A multi-sensor information fusion system and method based on blockchain

Abstract

The present invention relates to the field of information fusion technology, and specifically to a multi-sensor information fusion system and method based on blockchain, wherein the fusion method comprises the following steps: generating a sensor information group according to the information source in the information fusion model; classifying historical information fusion events; obtaining the amount of information transmitted to the information fusion model by each sensor and the amount of information adopted by the information fusion model, and calculating the information adoption rate; obtaining the expected information adoption rate according to the information adoption rates of historical key sensors and related sensors; when several sensors collect information, obtaining the actual information adoption rate according to the actual amount of information transmitted and the actual amount of information adopted by each sensor; determining the expected information adoption rate according to the sensor information of each sensor and the key sensors; and adjusting the information transmission process of each sensor to the information fusion model.

Description

Technical Field

[0001] The present invention relates to the field of information fusion technology, and specifically to a multi-sensor information fusion system and method based on blockchain. Background Art

[0002] Multi-sensor information fusion is an information processing process that uses computer technology to automatically analyze and synthesize information and data from multiple sensors or multiple sources under certain criteria to complete the required decision-making and estimation.

[0003] In most cases, the multi-sensor information fusion adopted is to mix the information collected by each sensor together, then extract features from all the information, and finally obtain the fused information; however, due to the large number of sensors collecting information, the collected information will become more, which leads to the generation of a lot of redundant information. Even if a lot of information is collected, not much information is actually available in the end. In addition, the influence of redundant information reduces the accuracy of the final fused information; at the same time, the finally generated fused information will be saved in the blockchain, which is difficult to modify again. Therefore, it is necessary to process the redundant information in time to obtain new fused information. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-sensor information fusion system and method based on blockchain to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a multi-sensor information fusion method based on blockchain, the fusion method comprising the following steps:

[0006] Step S100: setting an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fused information; generating a sensor information group according to the information source of each historical fusion information event in the information fusion model; and classifying the historical information fusion events by comparing the sensor information of any two groups of sensor information groups;

[0007] Step S200: Obtain each information fusion event in the history of similar information fusion events, the amount of information transmitted by each sensor to the information fusion model and the amount of information adopted by the information fusion model, and calculate the information adoption rate of each sensor; divide each sensor in an information fusion event into key sensors and related sensors according to the size of the information adoption rate, and obtain the expected information adoption rate of the key sensors and related sensors in an information fusion event according to the historical information adoption rates of the key sensors and related sensors;

[0008] Step S300: when a plurality of sensors collect information, the actual information adoption rate of each sensor is obtained according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model; the expected information adoption rate of the plurality of sensors is determined according to the sensor information of each sensor and the key sensor;

[0009] Step S400: By comparing the actual information adoption rate of each sensor with the expected information adoption rate, the information transmission process of each sensor to the information fusion model is adjusted.

[0010] Furthermore, the step S100 includes the following steps:

[0011] Step S101: reading each information fusion event in the historical information fusion events, extracting several features generated in the information fusion model for each information fusion event, capturing the information source of each feature in the information fusion model, obtaining the sensor information corresponding to each feature, and generating a sensor information group for an information fusion event;

[0012] Step S102: Obtain the sensor information group of each information fusion event in the historical information fusion events, select two sets of sensor information groups of any two information fusion events, compare whether the number of sensors in the two sets of sensor information groups is the same, if the number is the same, select any sensor information in one set of sensor information groups, read whether the same sensor information exists in the other set of sensor information groups, if the same sensor information exists, classify the two information fusion events corresponding to the two sets of sensor information groups into the same category.

[0013] Furthermore, the step S200 includes the following steps:

[0014] Step S201: selecting any two information fusion events in any category of information fusion events in historical information fusion events, and extracting fusion information generated by the two information fusion events in the information fusion model;

[0015] Step S202: Select one of the information fusion events and obtain the information transfer amount M transmitted by each sensor to the information fusion model in the information fusion model. i , where M i is the amount of information transmitted by the i-th sensor to the information fusion model; obtains the amount of information adopted by the information fusion model for each sensor when extracting features in the information fusion model C i , where C i is the amount of information adopted by the information fusion model for the information transmitted by the i-th sensor. The information adoption rate of the i-th sensor in the information fusion event is calculated as A i =Ci / M i ;

[0016] Step S203: The sensor with the highest information adoption rate in an information fusion event is taken as the key sensor of the information fusion event, and the remaining sensors are set as related sensors; information fusion events with the same key sensor in a type of information fusion event are set as the same information fusion event, and the information adoption rate A of the i-th sensor in the same information fusion event is set as i The highest; the total number of the same information fusion events with the highest information adoption rate of the i-th sensor in history is N i , according to the formula:

[0017] L i =N i -ceil(N i ×τ)

[0018] Among them, ceil() is the upward rounding function, τ is the data selection rate, and the number of event selections of the same information fusion event is calculated; the same information fusion event is selected according to the information adoption rate A i Sort from high to low, select L i The information adoption rate is taken as the expected information adoption rate G of the i-th sensor in the same information fusion event. i ;

[0019] Since the adoption rate of key sensors is high, it means that the accuracy of information transmission of key sensors is high. Therefore, the information transmission of key sensors is relatively stable. To set the expected information adoption rate of key sensors, it is only necessary to confirm the expected information adoption rate based on the highest adoption rates in the historical information fusion time.

[0020] Step S204: extract the information adoption rate of each relevant sensor in each same information fusion event, average the information adoption rate of each relevant sensor, and calculate the average information adoption rate in, is the average information adoption rate of the j-th sensor; the average information adoption rate of the j-th sensor as the expected information adoption rate of the jth sensor;

[0021] Since the fused information is finally stored in the blockchain, the unstable information adoption rate transmitted by the relevant sensors to the information transmission model will also be reflected in the blockchain. Therefore, averaging the information adoption rate of historical information fusion events can relatively accurately reflect the expected information adoption rate, and obtain the lowest information adoption rate while ensuring that the fused information is as unaffected by redundant information as possible, thereby reducing the difficulty of adjusting the relevant sensors.

[0022] Furthermore, the step S300 includes the following steps:

[0023] Step S301: Obtain the actual information transfer amount m of the information collected by the i-th sensor to the information fusion model i ; The information transmitted by each sensor is selected through the information fusion model, and the actual information adopted by the i-th sensor is set to c i , calculate the actual information adoption rate B of the i-th sensor i =c i / m i ;

[0024] Step S302: determine the information of each sensor that transmits information through the information fusion model, and generate an actual sensor information group; obtain the number of sensors in the actual sensor information group, and extract various information fusion events with the same number of sensors in the historical fusion events; select any sensor information in the actual sensor information group, and read whether there is the same sensor information in another group of sensor information groups. If any sensor information is the same, obtain several information fusion events with the same number of sensors and the same sensor information; extract a sensor with the highest actual information adoption rate, set the i-th sensor as the key sensor, and extract the information fusion event with the i-th sensor as the key sensor from several information fusion events with the same number of sensors and the same sensor information.

[0025] Furthermore, the step S400 includes the following steps:

[0026] Step S401: Obtain the expected information adoption rate G of the key sensor in the information fusion event where the key sensor is the i-th sensor i , let the actual information adoption rate of the i-th sensor be B i If B i <G i , then obtain the most recent n identical information fusion events in history, if all of the n identical information fusion events are B i <G i , obtain the non-adopting time interval of information transmitted by the information fusion model to the i-th sensor. When the sensor passes through the non-adopting time interval, it stops transmitting information to the information fusion model;

[0027] Since the information adoption rate of key sensors is maintained at a high level, the key sensors have good stability. When the information adoption rate is lower than the expected value, it may be due to other incidental factors and will not often cause problems with the results. Therefore, it is necessary to refer to the information adoption rates of the most recent times. If the adoption rate is lower than the expected value for many consecutive times, it can be said that there are problems in the information collection and information transmission process of the key sensors, and the key sensors need to be adjusted. Due to the stability of key sensors, the information collection period when the problem occurs will be relatively close, so directly stopping the information transmission during the invalid information period can ensure that the information adoption rate increases.

[0028] Step S402: Obtaining the expected information adoption rate of relevant sensors Set the actual information adoption rate of other sensors to B j , where j≠i, B j is the actual information adoption rate of the jth sensor; if Then the information fusion model obtains the non-adopted time interval T = {T1, T2, …, T p ,…,T q}, where T p is the pth unadopted time interval, T p ∈(min p ,max p ), min p is the minimum time of the pth unadopted time interval, max p is the maximum moment of the pth unadopted time interval, q is the number of unadopted time intervals;

[0029] Step S403: Obtain the total duration of the unadopted time interval T like or Then adjust the amount of information transmitted by the jth sensor to the information fusion model to Among them, m j is the actual information transmission amount m of the jth sensor i , calculate the new information transmission amount m′ of the jth sensor j ;

[0030] When the total length of the unadopted time interval is relatively continuous, no matter the total length is too long or too short, the information transmission process of the sensor can be stopped for the time interval;

[0031] Step S404: Select any three consecutive non-adopted time intervals, and set the pth non-adopted time interval T p ∈(min p ,maxp ), the p+1th unadopted time interval T p+1 ∈(min p+1 ,max p+1 ) and the p-1th unadopted time interval T p-1 ∈(min p-1 ,max p-1 ),like When , the amount of adopted information to be excluded when excluding unadopted information is set to x, according to the formula:

[0032]

[0033]

[0034] Calculate the amount of adopted information x that needs to be excluded when excluding unadopted information; adjust the j-th sensor to transmit information to the information fusion model every information transmission amount t, where,

[0035] When there is a certain periodicity in the non-adopted time interval, it is sufficient as long as the information transmission time of the sensor is periodic. Since the non-adopted time interval is not perfectly periodic, there is some valid information in the middle. Therefore, when excluding the non-adopted information, some valid information will be excluded. Therefore, calculating the amount of valid information excluded can ensure the periodic information transmission while ensuring that the information adoption rate can reach the expected information adoption rate, ensuring that the final result can be as accurate as possible.

[0036] In order to better implement the above method, a multi-sensor information fusion system is also proposed, and the fusion system includes a division module, an expected target setting module, a target confirmation module and an adjustment module;

[0037] The division module is used to set an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fusion information; generate a sensor information group according to the information source of each historical fusion information event in the information fusion model; and classify the historical information fusion events by comparing the sensor information of any two groups of sensor information groups;

[0038] The expected target setting module is used to obtain the information transmission amount transmitted by each sensor to the information fusion model and the information adoption amount of the information fusion model for each information fusion event in the history of similar information fusion events, and calculate the information adoption rate of each sensor; according to the size of the information adoption rate, each sensor in an information fusion event is divided into key sensors and related sensors, and according to the historical information adoption rates of key sensors and related sensors, the expected information adoption rates of key sensors and related sensors in an information fusion event are obtained;

[0039] The target confirmation module is used to obtain the actual information adoption rate of each sensor according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model when the multiple sensors are collecting information; and determine the expected information adoption rate of the multiple sensors according to the sensor information of each sensor and the key sensor;

[0040] The adjustment module is used to adjust the information transmission process transmitted by each sensor to the information fusion model by comparing the actual information adoption rate of each sensor with the expected information adoption rate.

[0041] Further, the division module includes a sensor information group generation unit and a classification unit;

[0042] The sensor information group generating unit is used to set an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fused information; generate a sensor information group according to the information source of each historical fusion information event in the information fusion model; the classification unit is used to classify the historical information fusion events by comparing the sensor information of any two groups of sensor information groups.

[0043] Further, the expected target setting module includes an information adoption rate calculation unit, a sensor division unit and an expected information adoption rate calculation unit;

[0044] The information adoption rate calculation unit is used to obtain the amount of information transmitted by each sensor to the information fusion model and the amount of information adopted by the information fusion model for each information fusion event in historical similar information fusion events, and calculate the information adoption rate of each sensor; the sensor classification unit is used to classify each sensor in an information fusion event into key sensors and related sensors according to the size of the information adoption rate; the expected information adoption rate calculation unit is used to determine the expected information adoption rate of the several sensors based on the sensor information of each sensor and the key sensors.

[0045] Further, the target confirmation module includes an actual information adoption rate unit and an expected information adoption rate confirmation unit;

[0046] The actual information adoption rate unit is used to obtain the actual information adoption rate of each sensor according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model when several sensors collect information; the expected information adoption rate confirmation unit is used to determine the expected information adoption rate of the several sensors according to the sensor information of each sensor and the key sensors.

[0047] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: (1) The present invention obtains the expected information adoption rate through the information adoption rate of each sensor in each historical information fusion event; by analyzing the historical information and excluding the information fusion events with low information adoption rate, the correct information adoption rate can be accurately obtained, which is convenient for timely exclusion of redundant information during subsequent comparison; (2) The present invention classifies the sensors according to the information adoption rate of each sensor. Since the key sensors have a high information adoption rate and transmit more accurate information, the processing of the key sensors is relatively direct; while for the related sensors, the reasons are discussed, and according to the time interval of the non-adopted information, it is determined whether the valid information is in the same time period or changes periodically, and different processing methods are given to make the final fusion information more accurate; (3) The present invention can accurately determine the expected target of each information fusion by matching the number of sensors and sensor information with the historical information fusion time, effectively adjust the information transmission amount of the sensor, and make the final fusion information more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0049] Figure 1 It is a schematic diagram of the steps of a multi-sensor information fusion method based on blockchain;

[0050] Figure 2 It is a structural diagram of a multi-sensor information fusion system based on blockchain. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] See also Figure 1 and Figure 2 The present invention provides a technical solution: a multi-sensor information fusion method based on blockchain, the fusion method includes the following steps:

[0053] Step S100: setting an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fused information; generating a sensor information group according to the information source of each historical fusion information event in the information fusion model; and classifying the historical information fusion events by comparing the sensor information of any two groups of sensor information groups;

[0054] Wherein, the step S100 includes the following steps:

[0055] Step S101: reading each information fusion event in the historical information fusion events, extracting several features generated in the information fusion model for each information fusion event, capturing the information source of each feature in the information fusion model, obtaining the sensor information corresponding to each feature, and generating a sensor information group for an information fusion event;

[0056] Step S102: obtaining the sensor information group of each information fusion event in the historical information fusion events, selecting two sensor information groups of any two information fusion events, comparing whether the number of sensors in the two sensor information groups is the same, if the number is the same, selecting any sensor information in one of the sensor information groups, reading whether the same sensor information exists in the other sensor information group, if the same sensor information exists, classifying the two information fusion events corresponding to the two sensor information groups into the same category;

[0057] Step S200: Obtain each information fusion event in the history of similar information fusion events, the amount of information transmitted by each sensor to the information fusion model and the amount of information adopted by the information fusion model, and calculate the information adoption rate of each sensor; divide each sensor in an information fusion event into key sensors and related sensors according to the size of the information adoption rate, and obtain the expected information adoption rate of the key sensors and related sensors in an information fusion event according to the historical information adoption rates of the key sensors and related sensors;

[0058] Wherein, the step S200 includes the following steps:

[0059] Step S201: selecting any two information fusion events in any category of information fusion events in historical information fusion events, and extracting fusion information generated by the two information fusion events in the information fusion model;

[0060] Step S202: Select one of the information fusion events and obtain the information transfer amount M transmitted by each sensor to the information fusion model in the information fusion model. i , where M iis the amount of information transmitted by the i-th sensor to the information fusion model; obtains the amount of information adopted by the information fusion model for each sensor when extracting features in the information fusion model C i , where C i is the amount of information adopted by the information fusion model for the information transmitted by the i-th sensor. The information adoption rate of the i-th sensor in the information fusion event is calculated as A i =C i / M i ;

[0061] Step S203: The sensor with the highest information adoption rate in an information fusion event is taken as the key sensor of the information fusion event, and the remaining sensors are set as related sensors; information fusion events with the same key sensor in a type of information fusion event are set as the same information fusion event, and the information adoption rate A of the i-th sensor in the same information fusion event is set as i The highest; the total number of the same information fusion events with the highest information adoption rate of the i-th sensor in history is N i , according to the formula:

[0062] L i =N i -ceil(N i ×τ)

[0063] Among them, ceil() is the upward rounding function, τ is the data selection rate, and the number of event selections of the same information fusion event is calculated; the same information fusion event is selected according to the information adoption rate A i Sort from high to low, select L i The information adoption rate is taken as the expected information adoption rate G of the i-th sensor in the same information fusion event. i ;

[0064] Step S204: extract the information adoption rate of each relevant sensor in each same information fusion event, average the information adoption rate of each relevant sensor, and calculate the average information adoption rate in, is the average information adoption rate of the j-th sensor; the average information adoption rate of the j-th sensor as the expected information adoption rate of the jth sensor.

[0065] Step S300: when a plurality of sensors collect information, the actual information adoption rate of each sensor is obtained according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model; the expected information adoption rate of the plurality of sensors is determined according to the sensor information of each sensor and the key sensor;

[0066] Wherein, the step S300 includes the following steps:

[0067] Step S301: Obtain the actual information transfer amount m of the information collected by the i-th sensor to the information fusion model i ; The information transmitted by each sensor is selected through the information fusion model, and the actual information adopted by the i-th sensor is set to c i , calculate the actual information adoption rate B of the i-th sensor i =c i / m i ;

[0068] Step S302: determine the information of each sensor that transmits information through the information fusion model, and generate an actual sensor information group; obtain the number of sensors in the actual sensor information group, and extract various information fusion events with the same number of sensors in the historical fusion events; select any sensor information in the actual sensor information group, and read whether there is the same sensor information in another group of sensor information groups. If any sensor information is the same, obtain several information fusion events with the same number of sensors and the same sensor information; extract a sensor with the highest actual information adoption rate, set the i-th sensor as the key sensor, and extract the information fusion event with the i-th sensor as the key sensor from several information fusion events with the same number of sensors and the same sensor information.

[0069] Step S400: by comparing the actual information adoption rate of each sensor with the expected information adoption rate, adjusting the information transmission process of each sensor to the information fusion model;

[0070] Wherein, the step S400 includes the following steps:

[0071] Step S401: Obtain the expected information adoption rate G of the key sensor in the information fusion event where the key sensor is the i-th sensor i , let the actual information adoption rate of the i-th sensor be B i If B i <G i , then obtain the most recent n identical information fusion events in history, if all of the n identical information fusion events are B i <G i , obtain the non-adopting time interval of information transmitted by the information fusion model to the i-th sensor. When the sensor passes through the non-adopting time interval, it stops transmitting information to the information fusion model;

[0072] Step S402: Obtaining the expected information adoption rate of relevant sensors Set the actual information adoption rate of other sensors to B j , where j≠i, Bj is the actual information adoption rate of the jth sensor; if Then the information fusion model obtains the non-adopted time interval T = {T1, T2, …, T p ,…,T q}, where T p is the pth unadopted time interval, T p ∈(min p ,max p ), min p is the minimum time of the pth unadopted time interval, max p is the maximum moment of the pth unadopted time interval, q is the number of unadopted time intervals;

[0073] Step S403: Obtain the total duration of the unadopted time interval T like or Then adjust the amount of information transmitted by the jth sensor to the information fusion model to Among them, m j is the actual information transmission amount m of the jth sensor i , calculate the new information transmission amount m′ of the jth sensor j ;

[0074] Step S404: Select any three consecutive non-adopted time intervals, and set the pth non-adopted time interval T p ∈(min p ,max p ), the p+1th unadopted time interval T p+1 ∈(min p+1 ,max p+1 ) and the p-1th unadopted time interval T p-1 ∈(min p-1 ,max p-1 ),like When , the amount of adopted information to be excluded when excluding unadopted information is set to x, according to the formula:

[0075]

[0076]

[0077] Calculate the amount of adopted information x that needs to be excluded when excluding unadopted information; adjust the j-th sensor to transmit information to the information fusion model every information transmission amount t, where,

[0078] A multi-sensor information fusion system based on blockchain, the fusion system includes a division module, an expected target setting module, a target confirmation module and an adjustment module;

[0079] The division module is used to set an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fusion information; generate a sensor information group according to the information source of each historical fusion information event in the information fusion model; and classify the historical information fusion events by comparing the sensor information of any two groups of sensor information groups;

[0080] The expected target setting module is used to obtain the information transmission amount transmitted by each sensor to the information fusion model and the information adoption amount of the information fusion model for each information fusion event in the history of similar information fusion events, and calculate the information adoption rate of each sensor; according to the size of the information adoption rate, each sensor in an information fusion event is divided into key sensors and related sensors, and according to the historical information adoption rates of key sensors and related sensors, the expected information adoption rates of key sensors and related sensors in an information fusion event are obtained;

[0081] The target confirmation module is used to obtain the actual information adoption rate of each sensor according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model when the multiple sensors are collecting information; and determine the expected information adoption rate of the multiple sensors according to the sensor information of each sensor and the key sensor;

[0082] The adjustment module is used to adjust the information transmission process transmitted by each sensor to the information fusion model by comparing the actual information adoption rate of each sensor with the expected information adoption rate.

[0083] Wherein, the division module includes a sensor information group generation unit and a classification unit;

[0084] The sensor information group generating unit is used to set an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fused information; generate a sensor information group according to the information source of each historical fusion information event in the information fusion model; the classification unit is used to classify the historical information fusion events by comparing the sensor information of any two groups of sensor information groups.

[0085] Wherein, the expected target setting module includes an information adoption rate calculation unit, a sensor division unit and an expected information adoption rate calculation unit;

[0086] The information adoption rate calculation unit is used to obtain the amount of information transmitted by each sensor to the information fusion model and the amount of information adopted by the information fusion model for each information fusion event in historical similar information fusion events, and calculate the information adoption rate of each sensor; the sensor classification unit is used to classify each sensor in an information fusion event into key sensors and related sensors according to the size of the information adoption rate; the expected information adoption rate calculation unit is used to determine the expected information adoption rate of the several sensors based on the sensor information of each sensor and the key sensors.

[0087] Wherein, the target confirmation module includes an actual information adoption rate unit and an expected information adoption rate confirmation unit;

[0088] The actual information adoption rate unit is used to obtain the actual information adoption rate of each sensor according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model when several sensors collect information; the expected information adoption rate confirmation unit is used to determine the expected information adoption rate of the several sensors according to the sensor information of each sensor and the key sensors.

[0089] Perform feature extraction on historical information fusion events, capture the information source of each feature in the information fusion model, and obtain the sensor information corresponding to each feature. The first information fusion event corresponds to sensors A, B, and C, the second information fusion event corresponds to sensors A and B, the third information fusion event corresponds to sensors B and C, and the fourth information fusion event corresponds to sensors A, B, and C. The first information fusion event and the fourth information fusion event are classified into the same category.

[0090] Assume that the sensor with the highest adoption rate in the first information fusion event is sensor A, and the sensor with the highest adoption rate in the fourth information fusion event is sensor B. For the first information fusion event, sensor A is taken as the key sensor, and for the fourth information fusion event, sensor B is taken as the key sensor. The information adoption rate of key sensor A is 95%, and the information adoption rate of key sensor B is 95%; the information adoption rate of the relevant sensors of the first information fusion event is 70%, and the information adoption rate of the relevant sensors of the fourth information fusion event is 70%;

[0091] When information collection is in progress, the information of the sensors that are collecting information is obtained, and the sensors that are collecting information are A, B, and C. When each sensor transmits information to the information fusion model, the information adoption rate of each sensor is calculated, where A is 94%, B is 75%, and C is 60%. Because A is a key sensor and 94% is less than 95%, the information fusion events of the three historical sensors ABC and A as the key sensor are selected. Because there are 0 information fusion events that meet the requirements, it is judged that the information adoption rate is lower than the expected information adoption rate, which is a special case, and the information transmission process of sensor A is not adjusted;

[0092] Therefore, the information adoption rate of C is 60%, which is less than 70%. Therefore, the non-adoption time intervals are 5s-10s, 15s-20s, and 25s-30s. Since 25-20-(15-10)=0, the amount of adopted information to be excluded when non-adopted information is set to x.

[0093]

[0094] The amount of adopted information that needs to be excluded when excluding unadopted information is calculated to be x = 33.3,

[0095] Therefore, every j-th sensor transmits information to the information fusion model every 37.4 pieces of information transmission.

[0096] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0097] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A multi-sensor information fusion method based on blockchain, characterized by: The fusion method comprises the following steps: Step S100: setting an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fused information; generating a sensor information group according to the information source of each historical fusion information event in the information fusion model; and classifying the historical information fusion events by comparing the sensor information of any two groups of sensor information groups; Step S200: Obtain each information fusion event in the history of similar information fusion events, the amount of information transmitted by each sensor to the information fusion model and the amount of information adopted by the information fusion model, and calculate the information adoption rate of each sensor; divide each sensor in an information fusion event into key sensors and related sensors according to the size of the information adoption rate, and obtain the expected information adoption rate of the key sensors and related sensors in an information fusion event according to the historical information adoption rates of the key sensors and related sensors; Step S300: when a plurality of sensors collect information, the actual information adoption rate of each sensor is obtained according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model; the expected information adoption rate of the plurality of sensors is determined according to the sensor information of each sensor and the key sensor; Step S400: by comparing the actual information adoption rate of each sensor with the expected information adoption rate, adjusting the information transmission process of each sensor to the information fusion model; The step S200 includes the following steps: Step S201: selecting any two information fusion events in any category of information fusion events in historical information fusion events, and extracting fusion information generated by the two information fusion events in the information fusion model; Step S202: Select one of the information fusion events and obtain the information transfer amount M transmitted by each sensor to the information fusion model in the information fusion model. i , where M i is the amount of information transmitted by the i-th sensor to the information fusion model; obtains the amount of information adopted by the information fusion model for each sensor when extracting features in the information fusion model C i , where C i is the amount of information adopted by the information fusion model for the information transmitted by the i-th sensor. The information adoption rate of the i-th sensor in the information fusion event is calculated as A i =C i / M i ; Step S203: The sensor with the highest information adoption rate in an information fusion event is taken as the key sensor of the information fusion event, and the remaining sensors are set as related sensors; information fusion events with the same key sensor in a type of information fusion event are set as the same information fusion event, and the information adoption rate A of the i-th sensor in the same information fusion event is set as i The highest; the total number of the same information fusion events with the highest information adoption rate of the i-th sensor in history is N i , according to the formula: L i =N i -ceil(N i ×τ) Among them, ceil() is the upward rounding function, τ is the data selection rate, and the number of event selections of the same information fusion event is calculated; the same information fusion event is selected according to the information adoption rate A i Sort from high to low, select L i The information adoption rate is taken as the expected information adoption rate G of the i-th sensor in the same information fusion event. i ; Step S204: extract the information adoption rate of each relevant sensor in each same information fusion event, average the information adoption rate of each relevant sensor, and calculate the average information adoption rate in, is the average information adoption rate of the j-th sensor; the average information adoption rate of the j-th sensor as the expected information adoption rate of the jth sensor.

2. According to a multi-sensor information fusion method based on blockchain in claim 1, it is characterized by: The step S100 includes the following steps: Step S101: reading each information fusion event in the historical information fusion events, extracting several features generated in the information fusion model for each information fusion event, capturing the information source of each feature in the information fusion model, obtaining the sensor information corresponding to each feature, and generating a sensor information group for an information fusion event; Step S102: Obtain the sensor information group of each information fusion event in the historical information fusion events, select two sets of sensor information groups of any two information fusion events, compare whether the number of sensors in the two sets of sensor information groups is the same, if the number is the same, select any sensor information in one set of sensor information groups, read whether the same sensor information exists in the other set of sensor information groups, if the same sensor information exists, classify the two information fusion events corresponding to the two sets of sensor information groups into the same category.

3. According to claim 2, a multi-sensor information fusion method based on blockchain is characterized in that: The step S300 includes the following steps: Step S301: Obtain the actual information transfer amount m of the information collected by the i-th sensor to the information fusion model i ; The information transmitted by each sensor is selected through the information fusion model, and the actual information adopted by the i-th sensor is set to c i , calculate the actual information adoption rate B of the i-th sensor i =c i / m i ; Step S302: determine the information of each sensor that transmits information through the information fusion model, and generate an actual sensor information group; obtain the number of sensors in the actual sensor information group, and extract various information fusion events with the same number of sensors in the historical fusion events; select any sensor information in the actual sensor information group, and read whether there is the same sensor information in another group of sensor information groups. If any sensor information is the same, obtain several information fusion events with the same number of sensors and the same sensor information; extract a sensor with the highest actual information adoption rate, set the i-th sensor as the key sensor, and extract the information fusion event with the i-th sensor as the key sensor from several information fusion events with the same number of sensors and the same sensor information.

4. According to a multi-sensor information fusion method based on blockchain in claim 3, it is characterized by: The step S400 includes the following steps: Step S401: Obtain the expected information adoption rate G of the key sensor in the information fusion event where the key sensor is the i-th sensor i , let the actual information adoption rate of the i-th sensor be B i If B i <G i , then obtain the most recent n identical information fusion events in history, if all of the n identical information fusion events are B i <G i , obtain the non-adopting time interval of information transmitted by the information fusion model to the i-th sensor. When the sensor passes through the non-adopting time interval, it stops transmitting information to the information fusion model; Step S402: Obtaining the expected information adoption rate of relevant sensors Set the actual information adoption rate of other sensors to B j , where j≠i, B j is the actual information adoption rate of the jth sensor; if Then the information fusion model obtains the non-adopted time interval T = T1, T2, ..., T p ,…,T q }, where T p is the pth unadopted time interval, T p ∈(min p ,max p ), min p is the minimum time of the pth unadopted time interval, max p is the maximum moment of the pth unadopted time interval, q is the number of unadopted time intervals; Step S403: Obtain the total duration of the unadopted time interval T like or Then adjust the amount of information transmitted by the jth sensor to the information fusion model to Among them, m j is the actual information transmission amount m of the jth sensor i , calculate the new information transmission amount m of the jth sensor j '; Step S404: Select any three consecutive non-adopted time intervals, and set the pth non-adopted time interval T p ∈(min p ,max p ), the p+1th unadopted time interval T p+1 ∈(min p+1 ,max p+1 ) and the p-1th unadopted time interval T p-1 ∈(min p-1 ,max p-1 ),like When , the amount of adopted information to be excluded when excluding unadopted information is set to x, according to the formula: Calculate the amount of adopted information x that needs to be excluded when excluding unadopted information; adjust the j-th sensor to transmit information to the information fusion model every information transmission amount t, where, 5. A multi-sensor information fusion system applied to a multi-sensor information fusion method based on blockchain according to any one of claims 1 to 4, characterized in that: The fusion system includes a division module, an expected target setting module, a target confirmation module and an adjustment module; The division module is used to set an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fusion information; generate a sensor information group according to the information source of each historical fusion information event in the information fusion model; and classify the historical information fusion events by comparing the sensor information of any two groups of sensor information groups; The expected target setting module is used to obtain the information transmission amount transmitted by each sensor to the information fusion model and the information adoption amount of the information fusion model for each information fusion event in the history of similar information fusion events, and calculate the information adoption rate of each sensor; according to the size of the information adoption rate, each sensor in an information fusion event is divided into key sensors and related sensors, and according to the historical information adoption rates of key sensors and related sensors, the expected information adoption rates of key sensors and related sensors in an information fusion event are obtained; The target confirmation module is used to obtain the actual information adoption rate of each sensor according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model when the multiple sensors are collecting information; and determine the expected information adoption rate of the multiple sensors according to the sensor information of each sensor and the key sensor; The adjustment module is used to adjust the information transmission process transmitted by each sensor to the information fusion model by comparing the actual information adoption rate of each sensor with the expected information adoption rate.

6. The multi-sensor information fusion system according to claim 5, characterized in that: The division module includes a sensor information group generation unit and a classification unit; The sensor information group generating unit is used to set an information fusion event as a process of putting the information collected by the sensor into the information fusion model for feature extraction and fusion to obtain fused information; generate a sensor information group according to the information source of each historical fusion information event in the information fusion model; the classification unit is used to classify the historical information fusion events by comparing the sensor information of any two groups of sensor information groups.

7. The multi-sensor information fusion system according to claim 5, characterized in that: The expected target setting module includes an information adoption rate calculation unit, a sensor division unit and an expected information adoption rate calculation unit; The information adoption rate calculation unit is used to obtain the amount of information transmitted by each sensor to the information fusion model and the amount of information adopted by the information fusion model for each information fusion event in historical similar information fusion events, and calculate the information adoption rate of each sensor; the sensor classification unit is used to classify each sensor in an information fusion event into key sensors and related sensors according to the size of the information adoption rate; the expected information adoption rate calculation unit is used to determine the expected information adoption rate of the several sensors based on the sensor information of each sensor and the key sensors.

8. The multi-sensor information fusion system according to claim 5, characterized in that: The target confirmation module includes an actual information adoption rate unit and an expected information adoption rate confirmation unit; The actual information adoption rate unit is used to obtain the actual information adoption rate of each sensor according to the actual information transmission amount transmitted by each sensor to the information fusion model and the actual information adoption amount of the information fusion model when several sensors collect information; the expected information adoption rate confirmation unit is used to determine the expected information adoption rate of the several sensors according to the sensor information of each sensor and the key sensors.

Images