CN116056030B

CN116056030B - Multi-sensor distributed detection method for environment backscattering under single channel

Info

Publication number: CN116056030B
Application number: CN202310018792.4A
Authority: CN
Inventors: 张高远; 陈开�; 冀保峰; 唐杰; 李兴旺; 韩瑽琤; 宋欢欢; 文红; 李永恩; 马聪芳; 张晓辉; 穆昱; 张冀; 张平; 孙力帆; 陶发展; 马华红; 吴红海; 谢萍; 张丽丽
Original assignee: Henan University of Science and Technology
Current assignee: Henan University of Science and Technology
Priority date: 2023-01-06
Filing date: 2023-01-06
Publication date: 2025-12-12
Anticipated expiration: 2043-01-06
Also published as: CN116056030A

Abstract

A multi-sensor distributed detection method assisted by environmental backscattering under single-channel conditions is proposed. A random binary information bit 'a' generated by a radio frequency (RF) source in the environment is transmitted through the channel and observed by I single-antenna local environmental backscattering sensors, simultaneously received by a fusion center configured with I antennas. When a binary information bit 'u' generated by the source is simultaneously observed by I independent local environmental backscattering sensors, a hard decision is made, and the RF signal is reflected based on the decision result. The reflected signal from each link is forwarded through a binary symmetric channel and then modulo-2 added with the RF signal received by the fusion center. The calculation result from each link is then transmitted to the fusion center through another binary symmetric channel. The fusion center derives a low-complexity decision metric based on the received signal and compares the decision metric with a decision threshold to obtain the detection result. This invention features low computational complexity, strong robustness, and high bandwidth utilization.

Description

Multi-sensor distributed detection method assisted by environmental backscattering under single channel

Technical Field

The invention relates to the technical field of information, in particular to a multi-sensor distributed detection method assisted by environmental backscattering under a single channel.

Background

The widely covered information-aware network is a key infrastructure for equipment manufacturing to convert and upgrade to 'digital', 'networked', 'intelligent'. The perception layer at the bottom layer of the network is mainly responsible for wireless access of equipment manufacturing local intelligent terminal equipment. On the premise that the network layer accurately and timely transmits data, the accuracy of data processing and the accuracy of data mining conclusion of the application layer depend on the quality of the data of the perception layer. Therefore, the reliable transmission of the data of the ubiquitous sensing network sensing layer is guaranteed, the method is the most effective path for realizing the reliable sensing of the equipment manufacturing time-space information from the information source, and the method is extremely important for the research of the method.

The data provided by a single sensor has not met the needs of equipment manufacturing digitization and intelligent development. Multiple sensors must be used to provide multi-feature observation information to make comprehensive, efficient, accurate and reasonable fusion decisions, estimates or decisions, and multi-source information fusion technology becomes a necessary choice. However, the conventional multi-sensor information fusion research mostly does not consider that cooperative communication is performed by means of relay nodes, and when the transmission distance is far, normal information transmission cannot be performed, i.e. the communication distance is greatly restricted, and the network coverage capability is insufficient. Only a small amount of researches on multi-relay decision fusion exist, but all the researches need to perfectly estimate the instantaneous Channel State Information (CSI) of each relay channel in each transmission link, so that the realization complexity is high and the engineering application is not easy. Moreover, most of the existing researches are developed under the parallel topological structure, so that more system bandwidth is consumed, and when the sensor data are more, bandwidth resources are consumed more.

In a wireless sensor network with strictly limited resources, each sensor node has no way to connect to an energy network due to its wireless nature, and the volume of the node is usually small and cannot carry a large-capacity battery. For some nodes in complex environments, the way to replace the battery for the node will bring more cost. If the energy supply of the relay node is insufficient, the whole wireless sensor network is easy to face the paralysis problem. When the fusion center performs decision fusion, if complex nonlinear operations such as logarithm and multiplication are needed, transmission delay is increased, and information transmission energy consumption is increased. It becomes particularly important how to reduce the computational complexity of the fusion center. Secondly, when the traditional wireless sensor network performs decision fusion in the fusion center, the CSI from the local sensor to the fusion center needs to be acquired, and in practical application, the estimation process of the CSI involves higher implementation complexity, high energy consumption and high energy consumption because the channel state is dynamically changed in real time. This is contrary to the concept of low complexity, low cost and low power consumption of wireless sensor networks. Moreover, when there is an error in the estimation of CSI, the performance of the entire system will be drastically reduced. That is, the detection system employing accurate CSI is not robust to CSI. Furthermore, most of the existing studies are deployed in parallel topologies, requiring that each local sensor be assigned a dedicated channel to transmit local observation information to the fusion center. When the number of local sensors is large, more bandwidth resources are consumed. The above-mentioned technical shortcomings limit to a certain extent the application depth and breadth of the multi-path multi-relay wireless sensor network in the reliable transmission of equipment manufacturing perception data.

Disclosure of Invention

In order to solve the technical problems, the invention provides the environment backscattering assisted multi-sensor distributed detection method under a single channel, which utilizes the characteristics of low energy consumption of a local sensor, low calculation complexity of a fusion center, strong decision fusion robustness and high bandwidth utilization rate by utilizing the multi-sensor to detect under the single channel as a multiple access channel based on an environment backscattering technology.

In order to achieve the technical aim, the technical scheme adopted is that the multi-sensor distributed detection method assisted by environmental backscattering under a single channel comprises the following steps:

Step S1, a radio frequency source in the environment randomly generates binary information bits a, the binary information bits a are observed by I single-antenna local environment backscatter sensors after channel transmission, and are simultaneously received by a fusion center configured with I antennas, wherein I is an odd number, and the received signals of the I single-antenna environment backscatter sensors are Fusing the received signal of the ith antenna of the center intoE _i represents the transmission error of the BSC between the radio frequency source and the I-th single-antenna local environment backscatter sensor, and Λ _i represents the transmission error of the BSC between the radio frequency source and the I-th antenna of the fusion center, wherein I is more than or equal to 1 and less than or equal to I;

Step S2, the binary information bit u generated by the information source is respectively and independently observed by the I single-antenna local environment backscattering sensors, based on the maximum likelihood criterion, the I-th single-antenna local environment backscattering sensor carries out hard judgment on the observed data sample to obtain x _i,x_i =0 or x _i =1, and according to the judgment result x _i, the I-th single-antenna local environment backscattering sensor determines how to reflect the radio frequency receiving signal b _i received in the step S1, when x _i =0, the b _i is reflected as is, and when x _i =1, the b _i is reflected after bit overturn;

Step S3, the reflected signals generated in the step S2 are received by the ith antenna of the fusion center after being forwarded by J-1 relay nodes, and are marked as y _i,y_i =0 or y _i =1, wherein J is the BSC signal number, y _i and the received signal c _i in the step S1 are subjected to modulo-two operation to obtain z _i, and then z _i of all transmission links and the channel error vector e are subjected to modulo-two operation at the fusion center to obtain the received information r for judgment, wherein r=0 or r=1;

s4, the fusion center extracts a low-complexity judgment metric value which does not contain channel state information according to the received information r;

and S5, comparing the low-complexity judgment metric value extracted in the step S4 with a judgment threshold to obtain a final detection result.

Further, in the step S3, the forwarding mode adopted by the relay node is amplification forwarding, and the amplification coefficient is 1.

Further, the method for extracting the low-complexity decision metric value by the fusion center in the step S4 is as follows:

Λ=2r-1

Where Λ represents a low complexity decision metric value that does not contain channel state information.

Further, the specific method for comparing the low complexity decision metric value with the decision threshold in step S5 is that when the low complexity decision metric value is greater than or equal to the decision threshold, the estimated value of the information u generated by the information source is obtained to be 1, and when the low complexity decision metric value is less than or equal to the decision threshold, the estimated value of the information u generated by the information source is obtained to be 0.

The invention has the beneficial effects that the invention provides the environment backscattering-assisted multi-sensor distributed detection method under the multiple access channel, the calculation is carried out through the established system detection model, the energy consumption of the local sensor is low based on the environment backscattering technology, the received information is obtained through special calculation in the fusion center, the low-complexity judgment metric value without any channel state information is calculated through a simple method, and finally the comparison detection is carried out.

Compared with the optimal decision fusion judgment method, the detection method provided by the invention has the characteristics of less performance loss, no need of any Channel State Information (CSI), low complexity, low cost and easiness in implementation. The method is very suitable for wireless sensor networks with strictly limited power consumption and computing capacity.

Drawings

FIG. 1 is a flow chart of the data fusion decision method of the present invention;

fig. 2 is a transmission model diagram of a communication system in an embodiment of the present invention;

FIG. 3 is a diagram of an equivalent system transmission model in an embodiment of the invention;

FIG. 4 is a diagram of an equivalent channel model from source to received value z _i in an embodiment of the present invention;

FIG. 5 is a plot of BER impact of the number of routes versus performance based on an accurate decision metric extraction method where the channels between the source to the local environmental backscatter sensor, the environmental RF source to the local environmental backscatter sensor, and the RF source to the antennas of the fusion center are all ideal, in an embodiment;

FIG. 6 is a graph of FER impact of the number of routes versus performance based on the accurate decision metric extraction method in an embodiment where the channels between the source to the local environment backscatter sensor, the ambient RF source to the local environment backscatter sensor, and the RF source to the antennas of the fusion center are all ideal;

FIG. 7 is a graph of BER impact of relay number on performance based on accurate decision metric extraction method in an embodiment where the channels between the source to the local environmental backscatter sensor, the environmental RF source to the local environmental backscatter sensor, and the RF source to the antennas of the fusion center are all ideal;

FIG. 8 is a graph of FER impact of relay number on performance based on an accurate decision metric extraction method for an embodiment where the channels between the source to the local environment backscatter sensor, the ambient RF source to the local environment backscatter sensor, and the RF source to the antennas of the fusion center are all ideal;

FIG. 9 is a graph of BER impact of relay number on performance based on a reduced decision metric extraction method for an embodiment where the channels between the source to the local environmental backscatter sensor, the environmental RF source to the local environmental backscatter sensor, and the RF source to the antennas of the fusion center are all ideal;

FIG. 10 is a graph of FER impact of relay number on performance based on a reduced decision metric extraction method for an embodiment where the channels between the source to the local environment backscatter sensor, the ambient RF source to the local environment backscatter sensor, and the RF source to the antennas of the fusion center are all ideal;

FIG. 11 is a BER impact graph of the number of routes versus performance based on a reduced decision metric extraction method where the channels between the source to the local environmental backscatter sensor, the environmental RF source to the local environmental backscatter sensor, and the RF source to the antennas of the fusion center are all ideal, in an embodiment;

FIG. 12 is a graph of FER impact of number of routes on performance based on a reduced decision metric extraction method for embodiments where the channels between the source to the local environment backscatter sensor, the ambient RF source to the local environment backscatter sensor, and the RF source to the fusion center antennas are all ideal;

FIG. 13 is a graph of BER impact based on a performance comparison of the best decision metric extraction method and the simplified decision metric extraction method in an embodiment where the channels between the source to the local environment backscatter sensors, the ambient RF source to the local environment backscatter sensors, and the RF source to the antennas of the fusion center are all ideal;

Fig. 14 is a graph of FER effect based on performance comparison of the best decision metric extraction method and the simplified decision metric extraction method in the case where the channels between the source to the local environment backscatter sensor, the ambient rf source to the local environment backscatter sensor, and the rf source to the antennas of the fusion center are all ideal in the embodiment.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all, embodiments of the present invention, and all other embodiments obtained by those skilled in the art without making any inventive effort based on the embodiments of the present invention are within the scope of protection of the present invention.

The decision fusion method for the environment backscattering assistance under the multiple access channel is realized based on a local single-antenna environment backscattering sensor, a relay node and a fusion center, as shown in fig. 1, and comprises the following steps of S1, randomly generating binary information bits a by a radio frequency source in the environment, wherein a=0 or a=1. a is observed by the local environment backscatter sensors of I single antennas after channel transmission, and is received by a fusion center of I antennas, wherein I is an odd number. Let the receiving value of the I (I is more than or equal to 1) th path single antenna environment back scattering sensor beFusing the receiving value of the ith antenna of the center to beE _i represents the transmission error of the BSC between the radio frequency source and the I-th single-antenna local environment backscatter sensor, and Λ _i represents the transmission error of the BSC between the radio frequency source and the I (1≤i≤I) antenna of the fusion center.

The application of the environment backscattering technology can well solve the problem of insufficient energy supply of the traditional wireless sensor network, and the specific implementation process is that the observed information source information is loaded on radio frequency signals widely existing in the space by using passive tags and reflected to a fusion center, and only a small amount of energy is consumed by a node part in the transmission process. The single-antenna local environment backscatter sensor in the method configures a single antenna for the sensor and is provided with a passive tag, and the local sensor has lower energy consumption in data transmission.

In step S2, the binary information bits u (u=0 or u=1) generated by the source are simultaneously and independently observed by the I single-antenna local environment backscatter sensors, respectively. Based on the maximum likelihood criterion, the i-th single-antenna local environment backscatter sensor makes a hard decision on the observed data sample to obtain x _i,x_i =0 or x _i =1. According to the decision result x _i, the i-th path single antenna local environment backscatter sensor decides how to reflect the radio frequency received signal b _i received in step S1. When x _i =0, b _i is reflected as it is, and when x _i =1, b _i is reflected after bit flipping.

S21, after the ith path of local environment backscattering sensor observes binary data H ₀ or H ₁ sent by a signal source, carrying out hard decision according to a maximum likelihood criterion to obtain information x _i;

and step S3, the reflected signal generated by the i-th path single antenna local environment backscatter sensor in step S2 is forwarded by J-1 relay nodes and then received by the i-th antenna of the fusion center, wherein y _i,y_i =0 or y _i =1 is recorded, and J is the number of BSC signals. In view of the inherent superposition characteristics of the multiple access channels (multiple sensors access single channel), y _i and the environmental radio frequency receiving signal c _i in step S1 are subjected to modulo-two operation to obtain z _i, and z _i (I is greater than or equal to 1) of all transmission links and the channel error vector e are subjected to modulo-two operation at the fusion center to obtain receiving information r for judgment, wherein r=0 or r=1.

S31, information reflected by an i-th path single antenna local environment backscatter sensor is forwarded by J-1 relay nodes to obtain y _i;

The relationship between x _i and y _i is:

Where e _i,j denotes a channel transmission error of the jth BSC of the ith transmission link, Representing a modulo two addition operation.

And S32, carrying out modulo-two operation on y _i and the environment radio frequency receiving signal c _i in the step S1 to obtain z _i, and carrying out modulo-two operation on all z _i (I is more than or equal to 1 and less than or equal to I) and the channel error vector e at a fusion center to obtain receiving information r for judgment.

In step S31, the relay node adopts an amplifying and forwarding strategy, and the amplifying coefficient is 1.

the method for extracting the judgment metric value by the fusion center in the step S4 comprises the following steps:

when the channel between the source and each single antenna local environment backscatter sensor is an ideal channel and the number of transmission paths I is an odd number, there are

Λ=2r-1 (2)

Where Λ represents the decision metric value extracted by the fusion center based on the received value r, without any instantaneous CSI, in case the channel between the source to the local environment backscatter sensor is ideal and the number of transmission paths I is odd.

S5, comparing the low-complexity decision metric value extracted in the step S4 with a decision threshold to obtain a final detection result (decision result).

Further, the method for comparing the low complexity decision metric value with the decision threshold in step S5 is as follows:

Where τ represents the decision threshold, The estimated value of the information u generated by the information source is represented, and the low complexity judgment metric value does not contain any instantaneous CSI, so that the judgment is simple and convenient to realize, and the detection judgment can be carried out through 0 and 1.

The system comprises an environment radio frequency source, each single-antenna local environment backscattering sensor, each independent antenna of a fusion center, a local environment backscattering sensor, a relay node and a channel between the relay node and the fusion center, wherein the channel between the relay node and the fusion center is BSC (binary symmetrical channel).

The extracting and processing method theoretical derivation of the low complexity decision metric value by the fusion center in the step S4 comprises the following steps:

a1 z _i at the fusion center can be expressed as

A2. let α _i＝P(E_i =1) be the probability of error transfer of BSC between the radio source and the ith local environment backscatter sensor, β _i＝P(Λ_i =1) be the probability of error transfer of BSC between the radio source and the ith antenna of the fusion center, and ε _i,j＝P(e_i,j =1) be the probability of error transfer of jth BSC for the ith transmission link, then it is known from equation (4):

Without loss of generality, let alpha _i＝ε_i,0,β_i＝ε_i,J+1 be

A3 As can be seen from the settlement results in equation (5), the statistical characteristics of the data transfer process from the binary data H ₀ or H ₁ sent from the source to z _i at the fusion center can be expressed as:

a3. at the fusion center, the received value r for the decision can be expressed as: Where e is the error vector introduced by the MAC at the fusion center. The task of the fusion center is to decide whether the binary data sent by the source is H ₀ or H ₁ according to the received value r. From the maximum likelihood criterion in logarithmic form, the expression for the best decision metric can be derived as:

let P (e=1) =γ, then when the fusion center receives a value r=1, a combination of formulas (6) and (7) is available, and formula (8) can be further expressed as

When the fusion center received value r=0, a combination of formulas (6) and (7) is available, and formula (8) can be further expressed as

Then combining formula (9) and formula (10) to obtain

Thus, the optimal decision metric final expression of the fusion center is obtained.

A4, in the implementation process of the optimal decision metric value, namely the formula (11), the fusion center needs to perfectly acquire two parameters representing the decision performance characteristics of each local environment backscattering sensor, namely the detection probability P _di and the false alarm probability P _fi, the instantaneous CSI of each relay BSC of each transmission link, namely the error transition probability epsilon _i,j, the error transition probability alpha _i of the BSC between the radio frequency source and each local environment backscattering sensor, the error transition probability beta _i of the BSC between the radio frequency source and the fusion center and the error transition probability gamma of the MAC, and the implementation process of the formula (11) involves a large number of logarithms and multiplications, so that the implementation complexity is high, the energy consumption is high, and therefore, the fusion detection method with low complexity and no instantaneous CSI is necessary is acquired.

A5, giving a sub-optimal decision metric value (the low-complexity decision metric value of the application) based on the optimal decision metric value, namely the theoretical derivation process of the formula (2).

A6. considering that the channel of the source to the local environment backscatter sensor is ideal, there is P _fi =0 and P _di =1 at this point.

When the fusion center receives the value r=1 and the number of transmission paths I is an odd number, the equation (9) can be written as

When the fusion center received value r=0 and the number of transmission paths I is an odd number, the equation (10) can be written as

The decision metric values obtainable by equations (12) and (13) in the case where the local sensor is ideal can be written as

A7 for ease of analysis, let epsilon _I+1 =γ, the decision metric value in equation (14) can be changed to

A8, using the relationship between the tanhθ and arctanh θ, i.e

And order theThen the logarithmic term in equation (15) may become:

A9 when the number of relays J-1 is large, ε _i to 0.5 can be seen from the formula (5), since Then delta _i to 0 at this time, formula (17) becomes

Then formula (15) can be approximated as

A10 is that for 1≤i≤I:

Order the Then there is

A11 when the number of relays J-1 is large, ε _i,j→0.5,η_i,j.fwdarw.0, formula (21) may be changed

A12, in combination with formulae (19) and (22), formula (15) can be written as

Note that the number of the components to be processed,

A13 since η _i,j and δ _i,j are both positive, then in formula (23)Positive values. Therefore, when the decision threshold τ is set to 0, the positive term in the reject (23) is discardedThe term does not affect the final decision result. The final sub-optimal decision metric value is obtained

Λ=2r-1 (24)

As shown in fig. 1, the workflow of each transmission link of the system is such that a binary information bit a randomly generated by a radio frequency source in the environment is received by a fusion center of I single antenna local environment backscatter sensors and I antennas configured. The source generates binary bit information u, and after being observed by the I local environment backscatter sensors simultaneously through the wireless channel, hard decisions are made according to maximum likelihood criteria. When the judgment result is 0, the received radio frequency signal is reflected out as it is, and when the judgment result is 1, the received radio frequency signal is reflected out after bit turning. The reflected signal of each link is transmitted through J-1 independent binary symmetrical channels, and then is subjected to modulo-two addition operation with the radio frequency signal received by the fusion center. The calculation result of each link finally reaches the fusion center through a binary symmetrical channel, the fusion center obtains a low-complexity decision metric value required by decision according to the received signals, and the final decision result is obtained after the decision metric value is compared with a decision threshold.

The method is suitable for wireless sensor networks with strictly limited power consumption and computing capacity, such as target detection under multiple sensors, defect detection under multiple sensors, fault detection under multiple sensors and the like, the area where the multiple sensors are located is observed through the multiple sensors, corresponding comparison results are obtained according to the method, whether targets, faults or defects exist or not is finally judged, the comparison method according to the formula (3) is determined according to actual conditions, for example, targets exist, 1 is finally detected, and no targets are 0.

The system error rate performance and the frame error rate performance when the best decision metric value extraction method is used are given in fig. 5 to 8, and the decision threshold τ is set to 0. In fig. 5 and 6, the relay number is set to 2, and the route number I is uniformly changed from 3 to 7. It can be seen from the figure that there is a threshold phenomenon, and when the probability of the error transition of the BSC is smaller than the threshold, the performance of the system changes more significantly, and when the probability is larger than the threshold, the system changes more slowly. And as the number of routes increases, the performance of the system tends to decrease. The detection performance in the relay number change scenario is given in fig. 7 and 8. As the number of relays increases, the system performance also decreases.

The error rate performance and the frame error rate performance of the system when using the sub-optimal decision metric extraction method are given in fig. 9 to 12, and the decision threshold τ is set to 0. The influence of the change of the route quantity and the change of the relay quantity on the detection performance of the system is respectively given. It can be observed that the curve trend in the case of the sub-optimal decision metric is very similar to the optimal decision metric.

The best decision metric value and the next best decision metric value are compared in simulation in the case where the channel from the source to the local sensor is ideal, and the decision threshold τ is set to 0 as shown in fig. 13 and 14. The relay number J is fixed to 2, and the selected route number I is changed from 3 to 7. As can be seen from fig. 13 and 14, the performance curves under the best and the next best metrics almost coincide, which illustrates that the sub-best decision metric method proposed by this patent has a performance loss within a small range with a large complexity reduction.

In summary, the low-complexity decision fusion method under the multiple access channel provided by the invention has the characteristics of high reliability, strong robustness, high bandwidth utilization rate and low computational complexity.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The environment backscattering assisted multi-sensor distributed detection method under the single channel is characterized by comprising the following steps of:

2. The method for multi-sensor distributed detection assisted by environmental backscatter in a single channel according to claim 1, wherein the forwarding mode adopted by the relay node in step S3 is amplification forwarding, and the amplification factor is 1.

3. The method for single channel environment backscatter assisted multi-sensor distributed detection of claim 1, wherein the method for fusion center extraction of low complexity decision metric in step S4 is:

Λ=2r-1

4. The method for performing the comparison between the low complexity decision metric and the decision threshold in step S5, wherein the estimated value of the source generation information u is 1 when the low complexity decision metric is greater than or equal to the decision threshold, and the estimated value of the source generation information u is 0 when the low complexity decision metric is less than or equal to the decision threshold.