_A1-τ_B1, Jin Ertong Direction of arrival AOA can be obtained by formula (9) by crossing geometrical relationship, shown in such as formula (10), be transferred to step 9；

Wherein：C represents the light velocity, and d is the distance between antenna A and B.

Step 5：When there is multipath effect, the direct signal for reaching label node by two antennas can be present not Same time delay, is 1 now in formula (7), when P is different with Q, wushu (7) is substituted into inside formula (6), the docking collection of letters number is arranged Obtain following expression：

Wherein：τ_k=τ_Ai-τ_BjThe delay inequality in antenna A the i-th paths and antenna B j-th strip path is represented, Π represents normal Number.

Receiving signal in a multi-path environment is and multidiameter delay difference τ_kRelated different cycles (f_a-f_b-βτ_k) signal it is folded Plus, because the amplitude of direct signal is stronger, the direct path delay inequality τ of usefully two antenna transmission signals is estimated AOA_k= τ_A1-τ_B1, corresponding signal frequency is：

Now the frequency range of key area is：(f_a-f_b± β d/c), it is assumed that f_a-f_b=Kf_s, f_sThe β d/c of ＞ 2, now The frequency range of key area is：

(K-1)f_s~(K+1) f_s, (13)

Most points are set in the range of these, step 8 is transferred to and carries out NDFT conversion, then be transferred to step 6；

Step 6：NDFT changes are carried out to signal, judge whether the signal peak after conversion exceedes defined threshold value, if Not less than defined threshold, then step 7 is transferred to, the corresponding frequency in position that otherwise signal peak occurs after search conversionAs f_RSSEstimate, substitute into formula (14) AOA can be obtained, be transferred to step 9；

Wherein：The spacing d ＜ c/2 β T of antenna_s。

Step 7：When the peak value of search is not above predetermined threshold value, then it is assumed that there is no line of sight signal under current environment, It is NLOS (non line of sight) environment, increases the scope of key area frequency, suboptimum AOA estimation is carried out into step 5；

Step 8：NDFT conversion is carried out to signal, FFT sampled point is equally distributed on unit circle, so base It can not go out realize the sampling at more small frequency interval in key area frequency in the method for FFT；NDFT transformed samples point is in list It is random distribution on circle of position, in key area, to be arranged on sampled point as much as possible, frequency interval as small as possible for frequency range NDFT changes, the first delay inequality for reaching footpath is estimated in multi-path environment with this, NDFT conversion is shown below：

Wherein：z₀,z₁,…,z_N-1It is any different N number of points chosen on unit circle；Formula (15) is write as the form of matrix (16) NDFT conversion quickly, is carried out by matrix form：

S=Z^Ts, (16)

Wherein：S=[S_NDFT(z₀),S_NDFT(z₁),…,S_NDFT(z_N-1)]^T, (17)

S=[s_RSS[0],s_RSS[1],…s_RSS[N-1]]^T, (18)

In (K-1) f_s~(K+1) f_sIn the range of sampling number be more than sampling number elsewhere；

Step 9：Kalman filtering processing is carried out to AOA, the influence that noise is caused to system is eliminated, further improves and estimate The precision of meter, output signal reaches direction AOA.

Compared with prior art, the invention has the advantages that：Using LFM signal systems, the big band by setting Width, can obtain high temporal resolution, so as to improve the precision of AOA estimations；In a multi-path environment, employ it is non-homogeneous from Dissipate Fourier transformation, to the point more than emphasis field frequency range sample, few points of sampling elsewhere, by observing frequency spectrum Fine mechanism estimates to improve the AOA under the resolution performance of frequency, more accurate progress multipath conditions；The present invention has simultaneously The advantages such as anti-multipath effect, high dynamic, high accuracy, low hardware complexity.

Brief description of the drawings

It is described herein to constitute the big part of the present invention as schemed to be used for proposing being explained further for the present invention, But do not constitute limitation of the invention.In the accompanying drawings：

Fig. 1 is the Wave arrival direction estimating method hardware composition frame chart proposed by the present invention based on linear FM signal

The Wave arrival direction estimating method transmission signal (solid line) based on linear FM signal that Fig. 2 is the present invention is believed with receiving The instantaneous frequency of number (dotted line)

Fig. 3 be the present invention the Wave arrival direction estimating method based on linear FM signal in NDFT (b figures) sampled points z put down Distribution on the unit circle of face

Fig. 4 is the Wave arrival direction estimating method implementing procedure figure based on linear FM signal of the present invention

Embodiment

Below in conjunction with the accompanying drawings and specific implementation case to the present invention carry out in further detail with complete explanation, state herein, Described specific implementation case is only used for explaining a kind of of the present invention, is not limitation of the invention.Further need exist for Illustrate, describe for convenience, it is not full content that part related to the present invention is illustrate only in accompanying drawing.

The present invention devises a kind of Wave arrival direction estimating method based on linear FM signal, is sent out simultaneously by two antennas LFM signals are penetrated, the receiving terminal docking collection of letters number is sampled, obtains line of sight transmission time delay difference by NDFT algorithms, pass through counterweight Important sampling is carried out in the range of point field frequency, the further influence for eliminating multipath effect to system improves AOA estimations Precision.The purpose of the present invention is to overcome existing AOA estimation techniques precision is not high in a multi-path environment, the larger shortcoming of error, if The a kind of of meter has antinoise signal multipath effect, high dynamic, high accuracy, the AOA methods of estimation of low hardware complexity.

In order to solve the problem of above-mentioned technical elements, present invention employs following technical scheme：

As shown in Figure 1, a kind of Wave arrival direction estimating method hardware composition frame chart based on linear FM signal, from figure As can be seen that the present invention is made up of the anchor node for installing two omni-directional antennas and the label node for installing an antenna；It is described Two antennas that anchor node is installed have fixed interval, the function with transmitting linear FM signal LFM, and two antennas Original frequency difference, chirp rate identical LFM signals are launched by an oscillator control simultaneously；The label node is to day The reception signal that line is received carries out A/D samplings, while carrying out signal down-converted, NDFT is carried out to pre-processing later data Conversion；Because indoor range is smaller, it is believed that transmission signal is propagated in the case of near field, as shown in Figure 2； Under LOS environment, the direct signal for reaching label node by two antennas can have different time delays, produce delay inequality, pass through It is estimated that carrying out the cycle of signal after NDFT conversion, further estimated shown in AOA such as formula (10) according to geometrical relationship；When When there is multipath effect, the direct signal for reaching label node by two antennas can have different time delays, while multipath Signal can also produce the time delay of complexity, therefore receive the superposition that signal is multi-signal, because the amplitude of direct signal is stronger, root The direct signal delay inequality corresponding reception signal period can be obtained according to the distance and sampling rate of transmitting antenna, to key area Frequency point at carry out compared with multiple spot NDFT, as shown in figure 3, find peak-peak corresponding to frequency, that is, direct signal Delay inequality, and then accurately estimate signal AOA, processing is filtered to AOA, the precision of estimation is further improved；When most When big peak value is not above predetermined thresholding, it is believed that current environment does not have direct signal, is non line of sight (NLOS) environment, need Expand the frequency range of key area, continue, using NDFT conversion, to obtain suboptimum AOA in a nlos environment.

Further, in present case, in order to allow the technical staff in association area to understand one kind of the present invention based on line Property FM signal Wave arrival direction estimating method, realized using chirp pulse signal AOA estimation function, while also carrying The overall flow figure of program implementation is gone out, as shown in Figure 4, the flow of this method comprises the following steps：

SS1. parameters are initialized：The original frequency f of the LFM signals of antenna A and B transmitting_a,f_bAnd chirp rate β, The repetition period T of LFM signals and two antennas apart from d, receive the sampling period T of signal_s, sample frequencySampling Points N so that f_a-f_b=Kf_s, f_sThe β d/c of ＞ 2, judge the threshold value Δ of the peak value after conversion；

SS2. anchor node periodic emission LFM signals, label node is adopted by A/D modules to the reception signal of antenna Sample, and down-conversion operation, while preserving the initial data that sampling is obtained；

SS3. the data to preservation carry out NDFT changes, and the frequency range to key area carries out important sampling, for example： (K-1)f_s~(K+1) f_s, by more points of being sampled at these Frequency points, improve frequency by observing the fine mechanism of frequency spectrum The resolution performance of rate；

SS4. search the peak value of the later data of conversion and record the Frequency point of peak value mappingJudge whether peak value surpasses Go out given threshold value：If beyond given threshold value Δ, it is believed that there is direct signal under the paths, being transferred to SS5, otherwise recognizing Direct signal is not present for current path, is obstructed path (NLOS), it is necessary to be missed through eliminating non line of sight by related mechanism Difference, while expanding key area frequency range, is transferred to SS3, finds the suboptimum AOA under current environment；

SS5. according to formula (12) f_RSS=f_a-f_b-β(τ_A1-τ_B1), when estimating line of sight of the signal after antenna A, B Prolong poor (τ_A1-τ_B1), according to formula (14)Estimate the direction of arrival AOA of signal；

SS6. to estimationThe filtering process such as Kalman, particle filter are carried out, the further noise that eliminates is caused to system Influence, improve estimation precision.

After above-mentioned step, the angle between label node and anchor node can be just obtained in real time in label node Degree, realizes and high-precision estimation is carried out to AOA.

The present invention is with low cost, can reduce the error that multipath effect is caused, and is measured by multiple actual experiment, this hair Bright angle measurement accuracy is higher, is particularly suitable for use in following scene：Parking lot traveling, pedestrian position automobile in market shopping etc. indoors The alignment system based on AOA in, can realize effective direction finding and positioning.

Finally it should be noted that：The above embodiments are merely illustrative of the technical solutions of the present invention rather than its limitations, institute The those of ordinary skill in category field should be understood；The either portion of techniques that can still be modified to the solution of the present invention is replaced Change, all non-scopes of patent protection for being used to limit the present invention, the scope of the present invention is still defined by claims；It is every to adopt Changed with the accompanying drawing content of the present invention similar structures that either specification is done, similarly should be included in the protection model of the patent In enclosing.

Claims (1)

1. a kind of Wave arrival direction estimating method based on linear FM signal, the device that this method is based on is complete by installing two The anchor node of directional antenna and the label node composition for installing an antenna；

Specifically, the distance between two antennas that the anchor node is installed are d, anchor node has transmitting linear frequency modulation LFM letters Number function, and two antennas control to launch simultaneously that original frequency is different, chirp rate identical by same oscillator LFM signals；

Characterized in that, this method specifically includes following steps：

Step one：The anchor node is two antennas that mounting distance is d, and transmitting original frequency is different, chirp rate identical LFM signals：

Wherein：The time domain waveform of antenna A transmission signals is：

The instantaneous frequency of the signal is：

f_A(t)=f_at+βt, (2)

The time domain waveform of antenna B transmission signals is：

The instantaneous frequency of the signal is：

f_B(t)=f_b+βt, (4)

Wherein：f_a＞ f_b, and

f_aFor antenna A original frequency, f_bFor antenna B original frequency, t represents the time, and β represents chirp rate,Represent antenna A initial phase,Represent antenna B initial phase；

Step 2：Because indoor range is smaller, it is believed that transmission signal is propagated in the case of near field, under LOS environment, The signal that label node is received to antenna carries out A/D samplings, while carrying out signal down-converted, the sampling period is T_s, sampling FrequencyReceiving signal is：

s_RSS=| | r (t) | |²+w(t), (6)

Wherein：W (t) represents the white noise received, and r (t) carries out the later signal of down coversion for the signal that receives of antenna:

Wherein：P and Q representative antennas A and B multipath number,τ_Ai、τ_BiRepresentative antennas A and B is in the i-th paths respectively On decay and time delay, wherein τ_Ai、τ_BiIt is antenna A and B LOS line of sight time delays, indoor environment has obvious multipath effect When, step 5 is transferred to, step 3 is otherwise transferred to；

Step 3：Under LOS environment, be present different time delays in the direct signal for reaching label node by two antennas, can produce Delay inequality；Now, P=1, Q=1 in formula (7), now inside wushu (7) substitution formula (6), receive signal and are obtained by arrangement：

Wherein：

θ is that the direction of arrival AOA, d of signal are the spacing of A, B antenna, and c represents the light velocity；

Under LOS environment, it is to have (f to receive signal_a-f_b-β(τ_A1-τ_B1)) sine wave of frequency, step 8 is transferred to, to receiving Signal carries out NDFT conversion；Step 4 is transferred to again；

Step 4：The cycle that signal is obtained in NDFT changes later is carried out to signalObtain delay inequality τ_A1-τ_B1, and then by several What relation can obtain direction of arrival AOA by formula (9), shown in such as formula (10), be transferred to step 9；

$\widehat{\mathrm{\θ}}=arccos\left(\frac{({\mathrm{\τ}}_{A1}-{\mathrm{\τ}}_{B1})\·c}{d}\right),---\left(10\right)$

Wherein：C represents the light velocity, and d is the distance between antenna A and B；

Step 5：When there is multipath effect, the direct signal for reaching label node by two antennas can exist different Time delay, is 1 now in formula (7), when P is different with Q, wushu (7) is substituted into inside formula (6), the docking collection of letters number is arranged Following expression：

Wherein：τ_k=τ_Ai-τ_BjThe delay inequality in antenna A the i-th paths and antenna B j-th strip path is represented, П represents constant；

Receiving signal in a multi-path environment is and multidiameter delay difference τ_kRelated different cycles (f_a-f_b-βτ_k) signal superposition, by It is stronger in the amplitude of direct signal, the direct path delay inequality τ of usefully two antenna transmission signals is estimated AOA_k=τ_A1- τ_B1, corresponding signal frequency is：

$f_{RSS}=f_a-f_b-\mathrm{\β}({\mathrm{\τ}}_{A1}-{\mathrm{\τ}}_{B1})=f_a-f_b-\mathrm{\β}\frac{dcos\mathrm{\θ}}{c},---\left(12\right)$

Now the frequency range of key area is：(f_a-f_b± β d/c), it is assumed that f_a-f_b=Kf_s, f_sThe β d/c of ＞ 2, now emphasis The frequency range in region is：

(K-1)f_s~(K+1) f_s, (13)

Most points are set in the range of these, step 8 is transferred to and carries out NDFT conversion, then be transferred to step 6；

Step 6：NDFT changes are carried out to signal, judge whether the signal peak after conversion exceedes defined threshold value, if do not surpassed Defined threshold is crossed, then is transferred to step 7, the corresponding frequency in position that otherwise signal peak occurs after search conversionIt is used as f_RSS's Estimate, AOA can be obtained by substituting into formula (14), be transferred to step 9；

$\widehat{\mathrm{\θ}}=arccos\left(\frac{(f_a-f_b-\hat{f})\·c}{\mathrm{\β}d}\right),---\left(14\right)$

Wherein：The spacing d ＜ c/2 β T of antenna_s；

Step 7：When the peak value of search is not above predetermined threshold value, then it is assumed that there is no line of sight signal under current environment, be NLOS (non line of sight) environment, increases the scope of key area frequency, and suboptimum AOA estimation is carried out into step 5；

Step 8：NDFT conversion is carried out to signal, FFT sampled point is equally distributed on unit circle, so based on FFT The method of conversion can not go out realize the sampling at more small frequency interval in key area frequency；NDFT transformed samples point is on unit circle It is random distribution, in key area, frequency range is arranged on sampled point as much as possible, frequency interval small NDFT as far as possible Change, the first delay inequality for reaching footpath in multi-path environment is estimated with this, and NDFT conversion is shown below：

$S_{NDFT}\left(z_k\right)=\underset{n=0}{\overset{N-1}{\Σ}}(s_{RSS}\[n\]\·z_k^{-n}),k=0,1,...,N-1,---\left(15\right)$

Wherein：z₀,z₁,…,z_N-1It is any different N number of points chosen on unit circle；Formula (15) is write as the form (16) of matrix, leads to Cross matrix form and quickly carry out NDFT conversion：

S=Z^Ts, (16)

Wherein：S=[S_NDFT(z₀),S_NDFT(z₁),…,S_NDFT(z_N-1)]^T, (17)

S=[s_RSS[0],s_RSS[1],…s_RSS[N-1]]^T, (18)

In (K-1) f_s~(K+1) f_sIn the range of sampling number be more than sampling number elsewhere；

Step 9：Kalman filtering processing is carried out to AOA, the influence that noise is caused to system is eliminated, further improves estimation Precision, output signal reaches direction AOA.