CN111769904B - Detection method for parallel transmission of multiple reflection devices in backscatter communication system - Google Patents

Detection method for parallel transmission of multiple reflection devices in backscatter communication system Download PDF

Info

Publication number
CN111769904B
CN111769904B CN202010579132.XA CN202010579132A CN111769904B CN 111769904 B CN111769904 B CN 111769904B CN 202010579132 A CN202010579132 A CN 202010579132A CN 111769904 B CN111769904 B CN 111769904B
Authority
CN
China
Prior art keywords
access point
reflection
signals
devices
signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010579132.XA
Other languages
Chinese (zh)
Other versions
CN111769904A (en
Inventor
梁应敞
杨刚
袁冬冬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN202010579132.XA priority Critical patent/CN111769904B/en
Publication of CN111769904A publication Critical patent/CN111769904A/en
Application granted granted Critical
Publication of CN111769904B publication Critical patent/CN111769904B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0078—Avoidance of errors by organising the transmitted data in a format specifically designed to deal with errors, e.g. location
    • H04L1/0083—Formatting with frames or packets; Protocol or part of protocol for error control
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
    • H04L1/0056—Systems characterized by the type of code used

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Error Detection And Correction (AREA)
  • Radio Transmission System (AREA)

Abstract

The invention belongs to the technical field of communication, and particularly relates to a method for detecting parallel transmission of multiple reflection devices in a backscattering communication system. The method of the invention is directed at a system of an access point and a plurality of reflection devices, after clustering is carried out on received signals through a maximum expectation algorithm, constellation points of the received signals are classified, and finally clustering of the constellation points is completed; and then, according to transition information of signal constellation points received at bit boundaries of different reflection devices, dividing the clustered clusters into a high level group and a low level group according to the condition that different reflection devices jump at the bit boundaries by utilizing the characteristic that FM0 codes jump at the level boundaries, and then decoding data of the reflection devices according to a corresponding division mode. The invention can realize the detection and decoding of the parallel transmission of a plurality of reflecting devices under the condition of not needing channel state information and pilot frequency.

Description

Detection method for parallel transmission of multiple reflection devices in backscatter communication system
Technical Field
The invention belongs to the technical field of communication, and particularly relates to a method for detecting parallel transmission of multiple reflection devices in a backscattering communication system.
Background
A backscatter communication system is a system that generates and transmits a radio frequency sinusoidal carrier wave by a reader to energize a nearby backscatter device (hereinafter referred to as a "reflector device"), i.e., a tag, and to carry the reflector device information back to the reader. The typical representative of the technology is RFID technology, an RFID reader sends a sinusoidal carrier to a tag, one part of a signal received by the tag is used for energy collection to meet the normal work of a tag circuit, and the other part of the signal is used for backscattering to transmit information of the tag to the reader. The backscattering communication technology can enable the label to get rid of the constraint of a battery, solve the energy problem of the sensor of the Internet of things and play an important role in the future Internet of things.
The existing backscatter communication system multi-tag transmission receiver algorithm has some defects. Taking the RFID system as an example, when multiple tags transmit in parallel in the RFID system, collision occurs between the multiple tags. At present, the mainstream processing method is to prevent collision of a plurality of labels in a time division multiplexing mode, and the method is divided into two major categories, namely a non-deterministic anti-collision algorithm and a deterministic anti-collision algorithm. The non-deterministic anti-collision algorithm represented by the ALOHA algorithm has the defect that the label starvation phenomenon is easy to occur, and the deterministic anti-collision algorithm represented by the binary search algorithm has the defect that when the number of labels is large, the depth of a query tree is very deep, and the reading and writing speed of a reader is influenced. Moreover, neither the deterministic anti-collision algorithm nor the deterministic anti-collision algorithm really realizes the parallel transmission detection of a plurality of labels, which limits the throughput performance of the system.
Disclosure of Invention
The invention mainly provides a signal detection method for multi-tag parallel transmission in a backscattering communication system.
The technical scheme adopted by the invention is as follows:
the backscattering communication system comprises an access point and a plurality of reflection devices, wherein the access point is provided with one or a plurality of antennas for sending and receiving signals, the access point sends downlink signals or pure carrier signals, the reflection devices select different backscattering coefficients to perform backscattering according to information bits, and the access point performs self-interference elimination and detects the signals of the reflection devices; the detection method is characterized by comprising the following steps:
s1, the access point sends a downlink signal; signal forms include, but are not limited to, pure carrier signals, OFDM signals, WiFi signals, cellular signals, or other radio frequency signals;
s2, the reflection equipment collects energy and performs backscattering on the received signal according to the information carried by the reflection equipment; the power reflection coefficient of the reflecting device is configured to be a fixed constant known to the access point, and the reflecting device absorbs and extracts energy from the signal transmitted by the access point. The reflection device reflects the incident signal back to the access point in different amplitudes and phases by continuously changing the impedance of the antenna, so that the incident signal is modulated, and the purpose of communicating with the access point is achieved. A plurality of reflection devices in the system can simultaneously reflect signals to the access point, and the access point recovers the information of each reflection device from the received superposed signals;
s3, the access point receives the back scattered signal, carries out self-interference elimination and detects the signal of the reflecting device: the reflection equipment reconstructs the self-interference signal in a digital domain and/or an analog domain and eliminates the self-interference signal from a received signal, and then the information of the reflection equipment is recovered by utilizing a machine learning clustering algorithm and the coding characteristics of the reflection equipment; the method specifically comprises the following steps:
when M reflection devices send signals simultaneously, the signals y (n) received by the access point are the superposition of the signals of all the reflection devices; assuming that the signal period of the reflection device is D times the period of the signal transmitted by the access point, the received signal of the qth receiving antenna of the access point is:
Figure GDA0003132386080000021
where p is the transmit power of the access point, αmIs the power reflection coefficient of the m-th reflecting device, fm,qRepresenting the channel response of the q-th receiving antenna of the access point to the m-th reflecting device, gm,qRepresenting the channel response from the mth reflecting device to the qth receiving antenna of the access point, s (n) representing the signal transmitted by the access point, xmIs the signal of the m-th reflecting device, uq(n) is the power σ2Additive gaussian noise of (a);
the receiver combines the continuous D symbols to establish the observed quantity on each antenna
Figure GDA0003132386080000022
By combining the (L-1) th to the lD-th symbols, where L is 1
Figure GDA0003132386080000023
Q is the number of receiver antennas; combining samples obtained in continuous LD s (n) symbol periods to obtain observation vector
Figure GDA0003132386080000024
Figure GDA0003132386080000025
The distribution of (a) is in accordance with a gaussian mixture model,
Figure GDA0003132386080000026
the mapped constellation points will appear as 2MA cluster, i.e.
Figure GDA0003132386080000027
Is in accordance with 2MAnd the element Gaussian mixture model is used for clustering the constellation points of the observation vectors:
let K be 2M，
Figure GDA0003132386080000031
The probability density function of (a) is:
Figure GDA0003132386080000032
wherein θ ═ pi1,π2,π3,...,πK,θ1,θ2,θ3,...,θK)，πkIs a weight coefficient representing the weight of the kth Gaussian component in the Gaussian mixture model, and
Figure GDA0003132386080000033
θk＝(μk,Σk)，μkis the mean of the Gaussian components, sigmakIs a matrix of the covariance,
Figure GDA0003132386080000034
is a gaussian distribution probability density function:
Figure GDA0003132386080000035
introduction of a variable of introduction gammal＝[γl1,γl2,γl3,...,γlk,...,γlK]Wherein L ═ 1,2, 3.., L, γlkIs defined as follows:
Figure GDA0003132386080000036
definition of
Figure GDA0003132386080000037
Obtaining log-likelihood functions of the fully observed data for the fully observed dataComprises the following steps:
Figure GDA0003132386080000038
clustering by using a maximum expectation algorithm, wherein the maximum expectation algorithm comprises a step E and a step M, the step E is a Q-function, and the Q-function is a log-likelihood function of complete data
Figure GDA0003132386080000039
At a given observation y and current parameter θiUnder the condition of introducing a variable gammalExpectation of conditional probability of (a): :
Figure GDA00031323860800000310
step M is to solve the function Q (theta )i) Maximum to θ; step E and step M are continuously executed in an iterative mode, and finally the estimated value theta is converged;
after obtaining the value obtained by the iteration of the maximum expectation algorithm, classifying the constellation points of the received signals y (n), and finally finishing the clustering of the constellation points;
collecting transition information of constellation points of received signals y (n) at bit boundaries of different reflection devices, and dividing clustered clusters into high-level groups and low-level groups according to the condition that different reflection devices jump at the bit boundaries by utilizing the characteristic that FM0 codes jump at the level at the bit boundaries;
and after different cluster dividing modes are obtained, decoding the data of the reflection equipment according to the corresponding dividing modes.
The invention has the beneficial effects that: the invention provides a parallel transmission detection method based on a maximum expectation algorithm and the digital coding hopping characteristic of reflection equipment, which converts the parallel detection problem into the clustering problem through a machine learning technology and can realize the detection and decoding of the parallel transmission of a plurality of reflection equipment under the condition of no need of channel state information and pilot frequency.
Drawings
FIG. 1: a schematic diagram of a backscatter communications system;
FIG. 2: a transmission frame structure chart of a parallel transmission backscattering communication system;
FIG. 3: a schematic diagram of a reflection apparatus;
FIG. 4: receiving a signal constellation diagram when the two reflection devices transmit in parallel;
FIG. 5: FM0 encoding schematic diagram;
FIG. 6: two reflection devices transmit schematic diagrams in parallel;
FIG. 7: receiving the transition condition of a signal at a constellation point at the 1-bit boundary moment of the reflection equipment;
FIG. 8: grouping process of the clusters by the reflection device 1;
FIG. 9: t is0A schematic time diagram;
FIG. 10: the grouping result of the clusters by the reflection apparatus 2;
FIG. 11: an error rate curve graph of a parallel transmission detection algorithm based on EM and DBSCAN clustering;
FIG. 12: and (3) an error rate comparison graph of the parallel detection algorithm based on EM clustering under different numbers of reflecting devices.
Detailed Description
The technical scheme of the invention is further described in detail by combining the accompanying drawings:
first, as shown in fig. 1, the backscatter communication system proposed by the present invention includes an access point (such as an RFID system reader/writer or an access point of other types of backscatter communication systems) and M (M ≧ 1) reflection devices;
as shown in fig. 2, all of the reflecting devices reflect signals to the access point at the same time;
as shown in fig. 3, each of the reflecting devices includes:
a backscatter antenna module: for receiving and reflecting signals of an environmental access point;
a backscatter modulation module: the load impedance of the antenna is changed according to the information symbol, so that backscattering modulation is realized;
a microcontroller module: a communication process for controlling the reflecting device;
the signal processor module: for the reflection device to perform basic signal processing, such as decoding of control signals, etc.;
radio frequency energy harvester and battery module: the system is used for collecting energy from incident signals and charging batteries to supply power to all modules;
other modules including units of storage, sensing, clocking, etc.
The reflecting device modulates its received incident signal by deliberately switching the load impedance to change the amplitude and/or phase of its backscattered signal, and the backscattered signal is received and ultimately decoded by the full-duplex access point.
The invention provides a backscatter communication and parallel detection method for parallel transmission of multiple reflection devices, which comprises the following steps:
s1, an access point sends a downlink signal;
s2, the reflection equipment collects energy and performs backscattering on a received signal according to information carried by the reflection equipment;
s3, receiving the backscattered signals by the access point, eliminating self-interference and detecting signals of the reflection equipment;
further, the signal form in step S1 includes, but is not limited to, an unmodulated carrier signal, an OFDM signal, a WIFI signal, a cellular signal, or other radio frequency signal:
the power reflection coefficient of the reflecting device in said step S2 is configured to be a fixed constant known to the access point, and the reflecting device absorbs the signal emitted by the access point and extracts energy therefrom. The reflection device reflects the incident signal back to the access point in different amplitudes and phases by continuously changing the impedance of the antenna, so that the incident signal is modulated, and the purpose of communicating with the access point is achieved. A plurality of reflecting devices in the system simultaneously reflect signals to the access point, and the access point recovers the information of each reflecting device from the received superposed signals.
In step S3, the reflection device reconstructs the self-interference signal in the digital domain and/or the analog domain and eliminates it from the received signal, and then uses the machine learning clustering algorithm and the coding features of the reflection device to recover the information of the plurality of reflection devices transmitted in parallel.
In the following, a signal processing flow at a data transmission stage is described in detail, and a parallel transmission detection method for multiple reflection devices is further proposed.
The system-related information is described as: s (n) information symbols, x, transmitted by the access pointmAnd an information symbol representing the mth reflecting device, and the symbol period of the reflecting device is assumed to be D times (D is more than or equal to 1) of the transmission symbol period of the access point. Assuming the channel to be quasi-static, using fm,qRepresenting the channel response of the q-th receiving antenna of the access point to the m-th reflecting device, gm,qRepresenting the channel response of the mth reflecting device to the qth receiving antenna of the access point. Power reflection coefficient of m-th reflecting device is represented by alphamAnd (4) showing.
Assuming that the transmit power of the access point is P, the reception of the backscatter signal from the m reflecting device by the qth receiving antenna of the access point can be written as:
Figure GDA0003132386080000061
since the M reflection devices reflect signals simultaneously, the signals received by the access point are a superposition of the signals of the respective reflection devices. Assuming that the receiver has Q receiving antennas, the received signal of the Q-th receiving antenna
Figure GDA0003132386080000062
Wherein u isq(n) is the power σ2Is the additive gaussian noise of (a) a,
Figure GDA0003132386080000063
the receiver combines successive D symbols, and since the signal s (n) is known to the receiver, an observation on each antenna can be established
Figure GDA0003132386080000064
Wherein s is*(n) represents the conjugate of the signal s (n). Tong (Chinese character of 'tong')After combining the (L-1) th to the (lD) th symbols, L is 1
Figure GDA0003132386080000065
Furthermore, samples obtained in consecutive LD s (n) symbol periods are combined and recorded as observation vectors
Figure GDA0003132386080000066
Note that, the above
Figure GDA0003132386080000067
Following the gaussian mixture model, the parallel transmission of two reflection devices and the use of a single antenna as a receiver will be described. When the number M of reflecting devices transmitted in parallel is 2,
Figure GDA0003132386080000068
can be written as:
Figure GDA0003132386080000069
typically, s (n) obeys distribution
Figure GDA00031323860800000610
And D is a large integer (reflecting device signal x in practical applications)mIs much lower than the symbol rate of the downlink signal s (n) sent by the access point), then
Figure GDA0003132386080000071
The following distribution is obeyed:
Figure GDA0003132386080000072
from the above, when the M reflection devices simultaneously transmit information, a signal is received
Figure GDA0003132386080000073
The distribution of (A) is in accordance with the Gaussian Mixture Model (GMM). Receiving a signal
Figure GDA0003132386080000074
The mapped constellation points will appear as 2MAnd (4) clustering. As shown in fig. 4, when M is 2, the signal is received
Figure GDA0003132386080000075
4 clusters are mapped on the constellation diagram. The dashed line with arrows in the figure represents the channel vector, and it can be seen that when the reflection device 1 transmits '1' and the reflection device 2 transmits '0', the constellation points of the received signals will be located in the cluster labeled 2 in the figure. Conversely, when the reflection device 1 transmits '0' and the reflection device 2 transmits '1', the reception signal is located in the cluster denoted by reference numeral 4. When both the reflection device 1 and the reflection device 2 transmit '1', the reception signal may be located in a cluster denoted by 1. When both reflecting devices transmit a '0', the constellation point of the received signal will be in the cluster labeled 3.
In step S3, the parallel detection algorithm first uses an Expectation Maximization (EM) algorithm to cluster constellation points of the received signal at the access point, and clusters the constellation points into a plurality of independent clusters. Then, the mapping relation between the clustered clusters and the data transmitted by each reflection device is obtained by utilizing the level jump characteristic of FM0, so that the information of each reflection device is decoded under the condition of no channel information and no pilot frequency.
The clustering process in step S3 is as follows:
Figure GDA0003132386080000076
is in accordance with 2MA meta GMM model. Let K be 2M，
Figure GDA0003132386080000077
The probability density function of (a) can be expressed as:
Figure GDA0003132386080000078
wherein θ ═ pi1,π2,π3,...,πK,θ1,θ2,θ3,...,θK)。πkIs a weight coefficient representing the weight of the kth Gaussian component in the GMM model, and
Figure GDA0003132386080000079
is a Gaussian distribution probability density function, θk＝(μk,Σk) The expression is as follows:
Figure GDA00031323860800000711
the EM algorithm introduces an implicit variable gammanThis way a more easily solved function is constructed.
Hidden variable gammal＝[γl1,γl2,γl3,...,γlk,...,γlK]Wherein L is 1,2, 3. Gamma raylkIs defined as follows:
Figure GDA0003132386080000081
at this time, define
Figure GDA0003132386080000082
The data were fully observed. The log-likelihood function of the fully observed data can be expressed as:
Figure GDA0003132386080000083
the EM algorithm is gradually close to the optimal solution through a continuous iterative updating mode. Each iteration is composed of steps E and M, and the steps E and M are continuously iterated to finally make the estimated value theta converge on a certain value.
Let thetaiAnd thetai+1Respectively represent the ith wheelThe value of θ before and after iteration. Step E in the EM algorithm is to find the Q function. The Q function is the log-likelihood function of the complete data
Figure GDA0003132386080000084
At a given observation y and current parameter θiUnder the condition of introducing a variable gammalExpectation of conditional probability of (a):
Figure GDA0003132386080000085
the M steps of the EM algorithm are to solve the function Q (theta )i) For maximum value of theta, i.e. solving for thetai+1。
Figure GDA0003132386080000086
Are used separately
Figure GDA0003132386080000087
And
Figure GDA0003132386080000088
representing the parameter theta to be updatedi+1Mean and variance, and weight coefficients.
The equation 7 is respectively aligned to muk，Σk，πkIf the partial derivative is found to be equal to 0, the deviation can be found
Figure GDA0003132386080000089
The results are as follows:
Figure GDA00031323860800000810
Figure GDA0003132386080000091
Figure GDA0003132386080000092
the EM algorithm must converge, but not necessarily to a globally optimal solution, but may also converge to a locally optimal solution. The EM algorithm is sensitive to the initial values, different initial values are selected, and the convergence results may be different. Therefore, the k-means + + algorithm can be used first for the received signal
Figure GDA0003132386080000093
Pre-clustering is carried out, then an initial value of the EM algorithm is calculated according to a clustering result, and a final classification result is obtained through iteration of the EM algorithm, so that the clustering accuracy is guaranteed as much as possible.
The process of acquiring the mapping relationship between the clusters and the reflection device data in step S3 is as follows:
the existing commercial reflective devices are encoded using FM0, and the rule for FM0 encoding is: if the data bit is '1', the level is flipped at the beginning of the data bit and continues until the beginning of the next data bit. If the data bit is '0', the level flips at the beginning of the data bit and again at an intermediate time of the data bit.
Fig. 5 is a schematic diagram of FM0 encoding, and it can be seen that whether the data is '0' or '1', the waveform has a flip at the beginning of each data bit. The time at which each data bit begins is referred to as the "bit boundary".
In addition, when a plurality of reflecting devices reflect signals simultaneously, the reflecting devices are not synchronous. Fig. 6 shows the parallel transmission of two reflecting devices with a single receiving antenna. In the figure, (a) is an original signal at the reflection apparatus 1, the period of the reflection apparatus 1 is 30us, and the response time delay is 10 us. (b) The period of the reflecting device 2 is 32us and the response time delay is 30us for the original signal at the reflecting device 2. (c) The backscatter signals generated for the reflecting device 1 and the reflecting device 2 propagate to the superimposed signal generated at the access point. The longer dotted line in the figure, and denoted t1,mIs the bit boundary time of the reflection device 1. Short dash-dot line, and mark t2,mIs the bit boundary time of the reflecting device 2. Can see twoThe bit boundaries of the reflecting device are not coincident most of the time.
This causes the received signal to flip in level at the bit boundary due to FM0 encoding
Figure GDA0003132386080000094
The constellation point of (a) makes a transition at the bit boundary instant of each reflection device. FIG. 7 is a bit boundary t of the reflection device 11,1～t1,6Temporal level flip causes
Figure GDA0003132386080000095
In the case of transition of constellation points between different clusters. As can be seen,
Figure GDA0003132386080000096
at constellation point t1,1At a time instant jumping from cluster labeled D to cluster labeled B, t1,2The moment jumps from C to A, t1,3And t1,4The moment jumps from B to D, t1,5The moment jumps from C to A, t1,6The time jumps from B to C.
Collecting received signals
Figure GDA0003132386080000104
Transition information of constellation points at bit boundaries of all reflecting devices. As shown in table 1 and table 2, the transition information of the constellation points of the reflection device 1 and the reflection device 2 at different bit boundary time instants, respectively.
Table 1: constellation point transition information of received signal at 1-bit boundary moment of reflecting equipment
Figure GDA0003132386080000101
Table 2: constellation point transition information of received signal at 2-bit boundary moment of reflecting equipment
Figure GDA0003132386080000102
As can be seen from the encoding characteristics of FM0, the clusters before and after the transition must represent opposite levels of the reflective device. For a reflective device, the clusters on the constellation diagram can be divided into 2 groups by their transition information at the bit boundary, where one group represents the high level of the reflective device m and the other group represents the low level of the reflective device m.
Fig. 8 is a process of grouping clusters on a constellation according to bit boundary transition information of the reflection apparatus 1 in table 1. (a) Indicating a bit boundary t at BD11,1Time of day, received signal
Figure GDA0003132386080000103
Transitions from the cluster labeled D to the cluster labeled B. This means that cluster B and cluster D represent opposite levels for BD 1. (b) Is shown at t1,2At that moment, the constellation point jumps from C to a, meaning that cluster C and cluster a represent opposite levels of BD 1. (c) Is shown at t1,6At that time, the constellation point jumps from B to C, and cluster C and cluster B represent opposite levels of BD 1. (d) It indicates that the cluster C and the cluster D have the same level since the cluster B is opposite in level to both the cluster C and the cluster D.
Furthermore, there is one time in each round of communication: all reflecting devices having no reflected signal, using T0Indicating this time. At T0At that time, all the reflective devices are in a charged state, and are low. As shown in fig. 9, by observing the received signal
Figure GDA0003132386080000113
Can be found at T0The time constellation points are present in the X groups, which means that the X groups represent the low level of the reflection device 1 and the Y groups represent the high level of the reflection device 1. Therefore, the clusters C and D represent that the data transmitted by the reflection apparatus 1 is '0', and the clusters a and B represent that the data transmitted by the reflection apparatus 1 is '1'.
Through the above steps, the mapping relationship between the clustered clusters and the transmission data of the reflection apparatus 1 is obtained without channel information and pilot. The decoding of the reflection device 1 can now be completed as shown in table 3:
table 3: reflective device 1 information decoding
Figure GDA0003132386080000111
The information of the reflection apparatus 1 is finally obtained as '01011010011010011001', and the decoding is completed.
The decoding step of the reflection apparatus 2 is the same as that of the reflection apparatus 1, and the clusters of constellation points are grouped according to the information of table 2, and the result is shown in fig. 10. It can be seen that the grouping of clusters on the constellation diagram is different for different reflection devices because the bit boundary transition information is different for each reflection device. The reflecting device 2 divides the clusters into two groups, P and Q, the P group including cluster a and cluster C, and the Q group including cluster B and cluster D. Due to the received signal
Figure GDA0003132386080000112
At T0The constellation points at the time of day fall in group Q, so a cluster in group Q indicates that the reflecting device 2 is low, and a cluster in group P indicates that the reflecting device 2 is high.
The above-described process of decoding information for the reflection apparatus 2 obtains information of the reflection apparatus 2 as '011010010101100101', and the decoding is completed as shown in table 4.
Table 4: reflective device 2 information decoding
Figure GDA0003132386080000121
Through the steps, when a plurality of reflection devices transmit in parallel, the receiver can cluster the constellation points through a clustering algorithm. Then, with the level jump characteristic of FM0 encoding, clusters are grouped according to the constellation point transition information of each reflection device at its bit boundary instant, and then the information of the reflection device is decoded.
The parallel detection algorithm of density clustering (DBSCAN) is used as a comparison scheme to compare with the parallel detection algorithm proposed by the patent.
DBSCAN is a density-based clustering algorithm. The DBSCAN uses the degree of closeness between data as a classification standard, and data in the same class should be closely connected. And dividing all the data into one-by-one categories according to the judgment standard of whether the data are tightly concentrated together, and finally finishing the clustering of the data.
The DBSCAN has two important parameters, one of which is the neighborhood radius eps, which means the size of the neighborhood radius of a certain data. Another parameter is the minimum number of objects minPts.
Assuming that there is a point P, if the number of data points in the neighborhood of the point P with eps as the radius is not less than minPts, the point P is called as the core object. If another point Q is within the eps neighborhood of P, then the P-Q direct density is said to be achievable. If the data point within the eps neighborhood of point Q is greater than minPts, then point Q is also the core object. If the point K is in the eps neighborhood of Q but not in the eps neighborhood of P, the direct density of K-Q is called reachable, and the indirect density of K-P is called reachable. And judging that K finds that the number of neighborhood data points in the eps radius is less than minPts, and if K is not a core object, the K is called a boundary point. The boundary points can no longer propagate down.
Step of DBSCAN clustering: firstly, a core object which is not classified is selected arbitrarily, all density-reachable points of the core object are found, namely, the data points are classified into a set when the core object is propagated outwards until boundary points are met and the data points can not be propagated any more, and then clustering of a category is completed. And then continuously selecting a new core object, searching density reachable points of the new core object, and then clustering. This process is repeated until all the core objects are classified and clustering is complete.
In the system, a plurality of reflecting devices transmit in parallel, and an access point is provided with one or two receiving antennas. The channel is a gaussian channel, and it is assumed that the channel coefficients do not change much during transmission. The convergence threshold epsilon of EM algorithm is 10-3And stopping iteration when the iterative update of the EM algorithm is smaller than the convergence threshold.
It can be seen from fig. 11 that under given conditions, no matter whether the receiver is a single antenna or a dual antenna, the bit error rate of the parallel detection algorithm based on EM clustering proposed by the present invention is lower than that of the algorithm using DBSCAN clustering.
Second, it can be seen thatNo matter the parallel detection algorithm of EM clustering or DBSCAN clustering, the error rate of the double-antenna receiver is lower than that of a single-antenna receiver, and the performance is better when the number of the antennas of the receiver is larger. The error rate of the algorithm is derived from the error code in the clustering process, and when the clusters of the constellation points are overlapped or aliased, the clustering accuracy of the single-antenna receiver on the constellation points is influenced. The multi-antenna can make the signal received by the receiver have higher dimensionality when the constellation point is in a certain dimensionality (namely receiving sample point (vector)
Figure GDA0003132386080000131
An element of) may be separated in another dimension when coincidence or aliasing occurs. Therefore, the multi-antenna receiver can improve the clustering accuracy, and the multi-antenna receiver has lower detection error rate than a single antenna.
In addition, the simulation result shows that the performance of the parallel detection algorithm based on the DBSCAN clustering is poor under the condition of low signal to noise ratio. For example, when the SNR is 7dB, the bit error rate of the parallel detection algorithm based on the DBSCAN clustering is about 50% in both the single-antenna receiver and the dual-antenna receiver, which means that the receiver cannot decode. This is because the constellation points of the received signal have a relatively diffuse distribution at low snr, and there are situations where there is interpenetration between clusters. According to the DBSCAN clustering principle, the fuzzy boundary between clusters can cause the density reachable relationship between core particles of different classes, so that two classes which should be independent are merged, and the final classification number is wrong, thereby causing the receiver to be completely unable to decode. Different from the DBSCAN algorithm, the EM algorithm is a soft clustering algorithm, and clustering of data is realized by calculating the responsivity of each data point to each category. Therefore, the EM algorithm can still guarantee the correct cluster number even when clusters coincide under the condition of low signal-to-noise ratio. Therefore, the parallel transmission detection algorithm of the multi-reflection device has great performance advantages.
Fig. 12 shows bit error rate curves of the parallel detection algorithm based on EM clustering when the receiver is equipped with 2 receiving antennas, respectively, when 2 reflection devices and 3 reflection devices transmit in parallel. It can be seen that the error rate of parallel communication of 3 reflecting devices under the same signal-to-noise ratio is higher than that of parallel communication of 2 reflecting devices. This is because the more reflecting devices that communicate in parallel, the denser the constellation points of the received signal, the closer the distance between clusters, increasing the probability of clustering errors.

Claims (1)

1. A detection method for parallel transmission of multiple reflection devices in a backscattering communication system comprises an access point and multiple reflection devices, wherein the access point is provided with one or more antennas for sending and receiving signals, the access point sends downlink signals or pure carrier signals, the reflection devices select different backscattering coefficients to perform backscattering according to information bits, and the access point performs self-interference elimination and detects signals of the reflection devices; the detection method is characterized by comprising the following steps:
s1, the access point sends a downlink signal;
s2, the reflecting devices collect energy, and the reflecting devices simultaneously reflect signals to the access point, wherein the power reflection coefficient of the emitting devices is a known fixed constant of the access point;
s3, the access point receives the backscattered signals and detects the signals of the reflecting equipment; the method specifically comprises the following steps:
when M reflection devices send signals simultaneously, the signals y (n) received by the access point are the superposition of the signals of all the reflection devices; the signal period of the reflection device is D times of the period of the signal transmitted by the access point, and then the received signal of the qth receiving antenna of the access point is:
Figure FDA0003132386070000011
where p is the transmit power of the access point, αmIs the power reflection coefficient of the m-th reflecting device, fm,qRepresenting the channel response of the q-th receiving antenna of the access point to the m-th reflecting device, gm,qTo representThe channel response of the mth reflecting device to the qth receiving antenna of the access point, s (n) representing the signal transmitted by the access point, xmIs the signal of the m-th reflecting device, uq(n) is the power σ2Additive gaussian noise of (a);
the receiver combines the continuous D symbols to establish the observed quantity on each antenna
Figure FDA0003132386070000012
By combining the (L-1) th to the lD-th symbols, where L is 1
Figure FDA0003132386070000013
Q is the number of receiver antennas; combining samples obtained in continuous LD s (n) symbol periods to obtain observation vector
Figure FDA0003132386070000014
Figure FDA0003132386070000015
The distribution of (a) is in accordance with a gaussian mixture model,
Figure FDA0003132386070000016
the mapped constellation points will appear as 2MA cluster, i.e.
Figure FDA0003132386070000017
Is in accordance with 2MAnd the element Gaussian mixture model is used for clustering the constellation points of the observation vectors:
let K be 2M，
Figure FDA0003132386070000021
The probability density function of (a) is:
Figure FDA0003132386070000022
wherein θ ═ pi1,π2,π3,...,πK,θ1,θ2,θ3,...,θK)，πkIs a weight coefficient representing the weight of the kth Gaussian component in the Gaussian mixture model, and
Figure FDA0003132386070000023
θk＝(μk,Σk)，μkis the mean of the Gaussian components, sigmakIs a matrix of the covariance,
Figure FDA0003132386070000024
is a gaussian distribution probability density function:
Figure FDA0003132386070000025
introduction of a variable of introduction gammal＝[γl1,γl2,γl3,...,γlk,...,γlK]Wherein L ═ 1,2, 3.., L, γlkIs defined as follows:
Figure FDA0003132386070000026
definition of
Figure FDA0003132386070000027
For the fully observed data, the log-likelihood function for obtaining the fully observed data is:
Figure FDA0003132386070000028
clustering is carried out by using a maximum expectation algorithm, wherein the maximum expectation algorithm comprises an expectation step and a maximization step, the expectation step is a Q function, and the Q function is a log likelihood function of complete data
Figure FDA0003132386070000029
At a given observation y and current parameter θiUnder the condition of introducing a variable gammalExpectation of conditional probability of (a): :
Figure FDA00031323860700000210
maximizing step as solving function Q (theta )i) Maximum to θ; continuously and iteratively executing the expectation step and the maximization step to finally converge the estimated value theta;
after obtaining the value obtained by the iteration of the maximum expectation algorithm, classifying the constellation points of the received signals y (n), and finally finishing the clustering of the constellation points;
collecting transition information of constellation points of received signals y (n) at bit boundaries of different reflection devices, and dividing clustered clusters into high-level groups and low-level groups according to the condition that different reflection devices jump at the bit boundaries by utilizing the characteristic that FM0 codes jump at the level at the bit boundaries;
and after different cluster dividing modes are obtained, decoding the data of the reflection equipment according to the corresponding dividing modes.
CN202010579132.XA 2020-06-23 2020-06-23 Detection method for parallel transmission of multiple reflection devices in backscatter communication system Active CN111769904B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010579132.XA CN111769904B (en) 2020-06-23 2020-06-23 Detection method for parallel transmission of multiple reflection devices in backscatter communication system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010579132.XA CN111769904B (en) 2020-06-23 2020-06-23 Detection method for parallel transmission of multiple reflection devices in backscatter communication system

Publications (2)

Publication Number Publication Date
CN111769904A CN111769904A (en) 2020-10-13
CN111769904B true CN111769904B (en) 2021-08-17

Family

ID=72721759

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010579132.XA Active CN111769904B (en) 2020-06-23 2020-06-23 Detection method for parallel transmission of multiple reflection devices in backscatter communication system

Country Status (1)

Country Link
CN (1) CN111769904B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112637103B (en) * 2020-12-18 2021-10-08 电子科技大学 Signal detection method of cooperative backscattering communication system
CN112668352B (en) * 2020-12-23 2022-04-22 中山大学 Environmental backscattering communication signal processing method
CN113014358B (en) * 2021-03-05 2022-06-28 苏州至善视界信息技术有限公司 Self-adaptive decoding method and decoding system applied to backscattering system
CN118201095A (en) * 2022-12-12 2024-06-14 维沃移动通信有限公司 Transmission method, device, terminal and network equipment

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2820847A1 (en) * 2010-12-07 2012-06-14 Hiok Nam Tay Auto-focus image system
US10454741B2 (en) * 2015-10-15 2019-10-22 Phasorlab, Inc. High-precision blind carrier synchronization methods for LTE SC-FDMA uplink
CN106685538B (en) * 2016-11-01 2020-05-01 清华大学 Environmental backscattering system and signal transmission method thereof
CN106526589B (en) * 2016-12-26 2018-11-30 浙江大学 A kind of radar target two-dimensional imaging method based on vortex electromagnetic wave
CN109060618A (en) * 2018-08-20 2018-12-21 国网冀北电力有限公司张家口供电公司 A kind of photovoltaic power generation power prediction device and method
CN109917378B (en) * 2018-12-26 2023-04-11 西安电子科技大学 Video SAR moving object detection method using space-time correlation

Also Published As

Publication number Publication date
CN111769904A (en) 2020-10-13

Similar Documents

Publication Publication Date Title
CN111769904B (en) Detection method for parallel transmission of multiple reflection devices in backscatter communication system
CN102124690B (en) MIMO receiver using ML depth-first and K-BEST detectors for SNR higher and lower ththreshold
CN107135017B (en) Signal transmitting and receiving method of backscattering communication system
CN106685538B (en) Environmental backscattering system and signal transmission method thereof
CN109889275B (en) Layered optical space modulation method suitable for turbulent flow channel
Zheng et al. DL-PR: Generalized automatic modulation classification method based on deep learning with priori regularization
CN101064579B (en) Method for detecting low-complexity globular decoding
CN103888217A (en) Sphere decoding detection method and device
CN107911193B (en) The symbol detection method of environment backscatter system based on Differential Manchester Encoding
CN106130615A (en) The activation antenna of generalized spatial modulation system and modulation symbol combined estimation method
CN109547376A (en) A kind of Modulation Identification method of the gauss hybrid models based on modulating characteristic
Zhang et al. Learning modulation filter networks for weak signal detection in noise
CN111682958B (en) Environmental backscattering signal detection method based on cluster analysis
Chen et al. End-to-end PSK signals demodulation using convolutional neural network
CN111641438B (en) System, method and processor readable medium for estimating number of RFID tags
Kim et al. A supervised-learning detector for multihop distributed reception systems
Li et al. Recovery of collided RFID tags with frequency drift on physical layer
Huang et al. A Physical Layer Algorithm for Estimation of Number of Tags in UHF RFID Anti-Collision Design.
CN110300409A (en) Limited piece long covert communications implementation method based on channel estimation
CN113343924B (en) Modulation signal identification method based on cyclic spectrum characteristics and generation countermeasure network
Rahman et al. Ensemble classifier based modulation recognition for beyond 5G massive MIMO (mMIMO) communication
Akyıldız et al. ML-aided collision recovery for UHF-RFID systems
Bag et al. Anti-collision algorithm for RFID system using adaptive Bayesian Belief Networks and it's VLSI Implementation
CN113935449A (en) Signal sending and receiving method of backscattering communication system
Perenda et al. Contrastive learning with self-reconstruction for channel-resilient modulation classification

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant