CN103873111B

CN103873111B - Narrowband Interference Suppression System and Method for Compressed Sensing Impulse UWB Receiver

Info

Publication number: CN103873111B
Application number: CN201410114209.0A
Authority: CN
Inventors: 王德强; 李智勇; 李国柱; 程金龙; 孟祥鹿
Original assignee: Shandong University
Current assignee: Shandong University
Priority date: 2014-03-25
Filing date: 2014-03-25
Publication date: 2015-10-21
Anticipated expiration: 2034-03-25
Also published as: CN103873111A

Abstract

The invention discloses a narrowband interference suppression system and method for a compressed sensing impulse ultra-wideband receiver. The method includes out-of-band noise filtering, pilot part narrowband interference estimation, pilot part narrowband interference suppression, signal correlation template reconstruction, There are six steps in the load part narrowband interference estimation and suppression, correlation demodulation. The method has two key points: one is to use compressed sensing technology to estimate the narrowband interference in each pilot symbol waveform one by one, and use a subtractor to subtract the corresponding narrowband interference observation sequence from the pilot observation sequence, which improves the signal correlation. Template estimation accuracy; the second is to use compressed sensing technology to estimate the narrowband interference in each load symbol waveform one by one, and use the subtractor to eliminate it from the load signal waveform, which improves the signal-to-interference ratio of the load signal. The method of the invention realizes the estimation and suppression of narrowband interference by using compressed sensing technology, greatly reduces the sampling rate, has good narrowband interference suppression effect, and improves the performance of the ultra-wideband receiver.

Description

Narrow-band interference suppression system and method of compressed sensing pulse ultra-wideband receiver

Technical Field

The invention relates to a narrow-band interference suppression system and method of a compressed sensing pulse ultra-wideband receiver, belonging to the field of broadband wireless communication.

Background

The nyquist sampling theorem states that: the original signal is recovered from the discrete signal without distortion, the sampling rate of which is at least 2 times the bandwidth of the signal. Compressed Sensing (CS) is an emerging signal processing framework. Unlike the conventional nyquist sampling theorem, the compressed sensing theory simultaneously compresses signals at a sampling rate much lower than the nyquist sampling theorem. The theory states that: for any compressible or sparse signal, the signal is projected onto a low-dimensional space through an observation matrix, and then the original signal is reconstructed or approximated through a series of reconstruction algorithms. Corresponding reconstruction algorithms include a basis Pursuit Algorithm (BP: Basic Pursuit Algorithm), a Matching Pursuit Algorithm (MP: Matching Pursuit Algorithm), an Orthogonal Matching Pursuit Algorithm (OMP: Orthogonal Matching Pursuit Algorithm), a subspace Pursuit Algorithm (SP: subspace Pursuit Algorithm), and the like.

The pulse ultra wide band (IR UWB) system is a new short-distance Wireless communication technology with high speed, low cost, low power consumption and good confidentiality, can be used for short-distance Wireless data networks such as Wireless Personal Area Networks (WPAN) and Wireless Body Area Networks (WBAN), and can also be used for systems such as radar ranging and radar imaging. Unlike conventional wireless communication technologies, impulse ultra-wideband uses a spectrum overlapping method to share currently used spectrum resources, i.e., it is coexisting with existing narrowband wireless systems. Although the Federal Communications Commission (FCC) limits the radiated power of an ultra-wideband system to ensure proper operation of an existing narrowband wireless communication system, other narrowband wireless communication systems inevitably interfere with an impulse ultra-wideband system because of their high radiated power relative to the ultra-wideband system. These interferences appear as a plurality of narrowband interferences within the ultra-wideband bandwidth. To ensure reliable communication of ultra-wideband systems, it is necessary to suppress interference of other communication systems. In one aspect, the classical narrowband interference suppression technique is a notch interference suppression technique based on the nyquist sampling theorem. The method realizes the estimation and the suppression of the interference in the frequency domain through FFT transformation, and then the estimation and the suppression are changed back to the time domain through IFFT so as to complete the suppression process of the whole interference. However, the time-frequency transform based on the FFT algorithm needs to sample the radio frequency interference signal at the nyquist rate, and at the same time, due to the high bandwidth (Ghz) characteristic of the impulse ultra-wideband system, an extremely high sampling rate is required. This undoubtedly increases the difficulty and cost of system hardware implementation. On the other hand, in recent years, only one article (skillful, Zhao sparkling, Zhou Chunhui, Wang Jing, Anjian ping) "pulse ultra wide band system narrow band interference estimation algorithm based on compressed sensing", instrument and meter report, volume 32, 3 rd of 3.2011) appears in domestic and foreign countries, and the article focuses on improvement of the narrow band interference OMP algorithm, and only introduces a short narrow band interference estimation process, but does not introduce the whole narrow band interference estimation and suppression process systematically.

Disclosure of Invention

The invention aims to solve the problems and provides a narrow-band interference suppression system and a narrow-band interference suppression method of a compressed sensing impulse ultra-wideband receiver.

In order to achieve the above object, the method comprises the steps of:

a narrow-band interference suppression system for a compressed sensing impulse ultra-wideband receiver, comprising

The broadband filter module filters out-of-band noise in the pulse ultra-wideband signal by utilizing a broadband filter;

a compressed sensing narrow-band interference estimation module, which obtains a narrow-band interference template from a pilot frequency part in the pulse ultra-wideband signal by using a compressed sensing theory;

the compressed sensing ultra-wideband correlation template estimation module eliminates the influence of narrow-band interference in pilot frequency by using a narrow-band interference template generated by the compressed sensing narrow-band interference estimation module, and then obtains an ultra-wideband signal correlation template by using a compressed sensing theory;

a signal delay module: properly delaying a load signal and sending the load signal to a load narrowband interference suppression module;

load observation storage module: temporarily storing the load observation sequences obtained by the observation module, and sequentially sending the observation sequences of each load signal to a narrow-band interference estimation template;

load narrowband interference suppression module: subtracting the narrow-band interference contained in the load signal estimated by the narrow-band interference estimation template from each load signal waveform by using a subtracter, and sending the load signal with the narrow-band interference suppressed to a relevant demodulation module;

and a relevant demodulation module: and performing correlation demodulation on the load signal output by the load narrowband interference suppression module after the narrowband interference is suppressed by using a correlator and a decision device by taking the signal template generated by the signal correlation template generation module as a correlation template.

The wideband filter module includes:

an ultra-wideband antenna module: receiving pulse ultra-wideband signals from a wireless channel and sending the signals to a receiving filter module;

a receiving filter module: and filtering out-of-band noise and interference on the pulse ultra-wideband signal.

The compressed sensing narrowband interference estimation module comprises:

an observation module: respectively observing the pilot frequency symbol waveform and the load symbol waveform, sending the pilot frequency observation sequence to a pilot frequency observation storage module, and sending the load observation sequence to a load observation storage module;

pilot frequency observation storage module: temporarily storing the observation sequence of each pilot frequency symbol waveform and sending the observation sequence to a narrow-band interference estimation module and a subtracter module;

a narrowband interference estimation module: and estimating narrow-band interference by utilizing an OMP algorithm, sending the narrow-band interference contained in the estimated pilot frequency symbol waveform to a pilot frequency narrow-band interference observation module, and sending the narrow-band interference contained in the estimated load signal waveform to a load narrow-band interference suppression module.

The compressed sensing ultra-wideband correlation template estimation module comprises:

pilot frequency narrowband interference observation module: observing the pilot frequency symbol narrowband interference estimated by the narrowband interference estimation module respectively, and sending a narrowband interference observation sequence obtained by observation to a narrowband interference observation storage module;

the narrow-band interference observation storage module: temporarily storing the narrow-band interference observation sequence and sending the sequence to a subtracter module;

a subtractor module: subtracting the corresponding narrow-band interference observation sequence from each pilot frequency observation sequence by using a subtracter to obtain a pilot frequency observation sequence for inhibiting narrow-band interference;

a vector averaging module: summing and averaging the pilot frequency observation sequence for inhibiting the narrow-band interference obtained in the subtracter module to obtain an average pilot frequency observation sequence so as to reduce the influence of additive white Gaussian noise and send the average pilot frequency observation sequence to a signal correlation template generation module;

a signal dependent template generation module: and reconstructing a signal correlation template by using an OMP algorithm according to the average pilot frequency observation sequence obtained by the vector averaging module, and sending the signal correlation template to a correlation demodulation module.

The narrow-band interference suppression method of the pulse ultra-wideband receiver based on compressed sensing comprises the following steps:

step (1): after the transceiver establishes communication, the receiving end receives the pulse ultra-wideband signal from the wireless channel, and the pulse ultra-wideband signal comprises two parts: the pilot frequency part and the data load part are used for filtering out-of-band noise superposed on the pulse ultra-wideband signal;

step (2): based on a compressed sensing theory, a receiving end observes each pilot frequency symbol waveform in the pilot frequency part of the pulse ultra-wideband signal in the step (1) by using a pre-generated Gaussian random matrix as an observation matrix to obtain a plurality of pilot frequency observation sequences, and then estimates narrowband interference contained in each pilot frequency symbol waveform;

and (3):

observing the narrow-band interference estimated in the step (2) by using the observation matrix in the step (2) respectively to obtain a plurality of narrow-band interference observation sequences;

subtracting the corresponding narrow-band interference observation sequences from the plurality of pilot frequency observation sequences obtained in the step (2) to obtain a plurality of pilot frequency observation sequences with narrow-band interference removed;

and (4): averaging the pilot frequency observation sequences obtained in the step (3) for removing the narrow-band interference to obtain an average pilot frequency observation sequence so as to reduce the influence of additive white Gaussian noise, and reconstructing a signal correlation template according to the obtained average pilot frequency observation sequence; the signal correlation template refers to: a local reference signal for use by a correlator in a correlation receiver;

and (5): observing each data symbol waveform in the data load part of the pulse ultra-wideband signal in the step (1) by using the observation matrix in the step (2) to obtain a data load observation sequence, then estimating narrowband interference contained in each data symbol waveform to obtain a narrowband interference waveform, and then subtracting the narrowband interference waveform from the load signal waveform by using a subtracter to eliminate the influence of the narrowband interference;

and (6): and (4) carrying out correlation demodulation on the load signal waveform obtained in the step (5) after the narrow-band interference is eliminated by using the signal correlation template generated in the step (4) through a correlation receiver to obtain a data symbol sequence.

The specific steps of the step (1) are as follows:

the ultra-wideband signal waveform obtained by the receiving end is assumed as follows:

wherein N is_pIndicates the number of pilot symbols, all pilot symbols are 1, N_sIndicates the number of load information symbols, T_sIs a symbol period, b_jIs a binary data modulation symbol and b_j∈{-1,1}，f_nb(t) and n (t) represent narrow-band interference and white gaussian noise respectively,t∈((i-1)T_s,iT_s),i＝1,2,...,N_prepresents the ith interfered pilot symbol waveform,t∈((j-1)T_s+N_pT_s,jT+N_pT_s),j＝1,2,...,N_srepresenting the jth disturbed load signal waveform.

The specific steps of the step (2) are as follows:

the receiving end utilizes a pre-generated Gaussian random observation matrix based on a compressed sensing theoryObserving each pilot frequency symbol waveform in the pilot frequency part of the pulse ultra-wideband signal in the step (1) to obtain N_pObservation sequence of pilot symbol waveformWherein y is_i,i＝1,2,…,N_pIs an M × 1 dimensional column vector, represents the observation sequence of the ith pilot symbol waveform, and its kth element is:

whereinRepresenting the ith interfered pilot signal waveform,for the k-th observed waveform in the observation matrix, N_pThe number of pilot frequencies is, and M is the number of observed waveforms;

when the narrowband interference contained in each pilot frequency symbol waveform is reconstructed, an Inverse Discrete Fourier Transform (IDFT) matrix is adopted as a sparse dictionary of the narrowband interference:

wherein,n is eachThe number of sampling points in the nyquist sampling of the pilot symbol waveform indicates the matrix conjugate transpose operation. Narrow-band interference in each pilot symbol period since the narrow-band interference is sparse in the frequency domaint∈((i-1)T_s,iT_s),i＝1,2,...,N_pExpressed as:

t∈((i-1)T_s,iT_s),i＝1,2,...,N_p。

therein, Ψ_nbIn order to have a narrow-band interference sparse dictionary,the projection coefficient vector of the narrow-band interference in the ith pilot symbol waveform is obtained, and because the narrow-band interference is sparse,only a few of the dominant elements.

From Gaussian random observation matrix phi and narrow-band interference sparse dictionary psi_nbThe reconstruction matrix can be obtained:

V_nb＝ΦΨ_nb

where phi is the observation matrix, Ψ_nbIs a narrow-band interference sparse dictionary.

From the reconstruction matrix V_nbAnd the obtained N_pObservation sequence ofThen estimating the narrow-band interference contained in each pilot frequency symbol waveform by using OMP algorithmWhereint∈((i-1)T_s,iT_s),i＝1,2,...,N_pRepresenting an estimate of the narrowband interference contained in the ith pilot symbol waveform.

The OMP algorithm comprises the following steps: the maximum iteration number of the OMP algorithm is assumed to be limited to K;

the first step is as follows: initialization: residual value r₀Y, index set(empty set), incremental matrixThe iteration time t is 1;

the second step is that: finding the residual value r_t-1And V_nbColumn V in_jColumn corresponding to inner product maximum

The third step: updating index setsUpdating delta matrices

The fourth step: obtained by least squares

The fifth step: updating residual values

And a sixth step: judging whether t is more than or equal to K, and if so, stopping iteration; if not, executing the second step.

Respectively using each pilot observation value y to the residual value y in the OMP algorithm_i,i＝1,2,...,N_pAlternatively, a projection coefficient vector of the narrow-band interference contained in each pilot symbol waveform is obtainedi＝1,2,...,N_pIs estimated byi＝1,2,...,N_pThus, the estimated narrowband interference is:

t∈((i-1)T_s,iT_s),i＝1,2,...,N_p。

therein, Ψ_nbIs a narrow strip of stemsThe disturbing-sparse dictionary is used for disturbing sparse dictionary,projecting coefficient vectors for narrow-band interference contained in the ith pilot symbol waveformi＝1,2,...,N_pIs estimated.

The specific steps of the step (3) are as follows:

using the Gaussian random observation matrix in the step (2)For the narrowband interference estimated in step (2)Respectively observing to obtain corresponding narrow-band interference observation sequencesWhereini＝1,2,...,N_pIs an Mx 1-dimensional column vector and represents the estimation of the narrow-band interference contained in the ith pilot symbol waveformObserving to obtain an observation sequence, wherein the kth element is as follows:

whereinRepresents an estimate of the narrowband interference contained in the ith pilot symbol waveform,for the k-th observed waveform in the observation matrix, N_pThe number of pilot frequencies is, and M is the number of observed waveforms;

a plurality of pilot frequency observation sequences obtained from the step (2)Subtracting the corresponding narrowband interference observation sequenceObtaining a pilot frequency observation sequence for removing the narrow-band interference;

wherein,represents the ithThe pilot symbol waveform suppresses the observation sequence after the narrowband interference.

The specific steps of the step (4) are as follows:

a plurality of pilot frequency observation sequences obtained in the step (3) after narrow-band interference suppression are obtainedAveraging to obtain average pilot observation sequenceTo reduce the effect of additive white Gaussian noise, i.e.

<math> <mrow> <mover> <mi>y</mi> <mo>&OverBar;</mo> </mover> <mo>=</mo> <mfrac> <mn>1</mn> <msub> <mi>N</mi> <mi>p</mi> </msub> </mfrac> <munderover> <mi>Σ</mi> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <msub> <mi>N</mi> <mi>p</mi> </msub> </munderover> <msub> <mi>y</mi> <mrow> <mi>i</mi> <mo>-</mo> <mi>f</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mn>1</mn> <msub> <mi>N</mi> <mi>p</mi> </msub> </mfrac> <munderover> <mi>Σ</mi> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <msub> <mi>N</mi> <mi>p</mi> </msub> </munderover> <mrow> <mo>(</mo> <msub> <mi>y</mi> <mi>i</mi> </msub> <mo>-</mo> <msubsup> <mi>y</mi> <mi>f</mi> <mi>i</mi> </msubsup> <mo>)</mo> </mrow> </mrow> </math>

Wherein, y_i-f,i＝1,2,...,N_pAnd showing the observation sequence after the ith pilot frequency waveform suppresses the narrow-band interference.

For signal correlation template reconstruction, the sparse dictionary is used as a feature vector sparse dictionary psi_gThe generation process is as follows:

the pulsed ultra-wideband signal g (t) received in step (1) is essentially a random process due to the time-varying nature of the pulsed ultra-wideband channel. Thus, its covariance function R (t- τ) is obtained over a large number of channel samples, i.e. it is

R(t-τ)＝E[g(t)g(τ+t)]

Let λ be₁＞λ₂＞λ₃＞...＞λ_NAnd representing the characteristic value of a Fredholm integral operator, wherein N is the number of sampling points when each pilot symbol waveform is subjected to Nyquist sampling. For R (t- τ) then:

<math> <mrow> <mo>&Integral;</mo> <mi>R</mi> <mrow> <mo>(</mo> <mi>t</mi> <mo>-</mo> <mi>τ</mi> <mo>)</mo> </mrow> <msub> <mi>u</mi> <mi>j</mi> </msub> <mrow> <mo>(</mo> <mi>τ</mi> <mo>)</mo> </mrow> <mi>dτ</mi> <mo>=</mo> <msub> <mi>λ</mi> <mi>j</mi> </msub> <msub> <mi>u</mi> <mi>j</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </mrow> </math>

wherein u is_j(t) is lambda_jCorresponding feature vector, and { u_j(t) is a complete set of orthogonal basis functions satisfying:

<math> <mrow> <mo>&Integral;</mo> <msub> <mi>u</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <msub> <mi>u</mi> <mi>j</mi> </msub> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>=</mo> <mtable> <mtr> <mtd> <mrow> <mfenced open='{' close=''> <mtable> <mtr> <mtd> <mn>0</mn> </mtd> <mtd> <mi>i</mi> <mo>&NotEqual;</mo> <mi>j</mi> </mtd> </mtr> <mtr> <mtd> <mn>1</mn> </mtd> <mtd> <mi>i</mi> <mo>=</mo> <mi>j</mi> </mtd> </mtr> </mtable> </mfenced> </mrow> </mtd> </mtr> </mtable> <mo>,</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>=</mo> <mn>1,2,3</mn> <mo>,</mo> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mo>,</mo> <mi>N</mi> </mrow> </math>

thus, [ u ]₁(t),u₂(t),u₃(t),...,u_N(t)]And forming a group of orthogonal bases of g (t), wherein the group of orthogonal bases are the feature vector sparse dictionary.

By the Gaussian random observation matrix in the step (2)And feature vector sparse dictionary Ψ_gThe correlation template reconstruction matrix is obtained, i.e.

V_g＝ΦΨ_g

Combining the average pilot frequency observation sequence obtained in the stepThe OMP algorithm can be used for reconstructing to obtain the signal correlation templateThe OMP algorithm process is as described in step (2).

The specific steps of the step (5) are as follows:

firstly, the Gaussian random observation matrix in the step (2) is utilizedObserving each load symbol waveform in the pulse ultra-wideband signal load part in the step (1) to obtain an observation sequence of each load symbol waveformWhereinIs an M multiplied by 1 dimension column vector, represents the observation sequence of the jth load symbol waveform, and the kth element is:

whereinRepresenting the jth disturbed load signal waveform,for the k-th observed waveform in the observation matrix, N_sThe number of load symbols is, and M is the number of observed waveforms;

then, according to the estimation method of the narrow-band interference contained in the pilot signal in the step (2), the estimation of the narrow-band interference contained in each load symbol waveform is obtained by utilizing an OMP algorithmRepresents an estimate of the narrow-band interference contained in the jth load symbol waveform, andexpressed as:

t∈((j-1)T_s+N_pT_s,jT_s+N_pT_s),j＝1,2,...,N_s。

therein, Ψ_nbIn order to have a narrow-band interference sparse dictionary,projection coefficient vector of narrow-band interference in jth load symbol waveform obtained by using OMP algorithmj＝1,2,...,N_sIs estimated.

Finally, the subtracter is used to eliminate the influence of narrow-band interference from the load signal waveform to obtain the load signal waveform s for suppressing the narrow-band interference_load(t)：

t∈((j-1)T_s+N_pT_s,jT_s+N_pT_s),j＝1,2,...,N_s

The specific steps of the step (6) are as follows:

the load signal waveform s obtained in the step (5) after the narrow-band interference is eliminated_load(t) using the signal correlation template generated in step (4)Symbol sequence obtained by coherent demodulation through coherent receiver j＝1,2,...,N_sSatisfy the requirement of

Wherein,representing the jth disturbed load signal waveform,representing an estimate of the narrowband interference contained in the jth load symbol waveform,as a related template, N_sIndicating the number of load information symbols.

The invention has the beneficial effects that:

the method of the invention realizes the estimation and the inhibition of the narrow-band interference through the compressed sensing technology, and has the following beneficial effects:

1. the sampling rate is greatly reduced through compressed sensing, the technical bottleneck faced by the traditional sampling theory is broken through, and the cost and the power consumption of a receiver are favorably controlled;

2. the compressed sensing and reconstruction algorithm is used for adaptively detecting and inhibiting random narrow-band interference, so that the estimation precision of a relevant template of a receiver is improved;

3. the detection reliability of the receiver is improved by accurately estimating and suppressing the high-power narrow-band interference superposed in the ultra-wide band signal.

Drawings

FIG. 1 is a block diagram of an embodiment of the method of the present invention;

FIG. 2 is a graph illustrating the variation of the bit error rate according to an embodiment of the present invention;

the system comprises an ultra-wideband antenna module 1, a receiving filter module 2, an observation module 3, a pilot frequency observation storage module 4, a narrow-band interference estimation module 5, a pilot frequency narrow-band interference observation module 6, a narrow-band interference observation storage module 7, a narrow-band interference observation storage module 8, a subtracter module 9, a vector averaging module 9, a signal correlation template generation module 10, a signal delay module 11, a load observation storage module 12, a load narrow-band interference suppression module 13 and a correlation demodulation module 14.

Detailed Description

The invention is further described with reference to the following figures and examples.

As shown in fig. 1, a block diagram of an embodiment of the inventive method, the modules function as follows:

ultra-wideband antenna module 1: receiving pulse ultra-wideband signals from a wireless channel and sending the signals to a receiving filter module 2;

reception filter module 2: filtering out-of-band noise and interference from the pulse ultra-wideband signal;

and an observation module 3: respectively observing the pilot frequency symbol waveform and the load symbol waveform, sending the pilot frequency observation sequence to the pilot frequency observation storage module 4, and sending the load observation sequence to the load observation storage module 12;

the pilot observation storage module 4: temporarily storing the observation sequence of each pilot frequency symbol waveform and sending the observation sequence to a narrow-band interference estimation module 5 and a subtracter module 8;

narrowband interference estimation module 5: estimating narrow-band interference by using an OMP algorithm, sending the narrow-band interference contained in the estimated pilot symbol waveform to a pilot narrow-band interference observation module 6, and sending the narrow-band interference contained in the estimated load signal waveform to a load narrow-band interference suppression module 13;

pilot frequency narrowband interference observation module 6: observing the pilot frequency symbol narrowband interference estimated by the narrowband interference estimation module 5 respectively, and sending the observed narrowband interference observation sequence to a narrowband interference observation storage module 7;

narrow-band interference observation storage module 7: temporarily storing the narrow-band interference observation sequence and sending the sequence to a subtracter module 8;

a subtractor module 8: subtracting the corresponding narrow-band interference observation sequence from each pilot frequency observation sequence by using a subtracter to obtain a pilot frequency observation sequence for inhibiting narrow-band interference;

vector averaging module 9: summing and averaging the pilot frequency observation sequence for suppressing the narrow-band interference obtained in the subtractor module 8 to obtain an average pilot frequency observation sequence so as to reduce the influence of additive white gaussian noise, and sending the average pilot frequency observation sequence to a signal correlation template generation module 10;

signal dependent template generation module 10: according to the average pilot observation sequence obtained by the vector averaging module 9, reconstructing a signal correlation template by using an OMP algorithm, and sending the signal correlation template to a correlation demodulation module 14;

the signal delay module 11: the load signal is delayed properly and sent to a load narrowband interference suppression module 13;

load observation storage module 12: temporarily storing the load observation sequences obtained by the observation module 3, and sequentially sending the observation sequences of each load signal to a narrow-band interference estimation template 5;

load narrowband interference suppression module 13: the subtracter is used for subtracting the narrow-band interference contained in the load signal estimated by the narrow-band interference estimation template 5 from each load signal waveform, and the load signal after the narrow-band interference is suppressed is sent to the relevant demodulation module 14;

correlation demodulation module 14: the signal template generated by the signal correlation template generation module 10 is used as a correlation template, and a correlator and a decision device are used for performing correlation demodulation on the load signal which is output by the load narrowband interference suppression module 13 and subjected to narrowband interference suppression.

Simulation parameters of the implementation example of the method of the invention are as follows: simulation environment: matlab7.13; transmitter basic pulse shape: a second derivative pulse shape of Gaussian; and (3) channel model: ieee802.15.3a CM1 channel model; the UWB modulation mode is BPSK; observation matrix: a Gaussian random matrix; narrow-band interference sparse dictionary: an inverse discrete Fourier transform matrix; sparse dictionary of UWB receive signal: a feature vector model sparse dictionary; pulse ultra-wideband signal power: -30 dBm; the narrow-band interference power is: -10 dBm; the CS reconstruction algorithm: an OMP algorithm; number of channels: 100, respectively; the number of pilot symbols of each data packet is 10; the total payload data symbol number is 1000000.

step (2): based on a compressed sensing theory, a receiving end observes each pilot frequency symbol waveform in the pilot frequency part of the pulse ultra-wideband signal in the step (1) by using a pre-generated Gaussian random matrix as an observation matrix to obtain a plurality of pilot frequency observation sequences, and then estimates narrowband interference contained in each pilot frequency symbol waveform by using an OMP algorithm;

and (3):

and (4): averaging the pilot frequency observation sequences obtained in the step (3) to obtain an average pilot frequency observation sequence so as to reduce the influence of additive white Gaussian noise, and reconstructing a signal correlation template by utilizing an OMP algorithm according to the obtained average pilot frequency observation sequence;

and (5): observing each data symbol waveform in the data load part of the pulse ultra-wideband signal in the step (1) by using the observation matrix in the step (2) to obtain a data load observation sequence, then estimating narrow-band interference contained in each data symbol waveform by using an OMP algorithm to obtain a narrow-band interference waveform, and then subtracting the narrow-band interference waveform from the load signal waveform by using a subtracter to eliminate the influence of the narrow-band interference;

The specific steps of the step (1) are as follows:

wherein N is_pIndicates the number of pilot symbols, N_sIndicates the number of load information symbols, T_sIs a symbol period, b_jIs a binary data modulation symbol and b_j∈{-1,}1，f_nb(t) and n (t) represent narrow-band interference and white gaussian noise respectively,t∈((i-1)T_s,iT_s),i＝1,2,...,N_prepresents the ith interfered pilot symbol waveform,t∈((j-1)T_s+N_pT_s,jT+N_pT_s),j＝1,2,...,N_srepresenting the jth disturbed load signal waveform.

The specific steps of the step (2) are as follows:

wherein,n is the number of sampling points when Nyquist sampling is carried out on each pilot symbol waveform, and represents matrix conjugate transpose operation. Narrow-band interference in each pilot symbol period since the narrow-band interference is sparse in the frequency domaint∈((i-1)T_s,iT_s),i＝1,2,...,N_pExpressed as:

t∈((i-1)T_s,iT_s),i＝1,2,...,N_p。

therein, Ψ_nbIn order to have a narrow-band interference sparse dictionary,is as followsThe projection coefficient vector of the narrow-band interference in the i pilot symbol waveforms, and because the narrow-band interference is sparse,only a few of the dominant elements.

V_nb＝ΦΨ_nb

The third step: updating index setsUpdating delta matrices

The fourth step: obtained by least squares

The fifth step: updating residual values

t∈((i-1)T_s,iT_s),i＝1,2,...,N_p。

therein, Ψ_nbIn order to have a narrow-band interference sparse dictionary,projecting coefficient vectors for narrow-band interference contained in the ith pilot symbol waveformi＝1,2,...,N_pIs estimated.

The specific steps of the step (3) are as follows:

wherein,i＝1,2,...,N_pand showing the observation sequence after the ith pilot symbol waveform suppresses the narrow-band interference.

The specific steps of the step (4) are as follows:

R(t-τ)＝E[g(t)g(τ+t)]

V_g＝ΦΨ_g

The specific steps of the step (5) are as follows:

firstly, the Gaussian random observation matrix in the step (2) is utilizedObserving each load symbol waveform in the pulse ultra-wideband signal load part in the step (1) to obtain an observation sequence of each load symbol waveformWhereinj＝1,2,…,N_sIs an M multiplied by 1 dimension column vector, represents the observation sequence of the jth load symbol waveform, and the kth element is:

whereinRepresenting the jth disturbed load signal waveform,for observing matrixMiddle k-th observed waveform, N_sThe number of load symbols is, and M is the number of observed waveforms;

t∈((j-1)T_s+N_pT_s,jT_s+N_pT_s),j＝1,2,...,N_s。

t∈((j-1)T_s+N_pT_s,jT_s+N_pT_s),j＝1,2,...,N_s

The specific steps of the step (6) are as follows:

Wherein,j＝1,2,...,N_srepresenting the jth disturbed load signal waveform,representing an estimate of the narrowband interference contained in the jth load symbol waveform,as a related template, N_sIndicating the number of load information symbols.

Fig. 2 is a plot of bit error rate versus signal-to-noise ratio ("suppression") generated by a simulated embodiment of the method of the present invention, and also shows a performance curve without narrowband interference suppression ("no suppression"). Comparing the two curves, the method of the invention can effectively estimate and restrain the narrow-band interference existing in the pulse ultra-wideband signal, and has good error rate performance.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.

Claims

1. The narrow-band interference suppression system of the impulse ultra-wideband receiver of compressed sensing, it is characterized in that, comprises

The broadband filter module includes: ultra-wideband antenna module: receives the pulsed ultra-wideband signal from the wireless channel, and sends the signal to the receiving filter module; receiving filter module: filters out-of-band noise and interference for the pulsed ultra-wideband signal;

The compressed sensing narrowband interference estimation module includes: an observation module: observe the pilot symbol waveform and the load symbol waveform respectively, and send the pilot observation sequence to the pilot observation storage module, and send the load observation sequence to the load observation storage module; Pilot observation storage module: temporarily store the observation sequence of each pilot symbol waveform, and send it to the narrowband interference estimation module and the subtractor module; narrowband interference estimation module: use the OMP algorithm to estimate narrowband interference, and convert the estimated pilot symbol waveform The narrowband interference contained in the signal is sent to the pilot narrowband interference observation module, and the narrowband interference contained in the estimated load signal waveform is sent to the load narrowband interference suppression module;

Compressed sensing ultra-wideband correlation template estimation module, including: pilot narrowband interference observation module: respectively observe the pilot symbol narrowband interference estimated by the narrowband interference estimation module, and send the observed narrowband interference observation sequence to the narrowband interference observation storage module ; Narrowband interference observation storage module: temporarily store the narrowband interference observation sequence, and send it to the subtractor module; Subtractor module: utilize the subtractor to subtract its corresponding narrowband interference observation sequence from each pilot observation sequence to obtain suppression The pilot observation sequence of narrowband interference; the vector average module: the pilot observation sequence of suppressing narrowband interference obtained in the subtractor module is summed and averaged to obtain the average pilot observation sequence, so as to reduce the influence of additive Gaussian white noise, and send it to the signal correlation template generation module; the signal correlation template generation module: according to the average pilot observation sequence obtained by the vector average module, utilize the OMP algorithm to reconstruct the signal correlation template, and send it to the correlation demodulation module;

Signal delay module: properly delay the load signal and send it to the load narrowband interference suppression module;

Load observation storage module: temporarily store the load observation sequence obtained by the observation module, and send the observation sequence of each load signal to the narrowband interference estimation module in turn;

Load narrow-band interference suppression module: use a subtractor to subtract the narrow-band interference contained in the load signal estimated by the narrow-band interference estimation module from each load signal waveform, and send the load signal after the narrow-band interference suppression to the relevant demodulation module;

Correlation demodulation module: take the signal template generated by the signal correlation template generation module as the correlation template, and use the correlator and the decision device to correlate and demodulate the load signal after the narrowband interference suppression output by the load narrowband interference suppression module.

2. A method based on the narrow-band interference suppression system of the impulse ultra-wideband receiver of compressed sensing as claimed in claim 1, is characterized in that, comprises the steps:

Step (1): After the sending and receiving parties establish communication, the receiving end receives the pulsed ultra-wideband signal from the wireless channel. The pulsed ultra-wideband signal includes two parts: the pilot part and the data load part, and filters out the superimposed pulsed ultra-wideband signal. out-of-band noise on

Step (2): Based on the compressive sensing theory, the receiving end uses the pre-generated Gaussian random matrix as the observation matrix to observe the waveform of each pilot symbol in the pilot part of the pulse UWB signal in step (1), and obtain several pilot observations sequence, and then estimate the narrowband interference contained in each pilot symbol waveform;

Step (3):

Observing the narrowband interference estimated in step (2) respectively with the observation matrix in step (2), obtaining several narrowband interference observation sequences;

Subtract the corresponding narrowband interference observation sequence from some pilot observation sequences obtained in step (2), obtain some pilot observation sequences that remove narrowband interference;

Step (4): Averaging several pilot observation sequences obtained in step (3) to obtain the average pilot observation sequence to reduce the influence of additive Gaussian white noise, and according to the obtained average pilot observation sequence, repeat Construct a signal-related template;

Step (5): Using the observation matrix in step (2), observe the waveforms of each data symbol in the data load part of the pulse UWB signal in step (1), obtain the data load observation sequence, and then analyze the waveforms of each data symbol in the waveform of each data symbol Estimate the narrow-band interference contained in it to obtain the narrow-band interference waveform, and then use the subtractor to subtract the narrow-band interference waveform from the load signal waveform to eliminate the influence of narrow-band interference;

Step (6): For the load signal waveform obtained in step (5) after eliminating narrowband interference, use the signal correlation template generated in step (4) to perform correlation demodulation by a correlation receiver to obtain a data symbol sequence.

3. the method for claim 2 is characterized in that, the concrete steps of described step (1) are:

Assume that the UWB signal waveform obtained by the receiving end is:

\begin{matrix} s the s ((t t)) = = {Σ Σ}_{i i = = 11}^{{N N}_{p p}} g g ((t t - - {iT i}_{s the s})) + + {Σ Σ}_{j j = = 11}^{{N N}_{s the s}} {b b}_{j j} g g ((t t - - {N N}_{p p} {T T}_{s the s} - - {jT J}_{s the s})) + + n no ((t t)) + + {f f}_{n no b b} ((t t)) \\ = = {Σ Σ}_{i i = = 11}^{{N N}_{p p}} {\overset{\cdot &Center Dot;}{g g}}_{i i} ((t t)) + + {Σ Σ}_{j j = = 11}^{{N N}_{s the s}} {\overset{\cdot &Center Dot;}{g g}}_{i i} ((t t)) \end{matrix}

Among them, N _p represents the number of pilot symbols, all pilot symbols are 1, N _s represents the number of load information symbols, T _s is the symbol period, b _j is the binary data modulation symbol and b _j ∈ {-1,1 }, f _nb (t) and n(t) represent narrow-band interference and Gaussian white noise respectively, t∈((i-1)T _s , iT _s ), i=1,2,...,N _p represents the ith pilot symbol waveform with interference, t∈((j-1)T _s +N _p T _s ,jT+N _p T _s ),j=1,2,...,N _s represents the jth load signal waveform with interference.

4. the method for claim 2 is characterized in that, the concrete steps of described step (2) are:

Based on the compressed sensing theory, the receiving end uses the Gaussian random observation matrix generated in advance Observe the waveforms of each pilot symbol in the pilot part of the pulse UWB signal in step (1), and obtain the observation sequence of N _p pilot symbol waveforms Where y _i , i=1,2,...,N _p is an M×1-dimensional column vector, which represents the observation sequence of the i-th pilot symbol waveform, and its k-th element is:

in Indicates the i-th interfered pilot signal waveform, is the kth observed waveform in the observation matrix, N _p is the number of pilots, and M is the number of observed waveforms;

When reconstructing the narrowband interference contained in each pilot symbol waveform, the inverse discrete Fourier transform IDFT matrix is used as the sparse dictionary of narrowband interference:

{Ψ Ψ}_{n no b b} = = [[{ψ ψ}_{11} ((t t)),, {ψ ψ}_{22} ((t t)),, ... ...,, {ψ ψ}_{N N} ((t t))]] = = \frac{11}{\sqrt{N N}} {[\begin{matrix} {W W}^{00} & {W W}^{00} & {W W}^{00} & . . . . . . & {W W}^{00} \\ {W W}^{00} & {W W}^{11} & {W W}^{22} & . . . . . . & {W W}^{N N - - 11} \\ {W W}^{00} & {W W}^{22} & {W W}^{44} & . . . . . . & {W W}^{22 ((N N - - 11))} \\ . . & . . & . . & . . & . . \\ . . & . . & . . & . . & . . \\ . . & . . & . . & . . & . . \\ {W W}^{00} & {W W}^{N N - - 11} & {W W}^{22 ((N N - - 11))} & . . . . . . & {W W}^{((N N - - 11)) ((N N - - 11))} \end{matrix}]}^{* *}

in, N is the number of sampling points when Nyquist sampling is performed on each pilot symbol waveform, and * represents the matrix conjugate transposition operation; since the narrowband interference is sparse in the frequency domain, the narrowband interference in each pilot symbol period t∈((i-1)T _s , iT _s ), i=1,2,...,N _p is expressed as:

{f f}_{n no b b}^{i i} ((t t)) = = {Σ Σ}_{k k = = 11}^{N N} {ψ ψ}_{k k} ((t t)) {θ θ}_{k k} = = {Ψ Ψ}_{n no b b} {θ θ}_{n no b b}^{i i},, t t &Element; &Element; ((((i i - - 11)) {T T}_{s the s},, {iT i}_{s the s})),, i i = = 11,, 22,, ... ...,, {N N}_{p p};;

Among them, Ψ _nb is a narrow-band interference sparse dictionary, is the projection coefficient vector of the narrowband interference in the i-th pilot symbol waveform, and since the narrowband interference is sparse, There are only a few dominant elements in

The reconstruction matrix is obtained from the Gaussian random observation matrix Φ and the narrow-band interference sparse dictionary Ψ _nb :

V _nb ＝ΦΨ _nb

Among them, Φ is the observation matrix, and Ψ _nb is the narrow-band interference sparse dictionary;

According to the reconstruction matrix V _nb and the obtained N _p observation sequences Then the narrowband interference contained in each pilot symbol waveform is estimated by OMP algorithm estimation in t∈((i-1)T _s , iT _s ), i=1, 2,..., N _p represents the estimation of the narrowband interference contained in the ith pilot symbol waveform.

5. the method for claim 2 is characterized in that, the concrete steps of described step (3) are:

Use the Gaussian random observation matrix in step (2) For the narrowband interference estimated in step (2) Observe separately to obtain the corresponding narrowband interference observation sequence in i=1,2,...,N _p is an M×1-dimensional column vector, which represents the estimation of the narrowband interference contained in the i-th pilot symbol waveform Perform observations to obtain an observation sequence, and its kth element is:

in Indicates the estimate of the narrowband interference contained in the ith pilot symbol waveform, is the kth observed waveform in the observation matrix, N _p is the number of pilots, and M is the number of observed waveforms;

Several pilot observation sequences obtained from step (2) Subtract the corresponding narrow-band interference observation sequence from , to obtain a pilot observation sequence that removes narrowband interference;

[[{y the y}_{11 - - f f},, {y the y}_{22 - - f f},, ... ...,, {y the y}_{{N N}_{p p} - - f f}]] = = \begin{matrix} [[{y the y}_{11},, {y the y}_{22},, ... ...,, {y the y}_{{N N}_{p p}}]] - - [[{y the y}^{11}_{f f},, {y the y}^{22}_{f f},, ... ...,, {y the y}_{f f}^{{N N}_{p p}}]] \\ [[{y the y}_{11} - - {y the y}^{11}_{f f},, {y the y}_{22} - -,, {y the y}_{f f}^{22},, . . . . . .,, {y the y}_{{N N}_{p p}} - - {y the y}_{f f}^{{N N}_{p p}}]] \end{matrix}

in, i=1,2,...,N _p represents the observation sequence of the i-th pilot symbol waveform after narrowband interference is suppressed.

6. the method for claim 2 is characterized in that, the concrete steps of described step (4) are:

Several pilot observation sequences after suppressing narrowband interference obtained in step (3) Perform an averaging operation to obtain the average pilot observation sequence To reduce the influence of additive white Gaussian noise, that is

\overset{&OverBar; &OverBar;}{y the y} = = \frac{11}{{N N}_{p p}} {Σ Σ}_{i i = = 11}^{{N N}_{p p}} {y the y}_{i i - - f f} = = \frac{11}{{N N}_{p p}} {Σ Σ}_{i i = = 11}^{{N N}_{p p}} (({y the y}_{i i} - - {y the y}_{f f}^{i i}))

Among them, y _if , i=1,2,...,N _p represents the observation sequence after the ith pilot waveform suppresses the narrowband interference;

For signal-related template reconstruction, the sparse dictionary used is the feature vector sparse dictionary Ψ _g , and its generation process is as follows:

Due to the time-varying nature of pulsed UWB channels, the pulsed UWB signal g(t) received in step (1) is essentially a random process; therefore, its covariance function R(t-τ ),Right now

R(t-τ)=E[g(t)g(τ+t)]

Assume that λ ₁ >λ ₂ >λ ₃ >...>λ _N represents the eigenvalue of the Fredholm integral operator, where N is the number of sampling points when each pilot symbol waveform is subjected to Nyquist sampling; for R(t -τ) then there are:

∫R(t-τ)u _j (τ)dτ＝λ _j u _j (t)

Among them, u _j (t) is the eigenvector corresponding to λ _j , and {u _j (t)} is a complete set of orthogonal basis functions, satisfying:

&Integral; &Integral; {u u}_{i i} ((t t)) {u u}_{j j} ((t t)) = = \{\begin{matrix} \begin{matrix} 00 & i i &NotEqual; &NotEqual; j j \end{matrix} \\ \begin{matrix} 11 & i i = = j j \end{matrix} \end{matrix},, i i,, j j = = 11,, 22,, 33,, ... ...,, N N

Therefore, [u ₁ (t),u ₂ (t),u ₃ (t),...,u _N (t)] constitute a set of orthogonal basis of g(t), which is is a sparse dictionary of feature vectors;

From step (2) the Gaussian random observation matrix and eigenvector sparse dictionary Ψ _g , to obtain the relevant template reconstruction matrix, namely

V _g = ΦΨ _g

Combined with the average pilot observation sequence obtained in this step The signal correlation template g(t), t∈(0,T _s ) is obtained by reconstructing with the OMP algorithm.

7. the method for claim 2 is characterized in that, the concrete steps of described step (5) are:

First, use the Gaussian random observation matrix from step (2) Observe the waveforms of each load symbol in the load part of the pulse UWB signal in step (1), and obtain the observation sequence of each load symbol waveform in j=1,2,...,N _s is an M×1-dimensional column vector, representing the observation sequence of the jth load symbol waveform, and its kth element is:

in Indicates the jth disturbed load signal waveform, is the kth observed waveform in the observation matrix, N _s is the number of load symbols, and M is the number of observed waveforms;

Then, according to the estimation method to the narrowband interference contained in the pilot signal in step (2), utilize the OMP algorithm to obtain the estimation of the narrowband interference contained in each load symbol waveform represents the estimate of the narrowband interference contained in the j-th loaded symbol waveform, and Expressed as:

{\overset{^^}{f f}}_{n no b b}^{j j} ((t t)) = = {Ψ Ψ}_{n no b b} {\overset{^^}{θ θ}}_{n no b b}^{j j},, t t &Element; &Element; ((((j j - - 11)) {T T}_{s the s} + + {N N}_{p p} {T T}_{s the s},, {jT J}_{s the s} + + {N N}_{p p} {T T}_{s the s})),, j j = = 11,, 22,, ... ...,, {N N}_{s the s};;

Among them, Ψ _nb is a narrow-band interference sparse dictionary, is the projection coefficient vector of the narrowband interference in the jth load symbol waveform obtained by using the OMP algorithm j = 1, 2, ..., estimates of N _s ;

Finally, use the subtractor to eliminate the influence of narrowband interference from the load signal waveform, and obtain the load signal waveform s _load (t) that suppresses narrowband interference:

{S S}_{l l o o a a d d} ((t t)) = = {Σ Σ}_{j j = = 11}^{{N N}_{s the s}} [[{\overset{\cdot &Center Dot;}{g g}}_{j j} ((t t)) - - {\overset{^^}{f f}}_{n no b b}^{j j} ((t t))]],, t t &Element; &Element; ((((j j - - 11)) {T T}_{s the s} + + {N N}_{p p} {T T}_{s the s},, {jT J}_{s the s} + + {N N}_{p p} {T T}_{s the s})),, j j = = 11,, 22,, ... ...,, {N N}_{s the s} . .