A logistics vehicle monitoring method and system based on feature fusion recognition

Abstract

The application acquires to-be-fused monitoring features, including the monitoring of the logistics vehicles that will execute the monitoring strategy in the middle logistics stage to the final logistics stage, the initial logistics stage is prior to the middle logistics stage, and the middle logistics stage is prior to the final logistics stage; generates the monitoring strategy for the monitoring of the logistics vehicles that will execute the monitoring strategy in the middle logistics stage to the final logistics stage according to the to-be-fused monitoring features; acquires the logistics vehicles that will execute the monitoring strategy between the middle logistics stage and the final logistics stage in the initial logistics stage, and generates the monitoring strategy for these logistics vehicles, that is, can monitor the logistics vehicles that execute the monitoring strategy, and updates the monitoring strategy for the logistics vehicles that execute the monitoring strategy. Since the application does not need to update the monitoring strategy for all historical active users, reduces the number of users that need to update the monitoring strategy, saves the computing resources, and can update the monitoring strategy for the logistics vehicles that will execute the monitoring strategy in time.

Description

Technical Field

[0001] This invention relates to the field of Internet technology, and in particular to a method and system for monitoring logistics vehicles based on feature fusion recognition. Background Technology

[0002] In recent years, with the rapid development of the social economy and third-party logistics, cargo transportation safety has increasingly become a crucial issue for enterprises and logistics companies. By outsourcing their logistics operations to third-party logistics contractors, companies can focus their time and energy on their core businesses, improving supply chain management and operational efficiency. Enterprise logistics supervision departments have a responsibility and a necessity to take measures to strengthen security precautions during cargo transportation. One effective approach is to fully utilize GPS technology and modern information technology to detect and eliminate potential security problems before they occur. Logistics information technology plays a vital role in assisting management and providing decision analysis.

[0003] Given the current state of the domestic logistics industry, in order to reduce logistics costs, manufacturing companies will inevitably choose multiple third-party logistics companies to handle their logistics outsourcing business. Even within the same logistics company, their vehicles may be equipped with GPS terminal products from different GPS service providers. In order to monitor the operational status of these vehicles, the company's logistics department will inevitably install GPS monitoring and command systems from each GPS operator. This leads to chaos in logistics management and makes the statistics of logistics data extremely cumbersome, making it impossible to have a clear understanding of logistics information and the status of transport vehicles. Summary of the Invention

[0004] This invention provides a logistics vehicle monitoring method and system based on feature fusion recognition. It can monitor logistics vehicles that execute monitoring strategies and update the monitoring strategies for the monitored logistics vehicles that execute the monitoring strategies. This reduces the number of users who need to update the monitoring strategies, saves computing resources, and can update the monitoring strategies for logistics vehicles that will execute the monitoring strategies in a timely manner.

[0005] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0006] In a first aspect, embodiments of the present invention provide a logistics vehicle monitoring method based on feature fusion recognition. The method includes: in the initial logistics stage, acquiring monitoring features to be fused, wherein the monitoring features to be fused include logistics vehicles monitored that will execute a monitoring strategy during the intermediate logistics stage to the final logistics stage, the initial logistics stage taking precedence over the intermediate logistics stage, and the intermediate logistics stage taking precedence over the final logistics stage; and generating a monitoring strategy for the logistics vehicles monitored that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage based on the monitoring features to be fused.

[0007] In conjunction with the first aspect, in a first possible implementation of the first aspect, before acquiring the monitoring features to be integrated in the initial logistics stage, the method further includes: acquiring alternative monitoring features to be integrated, wherein the alternative monitoring features to be integrated include logistics vehicles that will execute monitoring strategies within a preset stage and the execution cycle, the preset stage being the period for acquiring the alternative monitoring features to be integrated, the execution cycle being the period during which the logistics vehicles that will execute monitoring strategies execute monitoring strategies within the preset stage, and the initial logistics stage, the intermediate logistics stage, and the final logistics stage being within the preset stage.

[0008] In conjunction with the first possible implementation of the first aspect, in the second possible implementation of the first aspect, the step of acquiring the monitoring features to be integrated in the initial logistics stage specifically includes: in the initial logistics stage, reading the logistics vehicles whose execution cycle will execute the monitoring strategy from the intermediate logistics stage to the final logistics stage.

[0009] In conjunction with the first possible implementation of the first aspect, in the third possible implementation of the first aspect, the step of acquiring the candidate monitoring features to be fused specifically includes: acquiring a monitoring model, wherein the monitoring model is used to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period; inputting the parameters of each logistics vehicle into the monitoring model respectively, and calculating the probability that each logistics vehicle will execute the monitoring strategy within the preset period; if the probability that the first logistics vehicle will execute the monitoring strategy within the preset period is greater than the preset probability, then the first logistics vehicle is set as the logistics vehicle that will execute the monitoring strategy within the preset period, wherein the first logistics vehicle is any one of the logistics vehicles.

[0010] Combining the third possible implementation of the first aspect, in the fourth possible implementation of the first aspect, monitoring

[0011]

[0012] Where y(x) is the probability that the first logistics vehicle executes the monitoring strategy; β0, β1...β n The parameters of the monitoring model are: x1...x n These are the parameters of the first logistics vehicle.

[0013] In conjunction with the first aspect, in the fifth possible implementation of the first aspect, after generating the monitoring strategy for the logistics vehicle based on the monitoring features to be fused, the method further includes: before or during the intermediate logistics stage, sending the monitoring strategy to the monitored logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage; or, during the intermediate logistics stage to the final logistics stage, receiving a monitoring strategy request message sent by the monitored logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage, and sending the monitoring strategy to the monitored logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage.

[0014] Secondly, embodiments of the present invention provide a logistics vehicle monitoring system based on feature fusion recognition. The system includes: a first acquisition module, configured to acquire monitoring features to be fused during the initial logistics stage, wherein the monitoring features to be fused include logistics vehicles monitored that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage, wherein the initial logistics stage takes precedence over the intermediate logistics stage, and the intermediate logistics stage takes precedence over the final logistics stage; and a generation module, configured to generate the monitoring strategy for the logistics vehicles monitored that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage, based on the monitoring features to be fused.

[0015] In conjunction with the second aspect, in the first possible implementation of the second aspect, the system further includes: a second acquisition module, used by the first acquisition module to acquire candidate monitoring features to be integrated before acquiring the monitoring features to be integrated in the initial logistics stage, wherein the candidate monitoring features to be integrated include logistics vehicles that will execute monitoring strategies within a preset stage and the execution cycle, the preset stage being the period for acquiring the candidate monitoring features to be integrated, the execution cycle being the period during which the logistics vehicles that will execute monitoring strategies execute monitoring strategies within the preset stage, and the initial logistics stage, the intermediate logistics stage, and the final logistics stage being within the preset stage.

[0016] In conjunction with the first possible implementation of the second aspect, in the second possible implementation of the second aspect, the first acquisition module is specifically used to: in the initial logistics stage, read the logistics vehicles whose execution cycle will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage.

[0017] In conjunction with the first possible implementation of the second aspect, in the third possible implementation of the second aspect, the second acquisition module specifically includes: an acquisition submodule, used to acquire a monitoring model, wherein the monitoring model is used to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period; a calculation submodule, used to input the parameters of each logistics vehicle into the monitoring model respectively, and calculate the probability that each logistics vehicle will execute the monitoring strategy within the preset period; and a setting submodule, used to set the first logistics vehicle as the logistics vehicle that will execute the monitoring strategy within the preset period if the probability that the first logistics vehicle will execute the monitoring strategy within the preset period is greater than the preset probability, wherein the first logistics vehicle is any one of the logistics vehicles.

[0018] Combining the third possible implementation of the second aspect, in the fourth possible implementation of the second aspect, the monitoring model is as follows:

[0019] Where y(x) is the probability that the first logistics vehicle executes the monitoring strategy; β0, β1...β n The parameters of the monitoring model are: x1...x n These are the parameters of the first logistics vehicle.

[0020] In conjunction with the second aspect, in a fifth possible implementation of the second aspect, the system further includes: a receiving module, configured to receive a monitoring strategy request message sent by a logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage; and a sending module, configured to generate the monitoring strategy for the logistics vehicle based on the monitoring features to be fused, and then send the monitoring strategy to the logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage before or during the intermediate logistics stage; or, during the intermediate logistics stage to the final logistics stage, after the receiving module receives the monitoring strategy request message sent by the logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage, send the monitoring strategy to the logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage.

[0021] This invention provides a method and system for monitoring logistics vehicles based on feature fusion recognition. The method includes: in the initial logistics stage, acquiring monitoring features to be fused, wherein the monitoring features to be fused include logistics vehicles that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage, with the initial logistics stage taking precedence over the intermediate logistics stage, and the intermediate logistics stage taking precedence over the final logistics stage; and generating monitoring strategies for the logistics vehicles that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage based on the monitoring features to be fused.

[0022] Based on the description of the above embodiments, the present invention acquires logistics vehicles that will implement monitoring strategies from the initial logistics stage to the final logistics stage, generates monitoring strategies for these logistics vehicles, and updates the monitoring strategies for these vehicles. Since the technical solution of the present invention does not require updating monitoring strategies for all historically active users, it reduces the number of users requiring monitoring strategy updates, saves computing resources, and enables timely updates to monitoring strategies for logistics vehicles that will implement them. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings required for implementation in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart illustrating a logistics vehicle monitoring method based on feature fusion recognition provided in this embodiment of the invention. Figure 1 ;

[0025] Figure 2 A flowchart illustrating a logistics vehicle monitoring method based on feature fusion recognition provided in this embodiment of the invention. Figure 2 ;

[0026] Figure 3 A schematic diagram of the structure of a logistics vehicle monitoring system based on feature fusion recognition provided in an embodiment of the present invention. Figure 1 . Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Example 1

[0029] This invention provides a method for monitoring logistics vehicles based on feature fusion recognition, such as... Figure 1 The diagram shown illustrates the process of this method, including:

[0030] S101. The logistics vehicle monitoring system based on feature fusion recognition acquires the monitoring features to be fused during the initial logistics stage.

[0031] Among them, the monitoring features to be integrated include logistics vehicles that will implement monitoring strategies from the intermediate logistics stage to the final logistics stage, with the initial logistics stage taking priority over the intermediate logistics stage, and the intermediate logistics stage taking priority over the final logistics stage.

[0032] Specifically, the initial logistics phase is the period during which monitoring strategies will be implemented for logistics vehicles from the intermediate logistics phase to the final logistics phase. Since calculating the monitoring strategy requires a certain number of stages, the initial logistics phase takes precedence over the intermediate logistics phase to ensure that the calculation process of the monitoring strategy is completed before the intermediate logistics phase.

[0033] S102. The logistics vehicle monitoring system based on feature fusion recognition generates monitoring strategies for logistics vehicles that will execute monitoring strategies from the intermediate logistics stage to the final logistics stage, based on the monitoring features to be fused.

[0034] Specifically, the feature fusion-based logistics vehicle monitoring system generates monitoring strategies for logistics vehicles that will implement monitoring strategies during the intermediate to final logistics stages. Essentially, this system only updates monitoring strategies for the monitored vehicles it anticipates will execute those strategies, significantly reducing the computational load, shortening the computation cycle, and conserving computational resources.

[0035] For example, the monitoring strategy for logistics vehicles that will be monitored between 18:30 on December 9, 2022 (i.e., the intermediate logistics stage in this scheme) and 23:00 on December 9, 2022 (i.e., the final logistics stage in this scheme) is updated. The initial logistics stage can be any period before 18:30 on December 9, 2022. Optionally, the initial logistics stage can be set to 15:00 on December 9, 2022. Then, the logistics vehicle monitoring system based on feature fusion recognition will generate monitoring strategies for logistics vehicles that will be monitored between 15:00 on the initial logistics stage (15:00 on December 9, 2022) and 18:30 on the intermediate logistics stage (18:30 on December 9, 2022) and 23:00 on the final logistics stage (23:00 on December 9, 2022).

[0036] It should be noted that the initial logistics stage, intermediate logistics stage and final logistics stage can be preset, and this invention does not impose any restrictions on this.

[0037] Specifically, the interval between the intermediate logistics stage and the final logistics stage can be set to a fixed interval, such as one week, two weeks, or other durations; more specifically, the interval between the intermediate logistics stage and the final logistics stage can also be variable, in which case the specific values ​​of the intermediate logistics stage and the final logistics stage can be determined based on the historical access volume of the logistics vehicles.

[0038] For example, the update cycle of the monitoring strategy can be determined based on the frequency with which logistics vehicles execute the monitoring strategy. For instance, logistics vehicles typically execute the monitoring strategy more frequently between 10 PM and midnight, so the update cycle can be set one hour shorter. Conversely, logistics vehicles execute the monitoring strategy less frequently between 1 AM and 3 AM, so the update cycle can be one hour longer. This reduces the computational resources required for the feature fusion-based logistics vehicle monitoring system to generate monitoring strategies for logistics vehicles.

[0039] It should be added that the monitoring strategy generation process is completed before the intermediate logistics stage.

[0040] This invention provides a logistics vehicle monitoring method based on feature fusion recognition. In the initial logistics stage, monitoring features to be fused are obtained, including logistics vehicles that will implement monitoring strategies during the intermediate logistics stage to the final logistics stage. The initial logistics stage takes precedence over the intermediate logistics stage, and the intermediate logistics stage takes precedence over the final logistics stage. Based on the monitoring features to be fused, monitoring strategies are generated for the logistics vehicles that will implement monitoring strategies during the intermediate logistics stage to the final logistics stage.

[0041] Based on the description of the above embodiments, the present invention acquires logistics vehicles that will implement monitoring strategies from the initial logistics stage to the final logistics stage, generates monitoring strategies for these logistics vehicles, and updates the monitoring strategies for these vehicles. Since the technical solution of the present invention does not require updating monitoring strategies for all historically active users, it reduces the number of users requiring monitoring strategy updates, saves computing resources, and enables timely updates to monitoring strategies for logistics vehicles that will implement them.

[0042] Example 2

[0043] This invention provides a method for monitoring logistics vehicles based on feature fusion recognition, such as... Figure 2 The diagram shown illustrates the process of this method, including:

[0044] S201. Obtain the alternative monitoring features to be integrated.

[0045] Among them, the candidate monitoring features to be integrated include the logistics vehicles that will execute the monitoring strategy within a preset stage and the execution cycle. The preset stage is the period for obtaining the candidate monitoring features to be integrated, and the execution cycle is the period during which the logistics vehicles that will execute the monitoring strategy will execute the monitoring strategy within the preset stage. The initial logistics stage, intermediate logistics stage and final logistics stage are within the preset stage.

[0046] Specifically, the preset phase includes the initial logistics phase, the intermediate logistics phase, and the final logistics phase. The alternative monitoring features to be integrated include the logistics vehicles monitored that execute the monitoring strategy in each cycle within the preset phase; that is, the alternative monitoring features to be integrated include detailed information on which logistics vehicles will execute the monitoring strategy in which cycles.

[0047] The preset phase is the period for acquiring candidate monitoring features to be merged. Optionally, the preset phase can be set to one month (i.e., four weeks), meaning the program calculating candidate monitoring features can run at any fixed period of four weeks each month. Preferably, the program calculating candidate monitoring features can run once a month at 3:00 AM on Monday, i.e., selecting a period with low logistics vehicle access volume for extensive calculations, which can make full use of resources.

[0048] Specifically, S201 includes S201a-S201c.

[0049] S201a, Obtain the monitoring model.

[0050] The monitoring model is used to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period.

[0051] Preferably, the monitoring model is

[0052] Where y(x) is the probability that the first logistics vehicle executes the monitoring strategy; β0, β1...β n The parameters of the monitoring model are: x1...x n These are the parameters of the first logistics vehicle.

[0053] It should be added that the monitoring model y(x)=1 / 1+e-(β0+β1x1+...+βnxn) is a logistic regression model. The technical solution of this invention can also use other monitoring models to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset stage, such as matrix models.

[0054] For example, the parameters of a logistics vehicle may include: user attributes, such as user identifier, gender, driving experience, user tag, etc.; equipment attributes, such as equipment identifier, equipment model; application attributes, such as application identifier, application tag, application category, etc.; context attributes, such as user location information, etc.; period attributes, such as the stage to which the information to be monitored belongs, whether the current stage is a holiday or a weekday, weekday attributes (Monday, Tuesday, etc.), etc.; and the attributes of the information to be monitored, such as information identifier, information category, information tag, etc. The parameters of a logistics vehicle may also include combinations of user attributes, equipment attributes, and attributes of the information to be monitored, such as generating a combined attribute "gender-information identifier" by combining the user's gender and information identifier.

[0055] Furthermore, the monitoring model is used to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period. Specifically, it calculates the probability that historically active logistics vehicles marked in the database will execute the monitoring strategy within the preset period. The selection of historically active users can be selected from those within the last three months or those within the last month. The specific selection rules for historically active users can be preset, and this invention does not impose any restrictions on this.

[0056] For example, taking the logistic regression model y(x)=1 / 1+e-(β0+β1x1+...+βnxn) as an example, the process of obtaining the monitoring model is as follows:

[0057] The logistic regression model is: y(x)=1 / 1+e-(β0+β1x1+...+βnxn). Substituting the training data into this logistic regression model, we obtain the parameters β0, β1…βn of the logistic regression model.

[0058] It should be added that the training data is used to obtain the various parameters in the monitoring model. For example, in this embodiment of the invention, a logistic regression model is used to obtain the candidate monitoring features to be fused, and the parameters in the logistic regression model are trained from the training data.

[0059] It should be noted that the training data is obtained by combining the monitoring results of logistics vehicles implementing the monitoring strategy during a certain past period with the actual results of logistics vehicles implementing the monitoring strategy during that past period. Here, we will take the past two months (between 1:00 AM on October 19, 2022 and 1:00 AM on December 9, 2022) as an example for explanation.

[0060] Specifically, the selection of a certain period in the past can be set according to the actual situation.

[0061] For example, the monitoring results of the logistics vehicle monitoring system based on feature fusion recognition for the logistics vehicles that will be monitored during the past two months (between 1:00 AM on November 16, 2022 and 1:00 AM on December 9, 2022) are as follows:

[0062] driverid1, monitorstrategy1,6, sunday

[0063] driverid1,monitorstrategy1,18,sunday

[0064] driverid2,monitorstrategy1,20,sunday

[0065] driverid1,monitorstrategy1,6,monday

[0066] driverid1,monitorstrategy1,18,monday

[0067] driverid2,monitorstrategy1,21,monday

[0068] ...

[0069] Specifically, the meaning of the monitoring results for logistics vehicles that will be monitored between 1:00 AM on November 16, 2022 and 1:00 AM on December 9, 2022 is illustrated using the first line of data as an example. The first line of data consists of driverid1, monitorstrategy1, 6, and Sunday. Here, driverid1 is the user identifier, monitorstrategy1 is the application identifier, and 6 and Sunday are the periodic attributes of the logistics vehicle's access to the application. This line of data indicates that the logistics vehicle driverid1 will access the application monitorstrategy1 for 6 hours on Sunday.

[0070] For example, the results of the actual implementation of the monitoring policy for logistics vehicles between 1:00 AM on November 16, 2022 and 1:00 AM on December 9, 2022 are shown below:

[0071] driverid1,monitorstrategy1,6,sunday

[0072] driverid1,monitorstrategy1,18,sunday

[0073] driverid2,monitorstrategy1,20,sunday

[0074] driverid1,monitorstrategy1,18,monday

[0075] driverid2,monitorstrategy1,21,monday

[0076] ...

[0077] Specifically, the results of the monitoring policy actually executed by logistics vehicles between 1:00 AM on November 16, 2022 and 1:00 AM on December 9, 2022 are shown below. The meaning of the data is illustrated using the first row as an example: driverid1, monitorstrategy1, 6, sunday. Here, driverid1 is the user identifier, monitorstrategy1 is the application identifier, and 6 and sunday are the periodic attributes of the application execution by the logistics vehicle. This row indicates that logistics vehicle driverid1 accessed the application monitorstrategy1 for 6 hours on Sunday.

[0078] For example, training data is obtained based on the monitoring results of logistics vehicles that will implement the monitoring strategy between 1:00 AM on November 16, 2022 and 1:00 AM on December 9, 2022, and the logistics vehicles that actually implemented the monitoring strategy between 1:00 AM on November 16, 2022 and 1:00 AM on December 9, 2022, as shown below:

[0079] driver^driverid1,monitorstrategy^monitorstrategy1,time^6,week^sunday,1

[0080] driver^driverid1,monitorstrategy^monitorstrategy1,time^18,week^sunday,1

[0081] driver^driverid2,monitorstrategy^monitorstrategy1,time^20,week^sunday,1

[0082] driver^driverid1,monitorstrategy^monitorstrategy1,time^6,week^monday,0

[0083] driver^driverid1,monitorstrategy^monitorstrategy1,time^18,week^monday,1

[0084] driver^driverid2,monitorstrategy^monitorstrategy1,time^21,week^monday,1

[0085] ……

[0086] Specifically, each row of data represents a training data point. Taking the first training data point, driver^driverid1, monitor strategy^monitorstrategy1, time^6, week^sunday, 1, as an example, we can see that the logistics vehicle monitoring system based on feature fusion identification detects that logistics vehicle driverid1 will access the application monitorstrategy1 during the 6-hour period on Sunday. In fact, logistics vehicle driverid1 did access the application monitorstrategy1 during the 6-hour period on Sunday. Therefore, this data point is recorded as a positive example in the training data and marked as 1. Conversely, the fourth training data point, driver^driverid1, monitor strategy^monitorstrategy1, time^6, week^monday, 0, indicates that the logistics vehicle monitoring system based on feature fusion identification detects that logistics vehicle driverid1 will access the application monitorstrategy1 during the 6-hour period on Saturday. However, logistics vehicle driverid1 did not access the application monitorstrategy1 during the 6-hour period on Saturday. Therefore, this data point is recorded as a negative example in the training data and marked as 0. Training data is generated according to this pattern. Based on the training data, calculate the parameters β0, β1...βn in the monitoring model.

[0087] It should be added that the application attributes are included in the training data in order to distinguish the probability that logistics vehicles will execute monitoring strategies by accessing different applications, so that the logistics vehicle monitoring system based on feature fusion recognition can generate different monitoring strategies when logistics vehicles access different applications.

[0088] Optionally, the application attributes may not be included in the training data, so that the logistics vehicle monitoring system based on feature fusion recognition can generate the same monitoring strategy when the logistics vehicle accesses different applications.

[0089] It should be noted that for logistics vehicles that have not previously implemented monitoring strategies, there are no behavioral attributes, but unlabeled test data can be constructed based on other attributes in the parameters of the logistics vehicles.

[0090] In practice, training data will integrate various types of data, including user attributes; device attributes; application attributes; information attributes; context attributes; periodic attributes; user behavior attributes, such as whether the user clicked on the information displayed to them; and also combinations of user attributes, device attributes, application attributes, information attributes, context attributes, periodic attributes, and user behavior attributes.

[0091] S201b: Input the parameters of each logistics vehicle into the monitoring model and calculate the probability that each logistics vehicle will execute the monitoring strategy within the preset stage.

[0092] The specific function of the monitoring model here is to monitor the probability of a particular logistics vehicle executing a monitoring strategy within a certain period. That is, firstly, based on the training data, the parameters of the monitoring model are calculated; the monitoring model is...

[0093] Where y(x) is the probability that the first logistics vehicle executes the monitoring strategy; β0, β1...β n The parameters of the monitoring model are: x1...x n These are the parameters of the first logistics vehicle.

[0094] Furthermore, the parameters of logistics vehicles in the training data are used as the values ​​of each x in the monitoring model.

[0095] Specifically, the methods for generating parameters of the monitoring model based on training data are mainly divided into offline batch generation and online real-time generation. Offline batch generation refers to calculating the parameters of the monitoring model after obtaining all the training data.

[0096] Specifically, unlabeled test data is constructed based on the parameters of each logistics vehicle. This unlabeled test data is then fed into the monitoring model to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset phase. The format of the unlabeled test data is as follows:

[0097] driverid1,monitorstrategy1,6,tuesday

[0098] driverid1,monitorstrategy1,18,tuesday

[0099] ...

[0100] For example, based on unlabeled test data and a pre-calculated monitoring model, the probability of monitoring vehicle driverid1 executing the monitoring strategy in each hourly phase of the monitoring cycle is used to obtain candidate monitoring features to be fused. Substituting the first row of unlabeled data into the monitoring model yields the probability that the logistics vehicle will execute the monitoring strategy by applying monitor strategy1 in the 6-hour phase corresponding to Tuesday.

[0101] Using the above method, the probability of each logistics vehicle executing monitoring strategies by accessing different applications is calculated for each cycle hour within a preset phase.

[0102] S201c: If the probability that the first logistics vehicle will execute the monitoring strategy within the preset period is greater than the preset probability, then the first logistics vehicle is set as the logistics vehicle that will execute the monitoring strategy within the preset period.

[0103] The first logistics vehicle is any one of the various logistics vehicles.

[0104] Optionally, logistics vehicles with a probability greater than a preset probability of executing the monitoring strategy in a certain period are selected as the logistics vehicles that will execute the monitoring strategy in that period. The preset probability can be set in advance, and this invention does not limit it.

[0105] S202. In the initial logistics stage, acquire the monitoring features to be integrated.

[0106] The monitoring features to be integrated include logistics vehicles that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage, with the initial logistics stage taking precedence over the intermediate logistics stage, and the intermediate logistics stage taking precedence over the final logistics stage.

[0107] Specifically, in the initial logistics phase, the monitoring strategies for logistics vehicles whose execution cycles span from the intermediate logistics phase to the final logistics phase are read from the candidate monitoring features to be integrated.

[0108] S203. Based on the monitoring characteristics to be integrated, generate monitoring strategies for the logistics vehicles that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage.

[0109] Furthermore, after the logistics vehicle monitoring system based on feature fusion recognition obtains the monitoring features to be fused, it will generate a monitoring strategy for the logistics vehicles in the monitoring features to be fused.

[0110] Specifically, the process by which a feature fusion-based logistics vehicle monitoring system generates monitoring strategies for logistics vehicles that will implement monitoring strategies from the initial logistics stage to the final logistics stage is as follows:

[0111] (1) Obtain the parameters of the logistics vehicles that will implement the monitoring strategy from the intermediate logistics stage to the final logistics stage, and form unlabeled test data.

[0112] For example, the unlabeled test data, which is composed of the attributes of the logistics vehicle and the attributes of the information, can be: driverid1, 20, itemid1, chemical transport logistics vehicle; the meaning of this unlabeled test data is: driverid1 is the user identifier, 20 is the user's driving experience, itemid1 is the information identifier, and chemical transport logistics vehicle is the information category.

[0113] Based on the monitoring model and unlabeled test data, a monitoring strategy is generated for the logistics vehicles that will be monitored during the intermediate and final logistics stages.

[0114] For example, the parameters of the logistics vehicle in the unlabeled test data: driverid1, 20, itemid1, chemical transport logistics vehicle; are input into the monitoring model to calculate the probability that the logistics vehicle will execute this information during the intermediate logistics stage to the final logistics stage.

[0115] The probability of the logistics vehicle executing each piece of information is calculated using the above method, and the top n pieces of information with higher probabilities are monitored for the logistics vehicle.

[0116] Then, for each logistics vehicle that will implement the monitoring strategy during the intermediate logistics stage to the final logistics stage, the probability of each logistics vehicle executing each piece of information is calculated, and the top n pieces of information with the highest execution probability for each logistics vehicle are monitored and sent to the corresponding logistics vehicle. The size of n can be preset.

[0117] It should be noted that, in order to ensure the real-time updating of monitoring strategies for logistics vehicles by the feature fusion recognition-based logistics vehicle monitoring system, the update cycle of the monitoring strategy can be set to a relatively short period, such as one week, and even shorter when user traffic is high. Since the technical solution of this invention only needs to update the monitoring strategy for logistics vehicles that will execute the monitoring strategy during the intermediate to final logistics stages, setting a short update cycle is entirely feasible, thus ensuring the real-time updating of equipment monitoring.

[0118] S204. Logistics vehicles that will implement monitoring strategies during the intermediate logistics phase to the final logistics phase will obtain updated monitoring strategies.

[0119] Specifically, S204 can be S204a or S204b.

[0120] S204a. Before or during the intermediate logistics phase, send the monitoring policy to the monitored logistics vehicles that will execute the monitoring policy from the intermediate logistics phase to the final logistics phase.

[0121] S204b: During the intermediate logistics stage to the final logistics stage, receive monitoring policy request messages from logistics vehicles that will execute monitoring policies during the intermediate logistics stage to the final logistics stage, and send monitoring policies to the logistics vehicles that will execute monitoring policies during the intermediate logistics stage to the final logistics stage.

[0122] It should be noted that the logistics vehicle monitoring system based on feature fusion recognition will generate monitoring strategies for logistics vehicles during the initial logistics stage, intermediate logistics stage and final logistics stage. After the monitoring strategy is generated, there are two ways for the logistics vehicle to obtain the updated monitoring strategy, as shown in S204a and S204b.

[0123] Specifically, S204a involves a logistics vehicle monitoring system based on feature fusion recognition sending monitoring strategies to the monitored logistics vehicles that will execute the monitoring strategies between the intermediate and final logistics stages, either before or during the intermediate logistics stage.

[0124] Specifically, S205b involves a logistics vehicle monitoring system based on feature fusion recognition receiving monitoring policy request messages from logistics vehicles that will execute monitoring policies during the intermediate and final logistics stages, and sending monitoring policies to those logistics vehicles.

[0125] The implementation of the technical solution of this invention will be described below using a specific application scenario. Any four weeks in a month are denoted as 0-24. Assume a preset phase of four weeks, with a starting period of 3 hours, an initial logistics phase of 12 hours, a mid-logistics phase of 13 hours, and a final logistics phase of 14 hours. The monitoring strategy update cycle is set to a fixed one-week cycle. During the 3-hour period, the system monitors which users will execute the monitoring strategy and when over the next four weeks. During the initial logistics phase (12 hours), the monitoring strategy is updated for users who will execute the monitoring strategy between the 13-hour mid-logistics phase and the 14-hour final logistics phase.

[0126] Step 1: Within a 3-hour period, calculate the logistics vehicles that will implement the monitoring strategy and the cycle of each logistics vehicle implementing the monitoring strategy during the period from the current 3 hours to the next 3 hours.

[0127] Specifically, the parameter β of the monitoring model is calculated based on the training data, and then the probability of logistics vehicles in the monitoring database executing the monitoring strategy in a certain period is determined based on the obtained monitoring model. Logistics vehicles with a probability greater than the preset probability are identified as logistics vehicles that will execute the monitoring strategy in a certain period.

[0128] Step 2: In the initial logistics phase, identify the logistics vehicles that will be monitored from the intermediate logistics phase to the final logistics phase, and generate monitoring strategies for these logistics vehicles.

[0129] Specifically, the process of generating monitoring strategies for logistics vehicles that will be monitored during the intermediate and final logistics stages is as follows:

[0130] Based on the parameters of logistics vehicles, unlabeled test data is generated. The probability of logistics vehicle execution information is calculated based on the monitoring model and the unlabeled test data. The top n pieces of information with higher probabilities are then monitored and sent to the user.

[0131] The parameters of the monitoring model described above can be calculated using the training data shown below.

[0132] driverid1, 20, itemid1, chemical transport logistics vehicle, 1

[0133] driverid1, 20, itemid2, health, 0

[0134] driverid2, 60, itemid1, chemical transport logistics vehicle, 1

[0135] driverid2, 60, itemid2, health, 1

[0136] ...

[0137] The meaning of the data is explained using the first row as an example. The first row contains driverid1, 20, itemid1, "Chemical Transport Logistics Vehicle", 1. Here, driverid1 represents the user identifier, 20 represents the user's driving experience, itemid1 is the information identifier, "Chemical Transport Logistics Vehicle" represents the information category, and 1 represents the user's behavioral characteristic, specifically indicating that the user monitored the information and viewed it. This data constitutes a positive example in the training data. Similarly, the second row contains driverid1, 20, itemid2, "Health Care", 0. Here, driverid1, 20, itemid2, and "Health Care" have the same meaning as the first row. In this data, 0 indicates that the user monitored the information but did not view it, thus this data constitutes a negative example in the training data.

[0138] Based on this training data, the parameters of the monitoring model are calculated. This monitoring model can also execute a logistic regression model: y(x) = 1 / (1+e-(β0+β1x1+...+βnxn), where x represents the parameters of the logistics vehicle, and y(x) represents the probability of the user executing the monitoring strategy. To ensure real-time performance, the training of each parameter of the monitoring model can employ an incremental learning optimization method. This involves updating the parameters of the old monitoring model using newly added training data. The FTRL online learning algorithm mentioned earlier is an example of an incremental learning optimization method.

[0139] Then, based on the parameters of the logistics vehicles that will implement the monitoring strategy during the intermediate logistics stage to the final logistics stage, unlabeled test data is constructed. Based on the calculated monitoring model and the unlabeled test data, the probability of the logistics vehicle executing a certain information is calculated, and n pieces of information with a high probability of the logistics vehicle executing a certain information are monitored for the logistics vehicle.

[0140] It should be noted that the monitoring model for generating monitoring strategies for logistics vehicles can also use other models besides the logistic regression model mentioned above, such as matrix models.

[0141] Furthermore, when the monitoring model used to generate monitoring strategies for logistics vehicles is a logistic regression model within a linear model, the calculation cycle for the probability of a logistics vehicle executing a certain piece of information based on this monitoring model can be estimated. That is, based on the data of the logistics vehicles that will execute the monitoring strategy and the cycle for updating the monitoring strategy for each logistics vehicle, the computational resources required by the feature fusion-based logistics vehicle monitoring system for updating the monitoring strategy for those vehicles can be determined. Then, resources can be requested through cluster management tools, thus enabling effective resource utilization. Computational resources can be, but are not limited to, monitoring server resources and / or memory resources.

[0142] For example, if a monitoring server can update the monitoring policies of 100 logistics vehicles per minute, and the monitoring policies of 1,000 logistics vehicles need to be updated within one minute, then 10 monitoring servers need to be requested from the system; if each logistics vehicle requires 1M of memory to complete the calculation of the monitoring policy, then 1,000 logistics vehicles need to request 1,000M of memory.

[0143] This invention provides a logistics vehicle monitoring method based on feature fusion recognition. In the initial logistics stage, monitoring features to be fused are obtained, including logistics vehicles that will implement monitoring strategies during the intermediate logistics stage to the final logistics stage. The initial logistics stage takes precedence over the intermediate logistics stage, and the intermediate logistics stage takes precedence over the final logistics stage. Based on the monitoring features to be fused, monitoring strategies are generated for the logistics vehicles that will implement monitoring strategies during the intermediate logistics stage to the final logistics stage.

[0144] Based on the description of the above embodiments, the present invention acquires logistics vehicles that will implement monitoring strategies from the initial logistics stage to the final logistics stage, generates monitoring strategies for these logistics vehicles, and updates the monitoring strategies for these vehicles. Since the technical solution of the present invention does not require updating monitoring strategies for all historically active users, it reduces the number of users requiring monitoring strategy updates, saves computing resources, and enables timely updates to monitoring strategies for logistics vehicles that will implement them.

[0145] Example 3

[0146] This invention provides a logistics vehicle monitoring system based on feature fusion recognition, such as... Figure 3 The diagram shown is a structural schematic of the system, including:

[0147] The first acquisition module 10 is used to acquire monitoring features to be integrated during the initial logistics stage. The monitoring features to be integrated include logistics vehicles that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage. The initial logistics stage takes precedence over the intermediate logistics stage, and the intermediate logistics stage takes precedence over the final logistics stage.

[0148] The generation module 11 is used to generate monitoring strategies for logistics vehicles that will execute monitoring strategies from the intermediate logistics stage to the final logistics stage, based on the monitoring characteristics to be integrated.

[0149] The system also includes:

[0150] The second acquisition module is used to acquire alternative monitoring features to be integrated before acquiring the monitoring features to be integrated in the initial logistics stage. The alternative monitoring features to be integrated include the logistics vehicles that will execute the monitoring strategy within a preset stage and the execution cycle. The preset stage is the period for acquiring the alternative monitoring features to be integrated, and the execution cycle is the period during which the logistics vehicles that will execute the monitoring strategy will execute the monitoring strategy within the preset stage. The initial logistics stage, intermediate logistics stage and final logistics stage are within the preset stage.

[0151] The first acquisition module is specifically used in the initial logistics stage to read the logistics vehicles whose execution cycle is from the intermediate logistics stage to the final logistics stage and which will implement the monitoring strategy.

[0152] The second acquisition module specifically includes:

[0153] The acquisition submodule is used to acquire the monitoring model, which is used to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period.

[0154] The calculation submodule is used to input the parameters of each logistics vehicle into the monitoring model and calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period.

[0155] The setting submodule is used to set the first logistics vehicle as the logistics vehicle that will execute the monitoring strategy within the preset stage if the probability that the first logistics vehicle will execute the monitoring strategy within the preset stage is greater than the preset probability. Here, the first logistics vehicle is any one of the logistics vehicles.

[0156] The monitoring model is

[0157] Where y(x) is the probability that the first logistics vehicle executes the monitoring strategy; β0, β1...β n The parameters of the monitoring model are: x1...x n These are the parameters of the first logistics vehicle;

[0158] The system also includes:

[0159] The receiving module is used to receive monitoring strategy request messages sent by logistics vehicles that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage.

[0160] The sending module is used to generate a monitoring strategy for the logistics vehicles based on the monitoring characteristics to be integrated, and then send the monitoring strategy to the monitored logistics vehicles that will execute the monitoring strategy from the intermediate logistics stage to the final logistics stage, either before or during the intermediate logistics stage; or, during the intermediate logistics stage to the final logistics stage, the receiving module receives the monitoring strategy request message sent by the monitored logistics vehicles that will execute the monitoring strategy from the intermediate logistics stage to the final logistics stage, and sends the monitoring strategy to the monitored logistics vehicles that will execute the monitoring strategy from the intermediate logistics stage to the final logistics stage.

[0161] This invention provides a logistics vehicle monitoring system based on feature fusion recognition. A first acquisition module is used to acquire monitoring features to be fused during the initial logistics stage. These features include logistics vehicles that will execute monitoring strategies during the intermediate and final logistics stages, with priority given to the initial logistics stage over the intermediate logistics stage, and the intermediate logistics stage over the final logistics stage. A generation module is used to generate monitoring strategies for the logistics vehicles that will execute monitoring strategies during the intermediate and final logistics stages, based on the monitoring features to be fused.

[0162] Based on the description of the above embodiments, the present invention acquires logistics vehicles that will implement monitoring strategies from the initial logistics stage to the final logistics stage, generates monitoring strategies for these logistics vehicles, and updates the monitoring strategies for these vehicles. Since the technical solution of the present invention does not require updating monitoring strategies for all historically active users, it reduces the number of users requiring monitoring strategy updates, saves computing resources, and enables timely updates to monitoring strategies for logistics vehicles that will implement them.

[0163] Example 4

[0164] This invention provides a logistics vehicle monitoring system based on feature fusion recognition, comprising:

[0165] The processor is used to acquire monitoring features to be integrated during the initial logistics stage. These monitoring features include logistics vehicles that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage, with the initial logistics stage taking precedence over the intermediate logistics stage, and the intermediate logistics stage taking precedence over the final logistics stage. The processor is also used to generate monitoring strategies for the logistics vehicles that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage based on the monitoring features to be integrated.

[0166] The processor is also used to acquire alternative monitoring features to be integrated before acquiring the monitoring features to be integrated in the initial logistics stage. The alternative monitoring features to be integrated include the logistics vehicles that will execute the monitoring strategy within a preset stage and the execution cycle. The preset stage is the period for acquiring the alternative monitoring features to be integrated, and the execution cycle is the period during which the logistics vehicles that will execute the monitoring strategy execute the monitoring strategy within the preset stage. The initial logistics stage, intermediate logistics stage and final logistics stage are within the preset stage.

[0167] The processor is specifically used in the initial logistics phase to read the logistics vehicles whose execution cycle will be monitored during the intermediate logistics phase to the final logistics phase.

[0168] The processor is specifically used to acquire a monitoring model, wherein the monitoring model is used to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period; to input the parameters of each logistics vehicle into the monitoring model to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period; and to set the first logistics vehicle as the logistics vehicle that will execute the monitoring strategy within the preset period if the probability that the first logistics vehicle will execute the monitoring strategy within the preset period is greater than the preset probability, wherein the first logistics vehicle is any one of the logistics vehicles.

[0169] The monitoring model is

[0170] Where y(x) is the probability that the first logistics vehicle executes the monitoring strategy; β0, β1...β n The parameters of the monitoring model are: x1...x n These are the parameters of the first logistics vehicle.

[0171] A receiver is used to receive monitoring strategy request messages sent by logistics vehicles that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage.

[0172] The transmitter generates a monitoring strategy for the logistics vehicle based on the monitoring characteristics to be fused, and then sends the monitoring strategy to the monitored logistics vehicle that will execute the monitoring strategy from the intermediate logistics stage to the final logistics stage, either before or during the intermediate logistics stage; or, during the intermediate logistics stage to the final logistics stage, the receiver 21 receives a monitoring strategy request message sent by the monitored logistics vehicle that will execute the monitoring strategy from the intermediate logistics stage to the final logistics stage, and sends the monitoring strategy to the monitored logistics vehicle that will execute the monitoring strategy from the intermediate logistics stage to the final logistics stage.

[0173] The memory is used to store the attributes of logistics vehicles, including user attributes and information attributes.

[0174] This invention provides a logistics vehicle monitoring system based on feature fusion recognition. The processor is used to acquire monitoring features to be fused during the initial logistics stage. These features include logistics vehicles that will execute monitoring strategies during the intermediate and final logistics stages, with the initial logistics stage taking precedence over the intermediate logistics stage, and the intermediate logistics stage taking precedence over the final logistics stage. Based on the monitoring features to be fused, a monitoring strategy is generated for the logistics vehicles that will execute monitoring strategies during the intermediate and final logistics stages.

[0175] Those skilled in the art will understand that the contents disclosed herein can be varied and modified in many ways. For example, the various devices or components described above can be implemented in hardware, or in software, firmware, or a combination of some or all of the three.

[0176] This disclosure uses flowcharts to illustrate the steps of a method according to embodiments of this disclosure. It should be understood that the preceding or following steps are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes.

[0177] Those skilled in the art will understand that all or part of the steps in the above methods can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk. Optionally, all or part of the steps in the above embodiments can also be implemented by one or more integrated circuits. Accordingly, each module / unit in the above embodiments can be implemented in hardware or as a software functional module. This disclosure is not limited to any particular combination of hardware and software.

[0178] Unless otherwise defined, all terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. It should also be understood that terms such as those defined in a common dictionary should be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and not as having an idealized or highly formalized meaning, unless expressly defined herein.

[0179] The foregoing description is a description of the present disclosure and should not be construed as limiting it. Although several exemplary embodiments of the present disclosure have been described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novelty and advantages of the present disclosure. Therefore, all such modifications are intended to be included within the scope of the present disclosure as defined by the claims. It should be understood that the foregoing description is a description of the present disclosure and should not be construed as limiting it to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present disclosure is defined by the claims and their equivalents.

[0180] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0181] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A logistics vehicle monitoring method based on feature fusion recognition, characterized in that, include: In the initial logistics stage, monitoring features to be integrated are acquired, including logistics vehicles that will execute monitoring strategies from the intermediate logistics stage to the final logistics stage. The initial logistics stage takes precedence over the intermediate logistics stage, and the intermediate logistics stage takes precedence over the final logistics stage. Based on the monitoring features to be integrated, a monitoring strategy is generated for the logistics vehicles that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage. Before acquiring the monitoring features to be integrated during the initial logistics stage, the method further includes: Acquire candidate monitoring features to be integrated, wherein the candidate monitoring features to be integrated include the logistics vehicles that will execute the monitoring strategy within a preset stage and the execution cycle, the preset stage is the period for acquiring the candidate monitoring features to be integrated, the execution cycle is the period during which the logistics vehicles that will execute the monitoring strategy execute the monitoring strategy within the preset stage, and the initial logistics stage, the intermediate logistics stage and the final logistics stage are within the preset stage. In the initial logistics stage, acquiring the monitoring features to be integrated specifically includes: In the initial logistics phase, the logistics vehicles whose execution cycle will be used to implement the monitoring strategy during the intermediate logistics phase to the final logistics phase are read. The acquisition of candidate monitoring features to be merged specifically includes: Obtain a monitoring model, wherein the monitoring model is used to calculate the probability that each logistics vehicle will execute a monitoring strategy within a preset period; The parameters of each logistics vehicle are fed into the monitoring model to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period. If the probability that the first logistics vehicle will execute the monitoring strategy within a preset period is greater than the preset probability, then the first logistics vehicle is set as the logistics vehicle that will execute the monitoring strategy within the preset period, wherein the first logistics vehicle is any one of the logistics vehicles. 2.The method of claim 1, wherein, The monitoring model is ； Where y(x) is the probability that the first logistics vehicle executes the monitoring strategy; β0, β1...βn are the parameters of the monitoring model; x1...xn are the parameters of the first logistics vehicle, and the parameters of the logistics vehicle include a combination of user attributes, equipment attributes, and attributes of the information to be monitored. 3.The method of claim 1, wherein, After generating the monitoring strategy for the logistics vehicle based on the monitoring features to be fused, the method further includes: Before or during the intermediate logistics stage, the monitoring strategy is sent to the monitored logistics vehicles that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage. or, During the intermediate logistics stage to the final logistics stage, the system receives monitoring strategy request messages from logistics vehicles that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage, and sends the monitoring strategy to the logistics vehicles that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage.

4. A logistics vehicle monitoring system based on feature fusion recognition, characterized in that, include: The first acquisition module is used to acquire monitoring features to be integrated during the initial logistics stage. The monitoring features to be integrated include logistics vehicles that will execute monitoring strategies during the intermediate logistics stage to the final logistics stage. The initial logistics stage takes precedence over the intermediate logistics stage, and the intermediate logistics stage takes precedence over the final logistics stage. The generation module is used to generate the monitoring strategy for logistics vehicles that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage, based on the monitoring features to be integrated. The system also includes: The second acquisition module is used to acquire candidate monitoring features to be integrated before the first acquisition module acquires the monitoring features to be integrated in the initial logistics stage. The candidate monitoring features to be integrated include logistics vehicles that will execute monitoring strategies within a preset stage and the execution cycle. The preset stage is the period for acquiring the candidate monitoring features to be integrated. The execution cycle is the period during which the logistics vehicles that will execute monitoring strategies will execute monitoring strategies within the preset stage. The initial logistics stage, the intermediate logistics stage, and the final logistics stage are within the preset stage. The first acquisition module is specifically used for: In the initial logistics phase, the logistics vehicles whose execution cycle will be used to implement the monitoring strategy during the intermediate logistics phase to the final logistics phase are read. The second acquisition module specifically includes: The acquisition submodule is used to acquire the monitoring model, wherein the monitoring model is used to calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset stage; The calculation submodule is used to input the parameters of each logistics vehicle into the monitoring model and calculate the probability that each logistics vehicle will execute the monitoring strategy within a preset period. The configuration submodule is used to configure the first logistics vehicle as a logistics vehicle that will execute the monitoring strategy within the preset stage if the probability that the first logistics vehicle will execute the monitoring strategy within the preset stage is greater than the preset probability. The first logistics vehicle is any one of the logistics vehicles.

5. The feature fusion recognition based logistics vehicle monitoring system according to claim 4, characterized in that, The monitoring model is ； Where y(x) is the probability that the first logistics vehicle executes the monitoring strategy; β0, β1...βn are the parameters of the monitoring model; x1...xn are the parameters of the first logistics vehicle, and the parameters of the logistics vehicle include a combination of user attributes, equipment attributes, and attributes of the information to be monitored.

6. The feature fusion recognition based logistics vehicle monitoring system according to claim 5, wherein, The system also includes: The receiving module is used to receive monitoring strategy request messages sent by logistics vehicles that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage. The sending module is configured to generate the monitoring strategy for the logistics vehicle based on the monitoring features to be fused, and then send the monitoring strategy to the monitored logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage, either before or during the intermediate logistics stage; or, during the intermediate logistics stage to the final logistics stage, the receiving module receives a monitoring strategy request message sent by the monitored logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage, and then sends the monitoring strategy to the monitored logistics vehicle that will execute the monitoring strategy during the intermediate logistics stage to the final logistics stage.

Images