Moving target detection method and device based on unmanned platform

Abstract

The invention discloses a moving target detection method and device based on an unmanned platform. The method comprises the following steps: calculating a target detection demand change degree; calculating a maximum detection period according to the target detection demand change degree; calculating a target state set according to the current position of the moving target and the maximum detection period; according to the target state set, determining an unmanned platform combination, a next detection period and a grid position where the unmanned platform arrives; and according to the unmanned platform combination, the next detection period and the grid position where the unmanned platform arrives, moving target detection in the next period is carried out. According to the invention, moving target detection is realized, positioning precision and adaptability are improved, and energy consumption is reduced. The method can be widely applied to the technical field of target detection.

Description

Technical Field

[0001] The present invention relates to the field of target detection technology, and in particular to a method and device for detecting a moving target based on an unmanned driving platform. Background Art

[0002] Unmanned vehicle platform swarms perform real-time positioning and tracking of targets through cooperative sensing, path planning, and information sharing, emphasizing the need for high-precision target detection. Traditional target detection methods typically use fixed or mobile nodes to measure the distance or angle of a target, sending this distance or angle to a fusion center. This information, combined with the position coordinates of the detection nodes, is then used to calculate the target's position coordinates and velocity. However, the dynamic combination of the detection cycle and the location of the unmanned vehicle platforms results in insufficient positioning accuracy and low adaptability. Furthermore, the unmanned vehicles are powered by their own batteries, but their path planning accuracy is low, resulting in high energy consumption.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The embodiments of the present invention provide a method and device for detecting a moving target based on an unmanned driving platform, which effectively improves positioning accuracy and adaptability and reduces energy consumption.

[0005] In one aspect, an embodiment of the present invention provides a method for detecting a moving target based on an unmanned driving platform, comprising the following steps:

[0006] Calculate the target detection demand change;

[0007] Calculating a maximum detection period according to the target detection demand change degree;

[0008] Calculating a target state set according to the current position of the mobile target and the maximum detection period;

[0009] Determining, based on the target state set, an unmanned platform combination, a next detection cycle, and a grid position to be reached by the unmanned platform;

[0010] The next cycle of mobile target detection is performed according to the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform.

[0011] In some embodiments, calculating the target detection demand change degree includes:

[0012] Calculate the target direction change degree according to the moving speed and target moving rate of the moving target;

[0013] Calculating a target velocity change degree according to the target movement velocity;

[0014] Calculating a target change degree based on the target direction change degree and the target speed change degree;

[0015] The network change degree is calculated based on the link weights between the unmanned driving platforms, the network topology, and the total number of edges in the two network topologies at adjacent moments.

[0016] The target detection requirement change degree is calculated according to the weight coefficient, the target change degree and the network change degree.

[0017] In some embodiments, calculating the maximum detection period according to the target detection requirement change degree includes:

[0018] Calculating the pheromone at the current moment and the pheromone at the previous moment according to the target detection demand change degree;

[0019] Calculate the volatility coefficient based on the current pheromone and the previous pheromone;

[0020] The current pheromone is updated according to the volatility coefficient until the current pheromone is less than a preset volatility threshold, thereby obtaining the maximum detection period.

[0021] In some embodiments, the calculating the target state set according to the current position of the mobile target and the maximum detection period includes:

[0022] Dividing the maximum detection period according to a preset total number of periods to obtain a plurality of prediction periods;

[0023] Determining a current target state according to the current position of the moving target and the moving speed of the moving target;

[0024] The target state set is calculated according to the state transition matrix, the current target state and the multiple prediction cycles.

[0025] In some embodiments, determining the unmanned driving platform combination, the next detection cycle, and the grid position reached by the unmanned driving platform according to the target state set includes:

[0026] Rasterize the prediction space to obtain a rasterized spatial domain model;

[0027] Determining a reachable grid area based on the target state set, the detection radius, maximum flight speed, maximum lifting speed, maximum flight distance, horizontal navigation speed, and vertical lifting speed of the unmanned platform;

[0028] According to the reachable grid area and the obstacle grid area, invalid grids are removed from the gridded airspace model to obtain a target grid model;

[0029] Calculating a fusion covariance matrix according to the target grid model;

[0030] Determining the unmanned driving platform combination according to the fusion covariance matrix;

[0031] The next detection cycle and the grid position reached by the unmanned driving platform are determined according to the fusion covariance matrix.

[0032] In some embodiments, calculating the fusion covariance matrix according to the target grid model includes:

[0033] Calculating the target state of the detection system according to the state transfer matrix, the first target state and process noise;

[0034] Calculating an observation vector based on a state observation matrix, a target state of the detection system, and observation noise;

[0035] constructing a noise covariance matrix based on the observation noise;

[0036] Constructing a posterior estimation equation based on the predicted state, the posterior covariance, the state observation matrix, the noise covariance matrix and the observation vector;

[0037] Construct the posterior covariance equation based on the inverse matrix of the prediction error covariance matrix, the state observation matrix and the noise covariance matrix;

[0038] Fusing the posterior estimation equation and the posterior covariance equation to obtain a fused equation;

[0039] Performing a prediction on the fusion equation to obtain a predicted equation;

[0040] Inverting the predicted equation to obtain an inverse equation;

[0041] Performing time axis conversion on the inverse equation to obtain a time axis representation equation;

[0042] According to the grid information in the target grid model, the time axis representation equation is subjected to grid conversion to obtain the fusion covariance matrix.

[0043] In some embodiments, determining the unmanned driving platform combination according to the fusion covariance matrix includes:

[0044] Determining a preset number of target grids according to the fusion covariance matrix;

[0045] Smoothing the path points to obtain a smoothed path;

[0046] Constructing a smooth path constraint based on the smoothed path and a path that meets the smoothness requirements of the unmanned platform flight path;

[0047] Constructing a cost function according to the smooth path constraint, the preset number of target grids and the decision variables;

[0048] Calculate the energy consumption of the straight path based on the straight-line flight distance;

[0049] Calculate the corner energy consumption according to the corner path;

[0050] Calculating the energy consumption of the multi-angle arc approach path according to the corner energy consumption and the corner radius;

[0051] Calculating the circular arc path energy consumption according to the polygonal arc approximation path energy consumption and the arc of the circular arc path segment;

[0052] Calculating total energy consumption based on the corner energy consumption, the straight path energy consumption, and the arc path energy consumption;

[0053] The unmanned driving platform combination is determined based on the cost function and the total energy consumption, and the selection goal of the unmanned driving platform combination is to minimize the required total energy consumption.

[0054] In some embodiments, determining the next detection cycle and the grid position reached by the unmanned driving platform according to the fusion covariance matrix includes:

[0055] Calculating a first normalized trace according to the fusion covariance matrix;

[0056] Calculate the corrected normalized trace based on the energy consumption per unit distance and the first normalized trace;

[0057] determining the next detection cycle according to the corrected normalized trace;

[0058] The grid position reached by the unmanned driving platform is determined within the next detection cycle.

[0059] In some embodiments, the method further comprises:

[0060] Sending the current cycle detection strategy to the unmanned driving platform, wherein the current cycle detection strategy includes path planning, next cycle detection position and detection time;

[0061] The unmanned driving platform is used to detect the target according to the current cycle detection strategy to obtain the target distance and target angle;

[0062] Calculating detection target information according to the target distance and the target angle, the detection target information including position information and speed information;

[0063] Determine the next cycle detection strategy based on the detection target information.

[0064] On the other hand, an embodiment of the present invention provides a mobile target detection device based on an unmanned driving platform, comprising:

[0065] The first module is used to calculate the target detection demand change;

[0066] The second module is used to calculate the maximum detection period according to the target detection demand change;

[0067] A third module is configured to calculate a target state set based on the current position of the mobile target and the maximum detection period;

[0068] A fourth module is used to determine the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform according to the target state set;

[0069] The fifth module is used to perform mobile target detection in the next cycle based on the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform.

[0070] In another aspect, an embodiment of the present invention provides a computer device, comprising:

[0071] at least one processor;

[0072] at least one memory for storing at least one program;

[0073] When the at least one program is executed by the at least one processor, the at least one processor implements the method.

[0074] On the other hand, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described above is implemented.

[0075] The beneficial effects of the present invention are as follows:

[0076] The embodiment of the present invention first calculates the target detection demand change degree, calculates the maximum detection cycle based on the target detection demand change degree, then calculates the target state set based on the current position of the mobile target and the maximum detection cycle, and then determines the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform based on the target state set. Finally, according to the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform, the next cycle of mobile target detection is performed, so that mobile target detection can be achieved through different unmanned driving platform strategies, thereby improving positioning accuracy and adaptability and reducing energy consumption.

[0077] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the structures particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0079] Figure 1 This is a flow chart of a method for detecting a moving target based on an unmanned driving platform according to an embodiment of the present invention;

[0080] Figure 2 A schematic diagram of an application scenario of mobile target detection based on an unmanned driving platform according to an embodiment of the present invention;

[0081] Figure 3 A schematic diagram of a gridded airspace model according to an embodiment of the present invention;

[0082] Figure 4 This is a signaling flow chart for mobile target detection according to an embodiment of the present invention;

[0083] Figure 5 This is a structural diagram of a mobile target detection device based on an unmanned driving platform according to an embodiment of the present invention;

[0084] Figure 6 The figure is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0085] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are merely examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the appended claims.

[0086] It will be understood that the terms "first", "second", etc. used in this application may be used herein to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0087] The terms "at least one", "plurality", "each", "any", etc. used in this application include "at least one", "two" or more, "plurality" or "each", "any" or "any one", "each" or "any one" as used herein.

[0088] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0089] Before explaining the embodiments of the present application in detail, some of the nouns and terms involved in the embodiments of the present application are first explained. The nouns and terms involved in the embodiments of the present application are subject to the following explanations.

[0090] An autonomous driving platform is a technology system that enables vehicles or drones to autonomously complete driving tasks without human intervention. This platform relies on multiple sensors and advanced algorithms to perceive the surrounding environment, make decisions, and execute actions.

[0091] In related technologies, target detection and positioning typically involve measuring distances or distances and angles by unmanned platform nodes, which can be stationary or mobile. The system knows the node's location coordinates. After detecting the distance or distance and angle to the target, it sends this information to a fusion center, which uses methods such as Kalman filtering or particle filtering to obtain information such as the target's location coordinates and velocity. Unmanned platforms include drones and unmanned boats. Target positioning and tracking by a cluster of unmanned platforms utilizes cooperative sensing, path planning, and information sharing to achieve real-time positioning and tracking. This technology is commonly used in military reconnaissance, security monitoring, disaster relief, and other fields, offering high efficiency and flexibility. These applications emphasize the need for high-precision target detection. Typically, multiple unmanned platform nodes measure the distance to a target and transmit this information to a fusion center, which then calculates the target's location and velocity. During target tracking and positioning, the combination of the detection cycle and the location of the unmanned platform nodes is closely related to positioning accuracy. Compared to sensor network positioning on fixed unmanned platforms, unmanned platforms offer the advantage of flexible positioning. To fully leverage this advantage, how to properly adjust the unmanned platform's position for detection and accurate positioning is a pressing issue. Furthermore, unmanned platforms are powered by their own batteries, which consume power rapidly. Path planning also affects the platform's energy consumption during movement. Another pressing issue is how to complete positioning tasks with low energy consumption and extend positioning endurance. Traditional target detection methods typically use fixed or mobile nodes to measure the target's distance or angle, and then a fusion center processes the data to estimate the target's position and velocity. These methods suffer from insufficient positioning accuracy, high energy consumption, and poor adaptability on mobile unmanned platforms (such as drones or unmanned ships).

[0092] In view of this, this embodiment selects a grid based on the estimated error covariance matrix and the minimum energy consumption of the unmanned driving platform. This grid is also the position grid that the unmanned driving platform may reach in the next cycle. First, based on the change in target detection demand, the maximum detection cycle is determined to be T max ; Secondly, in T max The target position is predicted based on n different periods. Finally, the space is gridded and the covariance matrix of the single estimation error of the positioning of the unmanned driving platform after arriving at each grid is calculated. Then the fusion covariance matrix of the unmanned driving platform combination is calculated. The normalized trace of the unmanned driving platform reaching the grid position combination is calculated based on the fusion covariance. Combined with the comprehensive energy consumption of the unmanned driving platform, the optimal period T is determined. j and unmanned driving platform nodes, select appropriate grid positions for detection, improve positioning accuracy and reduce energy consumption of unmanned driving platforms.

[0093] The mobile target detection method based on an unmanned driving platform provided in the embodiment of the present application relates to the field of target detection technology. The mobile target detection method based on an unmanned driving platform provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application that implements a mobile target detection method based on an unmanned driving platform, etc., but is not limited to the above forms.

[0094] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0095] The following is a detailed explanation of the embodiments of the present application with reference to the accompanying drawings:

[0096] Figure 1 This is an optional flow chart of a mobile target detection method based on an unmanned driving platform provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S105.

[0097] Step S101, calculating the target detection demand change degree;

[0098] Step S102: Calculate the maximum detection period according to the target detection demand change degree;

[0099] Step S103: Calculate the target state set according to the current position of the mobile target and the maximum detection period;

[0100] Step S104: Determine the unmanned driving platform combination, the next detection cycle, and the grid position to be reached by the unmanned driving platform according to the target state set;

[0101] Step S105: Perform mobile target detection for the next cycle based on the unmanned driving platform combination, the next detection cycle, and the grid position reached by the unmanned driving platform.

[0102] Steps S101 to S105 shown in the embodiment of the present application implement mobile target detection, improve positioning accuracy and adaptability, and reduce energy consumption.

[0103] In some embodiments, in step S101, calculating the target detection requirement change degree may include but is not limited to the following steps:

[0104] Calculate the target direction change degree according to the moving speed and target moving rate of the moving target;

[0105] Calculate the target rate change according to the target moving rate;

[0106] Calculate the target change degree according to the target direction change degree and the target speed change degree;

[0107] The network change degree is calculated based on the link weights between the unmanned driving platforms, the network topology, and the total number of edges in the two network topologies at adjacent moments.

[0108] The target detection requirement change degree is calculated based on the weight coefficient, target change degree and network change degree.

[0109] In some embodiments, the current position of the mobile target and the moving speed of the mobile target can be first obtained through the unmanned driving platform node. The current position of the mobile target and the moving speed of the mobile target can also be obtained by other means, but not limited thereto. For example, in each detection cycle, the target can be detected by the unmanned driving platform node, and the position and moving speed of the target detected by each of them can be obtained respectively. The data fusion center fuses the detection data of each unmanned driving platform node, and finally obtains the current position of the mobile target and the moving speed of the mobile target. In addition, the application scenarios of the mobile target detection based on the unmanned driving platform of this embodiment are as follows: Figure 2 As shown, multiple unmanned driving platform nodes detect targets simultaneously and calculate target information in conjunction with the fusion center.

[0110] Then, the target direction change degree is calculated according to the moving speed and target moving rate of the moving target, wherein the calculation formula of the target direction change degree is: Where θ o(t, tT) is the target direction change, which means the ratio of the velocity direction angle between time t and time tT to 2π. Arccos is the inverse cosine function. is the moving speed of the moving target detected at the tth moment, is the moving speed of the moving target detected at the tTth moment, is the target moving speed detected at the tth moment, is the target movement speed detected at time tT. It can be understood that a series of calculations are performed using the speeds at these two moments to obtain the next detection cycle, and the target is detected to obtain the position and speed information of the next cycle. The initial speed is also detected by the system. Based on the target movement speed, the target speed change degree is calculated, where the calculation formula of the target speed change degree is: Where, v o (t, tT) is the target velocity change, which represents the ratio of the difference between the target velocity at time t and time tT to the velocity at time t. The target velocity change is calculated based on the target direction change and the target velocity change. The target velocity change is calculated as follows: OCD(t, tT) = θ o (t,tT)+v o (t, tT), where OCD(t, tT) is the target variation.

[0111] Then, the network change degree is calculated based on the link weight between the unmanned driving platforms, the network topology, and the total number of edges of the two network topologies at adjacent times. The calculation formula of the network change degree is: Where, Δ tp is the network change degree, |E1∪E2| is the network topology G at the tth moment t (V t ,E t ) and the network topology G at time tT t-T (V t-T ,E t-T ) The total number of edges, that is, the total number of edges of two network topologies at adjacent times, w t (u,v) is G t (V t ,E t ) in the network unmanned driving platform node u and the network unmanned driving platform node v, w t-T (u,v) is G t-T (V t-T ,E t-T) in the network unmanned driving platform node u and the network unmanned driving platform node v, that is, the link weight between the unmanned driving platforms, can be expressed as the distance between the network unmanned driving platform nodes. The greater the network change, the greater the network topology change, and the more frequent the network perception needs. According to the weight coefficient, the target change and the network change, the target detection demand change is calculated. The calculation formula of the target detection demand change is: S o (t,tT)=w1×OCD(t,tT)+w2×Δ tp , where S o (t, tT) is the change degree of target detection demand, w1 and w2 are weight coefficients. The greater the target change or the greater the network change, the more frequent the target detection needs.

[0112] In some embodiments, in step S102, calculating the maximum detection period according to the target detection demand change degree may include but is not limited to the following steps:

[0113] Calculate the current pheromone and the previous pheromone according to the change of target detection demand;

[0114] Calculate the volatility coefficient based on the current pheromone and the previous pheromone;

[0115] According to the volatility coefficient, the current pheromone is updated until the current pheromone is less than the preset volatility threshold, and the maximum detection period is obtained.

[0116] In some embodiments, the detection period needs to be adaptively changed as the target moves or the network changes. As the target detection demand change increases, the detection period needs to be shortened so that the target information can be detected in a timely manner; conversely, the detection period needs to be lengthened to save energy. The process of obtaining the target detection demand change is equivalent to the process of obtaining pheromones in the ant colony algorithm, and the value of the target detection demand change is equivalent to the value of the pheromone. Target detection update period T max Set the time for pheromone to evaporate. The faster the detection demand changes, the faster the pheromone evaporates. max The smaller the value, the shorter the detection interval; the slower the detection demand changes, the slower the pheromone evaporates, T max The value is relatively large, and the detection interval is lengthened. When the pheromone evaporates to a small enough time, it is considered that the pheromone has evaporated and the next target detection can be carried out. The pheromone at the current moment and the pheromone at the previous moment can be calculated according to the change degree of the target detection demand. The calculation formula of the pheromone at the current moment is: Q0(t)=S o (t,tT), where Q0(t) is the pheromone at the current moment, S o (t, tT) is the target detection demand change at the current moment. The calculation formula of pheromone at the previous moment is: Q0(tT)=So (tT,t-2T), where Q0(tT) is the pheromone at the previous moment, S o (tT, t-2T) is the target detection demand change at the previous moment. Based on the pheromone at the current moment and the pheromone at the previous moment, the volatility coefficient is calculated. The calculation formula of the volatility coefficient is: Where ρ is the volatility coefficient, ρ1, ρ2, ρ3, and α are all adjustable constants that can be set according to the scenario and environment. α can range from (0 to 1), and k represents the kth update. It can be understood that if the detection demand changes slightly and is relatively stable, the volatility time can be slightly longer and the value of ρ1 can be slightly smaller; if the detection demand changes moderately, the value of ρ2 can be set slightly larger than ρ1; if the detection demand changes significantly, the value of ρ3 can be set slightly larger than ρ2.

[0117] Then, according to the volatility coefficient, the current pheromone is updated until the current pheromone is less than the preset volatility threshold, and the maximum detection period is obtained. The calculation formula for updating the current pheromone is: Q k (t) = (1-ρ) k Q0(t), where Q k (t) is the pheromone at the current time of the kth update. k When ≈0.0001*Q0, it can be considered that the pheromone has evaporated. At this time, the number of iterations is the maximum detection cycle. It is understandable that each time the pheromone is updated, its value will become smaller, and this iteration will repeat until it is completely evaporated. The speed of pheromone evaporation depends on the volatility coefficient ρ, which is determined by comparing the value of the pheromone Q0(t) obtained this time with the value of the pheromone Q0(tT) obtained last time.

[0118] In some embodiments, in step S103, the target state set is calculated based on the current position of the mobile target and the maximum detection period, which may include but is not limited to the following steps:

[0119] Divide the maximum detection period according to the total number of preset periods to obtain multiple prediction periods;

[0120] Determine the target state at the current moment according to the current position and the moving speed of the moving target;

[0121] The target state set is calculated based on the state transition matrix, the current target state and multiple prediction cycles.

[0122] In some embodiments, it is usually necessary to determine at the kth moment the combination of unmanned driving platform nodes that perform target detection at the k+1th moment and which grids these unmanned driving platforms reach. Target prediction is performed at the kth moment to predict the position of the target at the k+1th moment. Since the effective detection distance of the unmanned driving platform node is within a certain range, the range in which the target can be detected can be determined after predicting the position at the k+1th moment. The maximum detection cycle can be divided according to the total number of preset cycles to obtain multiple prediction cycles. For example, the total number of preset cycles n can be determined according to the accuracy requirements. If the accuracy requirements are high, the value of n can be set to a larger value. 0-T max The second is divided into n parts and predicted separately Where j = 1, 2…n, n cycles. Based on n cycles, n predicted positions of the target can be predicted. Then, based on the current position and moving speed of the moving target, the current target state is determined. Then, based on the state transfer matrix, the current target state and multiple prediction cycles, the target state at the next moment is calculated, and then the target state set is obtained by combining them. For example, in the maximum detection cycle T max Within n different periods, the target position is predicted and the target prediction is performed. Assuming that the target is a slowly moving maneuvering point. For some complex motion models, a multi-model interaction method can be used, and an approximate uniform speed model can be used. The calculation formula for the target state at the next moment is: Where, is the target state at the next moment, x t is the target state at the current moment, is the state transfer matrix for the j-th prediction, x t =[x,v x ,y,v y ,z,v z ] T , represents the target state at the tth moment (the current target state), (x, y, z) is the current position of the moving target, (v x ,v y ,v z ) is the moving speed of the target in the x, y, and z directions. The t+Tth time of the jth prediction j The target state at a moment can be predicted based on The target position information and the detection range of the unmanned driving platform node are used to determine the position of the j-th predicted target and which grid units the unmanned driving platform can reach to detect the target. The target state at each moment is then combined to obtain the target state set.

[0123] In some embodiments, in step S104, determining the unmanned driving platform combination, the next detection cycle, and the grid position reached by the unmanned driving platform based on the target state set may include but is not limited to the following steps:

[0124] Step S201: rasterize the prediction space to obtain a rasterized spatial domain model;

[0125] Step S202: Determine the reachable grid area based on the target state set, the detection radius, maximum flight speed, maximum lifting speed, maximum flight distance, horizontal navigation speed, and vertical lifting speed of the unmanned platform;

[0126] Step S203: Remove invalid grids from the rasterized spatial model according to the reachable grid area and the obstacle grid area to obtain a target grid model;

[0127] Step S204: Calculate the fusion covariance matrix according to the target grid model;

[0128] Step S205: Determine the unmanned driving platform combination according to the fusion covariance matrix;

[0129] Step S206: Determine the next detection cycle and the grid position to be reached by the unmanned driving platform based on the fusion covariance matrix.

[0130] In some embodiments, the prediction space can be first rasterized to obtain a rasterized spatial domain model such as Figure 3 As shown, the gridded airspace model is exemplarily represented using a triplet array structure G = (Level, Address, Property), where Level represents the grid cell height code, Address represents the address code, and Property represents the attribute set. The height code represents the grid cell's spatial height information, the address code represents its latitude and longitude information, and the attribute code represents the various data types and attribute characteristics of the grid cell. This airspace rasterization model has both spatial and attribute implications, offering significant advantages in the positioning and deployment of unmanned platform nodes. Based on the GeoSOT grid, the airspace rasterization model facilitates the conversion between latitude and longitude coordinates and grid codes, providing a comprehensive representation of the airspace. Its unique encoding scheme ensures inherent uniqueness, eliminating the need for any additional identification code, making data retrieval easier. This achieves digital airspace. Grid cells can serve as both spatial indexes for airspace information and as components of the airspace for various needs. This method allows for the description of any location in the airspace, enabling control from a single map at the control center.

[0131] Then, based on the target state set, the detection radius of the unmanned driving platform, the maximum flight speed, the maximum lifting speed, the maximum flight distance, the horizontal navigation speed and the vertical lifting speed, the reachable grid area is determined, and then based on the reachable grid area and the obstacle grid area, the invalid grids of the rasterized airspace model are removed to obtain the target grid model. For example, due to the flight capability of the unmanned driving platform and the obstruction of the surrounding environment, there are some grids that the unmanned driving platform cannot reach or does not need to reach, and these grids can be removed. The grids that the unmanned driving platform can reach can be determined based on the capabilities of the unmanned driving platform. The detection radius, maximum flight speed, maximum lifting speed and maximum flight distance of the unmanned driving platform are all limited. These are the capabilities of the unmanned driving platform. Set the prediction t+T j The target's arrival position at time O, the detection radius of the unmanned driving platform is R, and the sphere with O as the center and R radius is the space where the target can be detected. The grid in this space is effective. The unmanned driving platform's horizontal navigation speed and vertical lifting and lowering speed in the air are also limited, so at T j The maximum distance that can be reached is also limited. For the grid in the spherical space with a radius of R, due to the limitation of the unmanned driving platform, the maximum distance that can be reached in T j The unmanned driving platform cannot reach certain grids within a certain time, and these grids need to be removed. At the same time, grids that have obstacles to detection and communication can also be removed. During the detection process of the unmanned driving platform, it has a certain perception of the environment and surrounding obstacles, and predicts that the unmanned driving platform T j After reaching the target location, due to obstacles, the detection capability of the target within a certain range is determined to be very poor. At this time, these grids are removed. Finally, based on the target grid model, the fusion covariance matrix is calculated. Based on the fusion covariance matrix, the unmanned driving platform combination is determined. Based on the fusion covariance matrix, the next detection cycle and the grid position reached by the unmanned driving platform are determined.

[0132] In some embodiments, in step S204, calculating the fusion covariance matrix according to the target grid model may include but is not limited to the following steps:

[0133] Calculating the target state of the detection system according to the state transfer matrix, the first target state and process noise;

[0134] Calculate the observation vector based on the state observation matrix, the target state of the detection system and the observation noise;

[0135] According to the observation noise, the noise covariance matrix is constructed;

[0136] Construct the posterior estimation equation based on the predicted state, posterior covariance, state observation matrix, noise covariance matrix and observation vector;

[0137] Construct the posterior covariance equation based on the inverse matrix of the prediction error covariance matrix, the state observation matrix and the noise covariance matrix;

[0138] The posterior estimation equation and the posterior covariance equation are fused to obtain a fusion equation;

[0139] Make a prediction on the fusion equation to obtain the predicted equation;

[0140] Inverse the predicted equation to obtain the inverse equation;

[0141] Perform time axis transformation on the inverse equation to obtain the time axis representation equation;

[0142] According to the grid information in the target grid model, the time axis representation equation is transformed into a grid to obtain the fusion covariance matrix.

[0143] In some embodiments, the target state of the detection system can be calculated based on the state transfer matrix, the first target state and the process noise, wherein the calculation formula of the target state of the detection system is: k+1 =Φ k x k +v k ,k=0,1,…,n, where x k+1 is the target state of the detection system, that is, the target state at time k+1, Φ k is the state transfer matrix, x k is the first target state, that is, the target state at time k, v k is the process noise, and n is the total number of preset cycles. According to the state observation matrix, the target state of the detection system and the observation noise, the observation vector is calculated. The calculation formula of the observation vector is: Where, (*) ′ represents transpose, is the observation vector, is the state observation matrix, is the observation noise (matrix), N is the total number of grids, is the observation value corresponding to a single target. According to the observation noise, the noise covariance matrix is constructed, where the expression of the noise covariance matrix is: Where R k is the noise covariance matrix, is the i-th element in the noise covariance matrix, that is, the observation noise covariance corresponding to a single target.

[0144] Then, the posterior estimation equation is constructed based on the predicted state, posterior covariance, state observation matrix, noise covariance matrix and observation vector, where the expression of the posterior estimation equation is: Where x k / k is the posterior estimate, x k / k-1 is the predicted state, P k / k is the posterior covariance. According to the inverse matrix of the prediction error covariance matrix, the state observation matrix and the noise covariance matrix, the posterior covariance equation is constructed, where the expression of the posterior covariance equation is: Where, is the inverse of the posterior covariance, is the inverse matrix of the prediction error covariance matrix.

[0145] P k / k =E[(x k / k -x k )(x k / k -x k ) ′ ∣∣Y k ], P k / k-1 =E[(x k / k-1 -x k )(x k / k-1 -x k ) ′ ∣∣Y k-1 ].

[0146] Then the posterior estimation equation and the posterior covariance equation are fused to obtain the fusion equation, where the expression of the fusion equation is: Make a prediction on the fusion equation to obtain the predicted equation, where the expression of the predicted equation is: Furthermore, after making a prediction, you can get: Where x k-1 / k-1 is the posterior estimate at time k-1, x k / k-1 is the state prediction value at time k based on the estimation at time k-1, that is, the prior estimate. k is the state transition matrix, which describes how the system evolves from time k-1 to time k. The predicted equation is inverted to obtain the inverse equation, where the expression of the inverse equation is: Where, P k+1 / k+1 is the posterior fusion covariance at time k+1.

[0147] Finally, the inverse equation is transformed into a time axis to obtain the time axis representation equation, where the expression of the time axis representation equation is: In order to further select the cycle and grid with better detection performance and energy consumption of the unmanned driving platform from the candidate cycles and grids, the tracking effect and energy consumption of different cycle and grid position combinations are predicted. Note that in the above derivation, it is not necessary to obtain measurement information to calculate the estimation error covariance P of the extended Kalman filter algorithm, which means that t+T can be obtained at time t. j The estimated error covariance matrix P of each unmanned driving platform node at each moment can be used to select the unmanned driving platform node based on the estimated error covariance matrix. Since predictions need to be made separately within j periods, a grid represents a position. Based on the grid information in the target grid model, the time axis representation equation can be grid-converted to obtain the fused covariance matrix. The expression of the fused covariance matrix is: Where, is the predicted t+T j The fusion covariance matrix of N grids selected at any moment, where i is the i-th grid.

[0148] In some embodiments, in step S205, determining the unmanned driving platform combination according to the fusion covariance matrix may include but is not limited to the following steps:

[0149] Determine a preset number of target grids according to the fusion covariance matrix;

[0150] Smoothing the path points to obtain a smoothed path;

[0151] Constructing smooth path constraints based on the smoothed path and the path that meets the smoothing requirements of the unmanned platform flight path;

[0152] Construct a cost function based on the smooth path constraints, a preset number of target grids, and decision variables;

[0153] Calculate the energy consumption of the straight path based on the straight-line flight distance;

[0154] Calculate the corner energy consumption according to the corner path;

[0155] Calculate the energy consumption of the multi-angle arc approach path based on the corner energy consumption and corner radius;

[0156] Calculate the energy consumption of the circular arc path based on the energy consumption of the polygonal arc approximation path and the arc of the circular arc path segment;

[0157] Calculate the total energy consumption based on corner energy consumption, straight path energy consumption and arc path energy consumption;

[0158] According to the cost function and total energy consumption, the unmanned driving platform combination is determined, and the selection goal of the unmanned driving platform combination is to minimize the required total energy consumption.

[0159] In some embodiments, a preset number of target grids can be determined based on the fusion covariance matrix. After finding N target grids, N unmanned driving platforms need to be selected from the unmanned driving platform group to determine their flight paths. The path planning of the unmanned driving platform in three-dimensional space is to grid the three-dimensional space, and the starting point and end point of the unmanned driving platform belong to different grids. Several grids between the starting point and the end point are found so that the connection line of all points meets the established constraints. On this basis, the broken line path is smoothed to generate a path that the unmanned driving platform can actually fly. The path points can be smoothed to obtain a smoothed path p(x), and a smoothed path constraint is constructed based on the smoothed path and the path that meets the smoothing requirements of the unmanned driving platform flight path. The expression of the smoothed path constraint is: p(x)∈P, where p(x) is the path point smoothing, and P is the path that meets the smoothing requirements of the unmanned driving platform flight path. A cost function is then constructed based on the smooth path constraints, a preset number of target grids, and decision variables. The cost function is expressed as: minC = f(x), where C is the cost function derived from the decision variables, representing the degree of fit (a smaller C value indicates a better fit), and x represents the decision variable containing all path point information. It can be understood that the ultimate goal of path planning is to minimize the path point cost function while achieving a smooth path that meets the flight requirements of the unmanned platform.

[0160] After the unmanned platform path is planned, a path combination with the lowest energy consumption can be selected based on the planned path combination and energy estimation model. In the energy consumption estimation, the total energy consumption can be obtained by summing the energy consumption of the corner path, straight path, and arc path in the broken line path. The energy consumption of the straight line path can be calculated based on the straight line flight distance. Generally, the straight line flight energy consumption estimation model is basically proportional to the straight line flight distance. The calculation formula for the straight line path energy consumption is: g v =kd, where g v is the energy consumption of the straight path, k is the energy consumption coefficient of the straight path, which is determined by the drone parameters, and d is the straight path in the path. According to the corner path, the corner energy consumption is calculated. For the broken line path, the larger the steering angle, the greater the energy consumption. The actual data of the steering angle and energy consumption are very close to the calculated quadratic curve. Therefore, the calculation formula of the corner energy consumption that can be expressed by this quadratic polynomial is: v (θ)=aθ 2 +bθ, where, f v(θ) is the corner energy consumption at a certain speed, a and b are both corner energy consumption coefficients, which are determined by the drone parameters, and θ is the corner path in the broken line path in path l. According to the corner energy consumption and corner radius, the energy consumption of the multi-angle arc approximation path is calculated. For the circular arc path, it can be compared with the broken line path. When turning the same angle, it consumes more energy to complete it with a large corner than to complete it with multiple small corners. When the small corners are further subdivided, the extreme case is the arc. The calculation formula for the multi-angle arc approximation (the arc of unit length of the unmanned platform flight) path energy consumption is: Where, f e (r) is the energy consumption of the polygonal arc approach path at a certain speed, X and Y are the energy consumption coefficients of the polygonal arc approach path, which are determined by the drone parameters, and r is the arc radius of the arc path segment in the path. According to the energy consumption of the polygonal arc approach path and the arc of the arc path segment, the energy consumption of the arc path is calculated, where the calculation formula of the arc path energy consumption is: e v =f e (r)·m, where e v is the energy consumption of the arc path at a certain speed, and m is the arc length of the arc path segment in the path.

[0161] Finally, the total energy consumption is calculated based on the corner energy consumption, straight path energy consumption and arc path energy consumption. The calculation formula for the total energy consumption is: E = ∑f v (θ)+∑g v (d)+∑e v (r, m), where E is the total energy consumption of the unmanned driving platform flying over the path l. According to the cost function and the total energy consumption, the unmanned driving platform combination is determined, and the selection target of the unmanned driving platform combination is to minimize the required total energy consumption. For example, after finding N target grids, N unmanned driving platforms are selected from the unmanned driving platform group to reach N grids. First, the unmanned driving platform path planning is performed, and then the energy consumption is predicted based on the planned path, and the lowest combination of N unmanned driving platforms is selected. The total energy consumption generated by different unmanned driving platforms allocating different grids is inconsistent. In the process of path planning, in addition to considering obstacle avoidance to reach the destination, it is also necessary to consider the lowest overall energy consumption. Find the path combination with the lowest energy consumption, and its total energy consumption can be expressed as E m The energy consumption per unit distance can be expressed as

[0162] In some embodiments, in step S206, determining the next detection cycle and the grid position reached by the unmanned driving platform based on the fusion covariance matrix may include but is not limited to the following steps:

[0163] Calculate the first normalized trace based on the fusion covariance matrix;

[0164] Calculate the corrected normalized trace based on the energy consumption per unit distance and the first normalized trace;

[0165] Determine the next detection cycle based on the corrected normalized trace;

[0166] The grid position reached by the unmanned driving platform is determined in the next detection cycle.

[0167] In some embodiments, the first normalized trace can be calculated based on the fusion covariance matrix. For example, after the fusion covariance matrix is calculated, the common practice is to calculate the trace of the fusion covariance. However, in this embodiment, the fusion covariance has both position error and speed error, and the dimensions are not unified. Therefore, the normalized trace can be introduced to define the quality of the unmanned driving platform node, and the normalized traces of different unmanned driving platform node combinations with different periods are calculated. The calculation formula of the first normalized trace is: Where S is the first normalized trace, p 11 is the element of the first row and first column of the fusion covariance matrix P, p ii is the element in the ith row and ith column of the fusion covariance matrix P, T j is the jth period. However, this only considers the impact of the estimation error covariance, without considering the impact of the unmanned driving platform energy consumption. It can be concluded that the energy consumption per unit distance is Therefore, the corrected normalized trace is introduced. The corrected normalized trace can be calculated based on the energy consumption per unit distance and the first normalized trace. The calculation formula of the corrected normalized trace is: Where W is the normalized trace, E ma The minimum energy consumption of the unmanned driving platform to reach N grid combinations predicted for the ath time, l is the path. Finally, based on the corrected normalized trace, the next detection period is determined, and the grid position reached by the unmanned driving platform is determined within the next detection period. For example, the corrected normalized trace W of different unmanned driving platform position combinations with different periods is calculated, and the unmanned driving platform position combination with the smallest W is selected. At the same time, the period T of the smallest W is also determined. j As the next detection cycle, and determine T j The grid position reached by the unmanned driving platform at any time so that target positioning can be performed later.

[0168] In some embodiments, in step S105, the next cycle of mobile target detection can be performed based on the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform. For example, after calculating the next detection cycle, which drone nodes perform the next cycle of detection (i.e., the unmanned driving platform combination), and the positions where these unmanned driving platforms should appear in the next cycle (i.e., the grid positions reached by the unmanned driving platform), these calculations are performed in the fusion center. After the results are calculated, the fusion center sends instructions to the unmanned driving platform nodes participating in the detection in the next cycle, and the unmanned driving platform nodes are instructed to perform detection in the next cycle. The results of the detection will be sent back to the fusion center, and the coordinate position and speed of the mobile target will be located by calculation in the fusion center, and the next cycle of mobile target detection can be repeated subsequently.

[0169] In some embodiments, the method further comprises:

[0170] Send the current cycle detection strategy to the unmanned driving platform. The current cycle detection strategy includes path planning, next cycle detection location and detection time;

[0171] The unmanned driving platform is used to detect targets according to the current cycle detection strategy and obtain target distance and target angle;

[0172] Calculate the detection target information according to the target distance and target angle, and the detection target information includes position information and speed information;

[0173] Determine the detection strategy for the next cycle based on the detection target information.

[0174] In some embodiments, the signaling process of mobile target detection is as follows: Figure 4 As shown, the current cycle detection strategy can be first sent to the unmanned driving platform through the data fusion center. The current cycle detection strategy includes path planning, the detection location for the next cycle, and the detection time. The unmanned driving platform is used to detect the target according to the current cycle detection strategy, obtain the target distance and target angle, and transmit it back to the data fusion center. It is understandable that the detected information can include only the target distance or both the target distance and target angle. Then, based on the target distance and target angle, the data fusion center calculates the detection target information. The detection target information includes position information and speed information. Finally, based on the detection target information, the detection strategy for the next cycle is determined. By predicting the location information of targets in different cycles, the grid within the effective detection range of each cycle is determined. The next cycle detection strategy is determined based on energy consumption and the estimation error covariance matrix. It includes the next network detection cycle, the unmanned driving platform combination for detecting targets, and the location of the next cycle. The next cycle detection strategy is then sent to the unmanned driving platform for the next cycle detection.

[0175] The beneficial effects of implementing the embodiments of the present invention include: the embodiments of the present invention first calculate the target detection demand change degree, calculate the maximum detection cycle based on the target detection demand change degree, then calculate the target state set based on the current position of the mobile target and the maximum detection cycle, and then determine the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform based on the target state set, and finally perform the next cycle of mobile target detection based on the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform, so that mobile target detection can be achieved through different unmanned driving platform strategies, thereby improving positioning accuracy and adaptability and reducing energy consumption. At the same time, this embodiment is based on spatial gridding, and the unmanned driving platform low-energy path planning jointly integrates the estimated error covariance to determine the appropriate cycle and detection position.

[0176] like Figure 5 As shown, an embodiment of the present invention further provides a mobile target detection device based on an unmanned driving platform, comprising:

[0177] The first module 801 is used to calculate the target detection demand change degree;

[0178] The second module 802 is used to calculate the maximum detection period according to the target detection demand change;

[0179] The third module 803 is used to calculate the target state set according to the current position of the mobile target and the maximum detection period;

[0180] The fourth module 804 is used to determine the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform according to the target state set;

[0181] The fifth module 805 is used to perform mobile target detection in the next cycle based on the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform.

[0182] The contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0183] like Figure 6 As shown, an embodiment of the present invention further provides a computer device, including:

[0184] at least one processor 901;

[0185] At least one memory 902, configured to store at least one program;

[0186] When at least one program is executed by at least one processor, the at least one processor implements Figure 1The method shown.

[0187] The contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0188] The embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program, which is executed by a processor to implement Figure 1 The method shown.

[0189] The contents of the above method embodiments are all applicable to the present storage medium embodiment. The functions specifically implemented by the present storage medium embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0190] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A mobile target detection method based on an unmanned driving platform, characterized in that: The following steps are involved: Calculate the target detection demand change; Calculating a maximum detection period according to the target detection demand change degree; Calculating a target state set according to the current position of the mobile target and the maximum detection period; Determining, based on the target state set, an unmanned platform combination, a next detection cycle, and a grid position to be reached by the unmanned platform; The next cycle of mobile target detection is performed according to the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform.

2. The method according to claim 1, characterized in that The calculating target detection demand change degree includes: Calculate the target direction change degree according to the moving speed and target moving rate of the moving target; Calculating a target velocity change degree according to the target movement velocity; Calculating a target change degree based on the target direction change degree and the target speed change degree; The network change degree is calculated based on the link weights between the unmanned driving platforms, the network topology, and the total number of edges in the two network topologies at adjacent moments. The target detection requirement change degree is calculated according to the weight coefficient, the target change degree and the network change degree.

3. The method according to claim 1, characterized in that The calculating the maximum detection period according to the target detection requirement change degree includes: Calculating the pheromone at the current moment and the pheromone at the previous moment according to the target detection demand change degree; Calculate the volatility coefficient based on the current pheromone and the previous pheromone; The current pheromone is updated according to the volatility coefficient until the current pheromone is less than a preset volatility threshold, thereby obtaining the maximum detection period.

4. The method according to claim 1, wherein The calculating of the target state set according to the current position of the mobile target and the maximum detection period includes: Dividing the maximum detection period according to a preset total number of periods to obtain a plurality of prediction periods; Determining a current target state according to the current position of the moving target and the moving speed of the moving target; The target state set is calculated according to the state transition matrix, the current target state and the multiple prediction cycles.

5. The method according to claim 1, wherein Determining the unmanned driving platform combination, the next detection cycle, and the grid position reached by the unmanned driving platform according to the target state set includes: Rasterize the prediction space to obtain a rasterized spatial domain model; Determining a reachable grid area based on the target state set, the detection radius, maximum flight speed, maximum lifting speed, maximum flight distance, horizontal navigation speed, and vertical lifting speed of the unmanned platform; According to the reachable grid area and the obstacle grid area, invalid grids are removed from the gridded airspace model to obtain a target grid model; Calculating a fusion covariance matrix according to the target grid model; Determining the unmanned driving platform combination according to the fusion covariance matrix; The next detection cycle and the grid position reached by the unmanned driving platform are determined according to the fusion covariance matrix.

6. The method according to claim 5, characterized in that The step of calculating the fusion covariance matrix according to the target grid model includes: Calculating the target state of the detection system according to the state transfer matrix, the first target state and process noise; Calculating an observation vector based on a state observation matrix, a target state of the detection system, and observation noise; constructing a noise covariance matrix based on the observation noise; Constructing a posterior estimation equation based on the predicted state, the posterior covariance, the state observation matrix, the noise covariance matrix and the observation vector; Construct the posterior covariance equation based on the inverse matrix of the prediction error covariance matrix, the state observation matrix and the noise covariance matrix; Fusing the posterior estimation equation and the posterior covariance equation to obtain a fused equation; Performing a prediction on the fusion equation to obtain a predicted equation; Inverting the predicted equation to obtain an inverse equation; Performing time axis conversion on the inverse equation to obtain a time axis representation equation; According to the grid information in the target grid model, the time axis representation equation is subjected to grid conversion to obtain the fusion covariance matrix.

7. The method according to claim 5, characterized in that The step of determining the unmanned driving platform combination according to the fusion covariance matrix includes: Determining a preset number of target grids according to the fusion covariance matrix; Smoothing the path points to obtain a smoothed path; Constructing a smooth path constraint based on the smoothed path and a path that meets the smoothness requirements of the unmanned platform flight path; Constructing a cost function according to the smooth path constraint, the preset number of target grids and the decision variables; Calculate the energy consumption of the straight path based on the straight-line flight distance; Calculate the corner energy consumption according to the corner path; Calculating the energy consumption of the multi-angle arc approach path according to the corner energy consumption and the corner radius; Calculating the circular arc path energy consumption according to the polygonal arc approximation path energy consumption and the arc of the circular arc path segment; Calculating total energy consumption based on the corner energy consumption, the straight path energy consumption, and the arc path energy consumption; The unmanned driving platform combination is determined based on the cost function and the total energy consumption, and the selection goal of the unmanned driving platform combination is to minimize the required total energy consumption.

8. The method according to claim 5, characterized in that Determining the next detection cycle and the grid position reached by the unmanned driving platform according to the fusion covariance matrix includes: Calculating a first normalized trace according to the fusion covariance matrix; Calculate the corrected normalized trace based on the energy consumption per unit distance and the first normalized trace; determining the next detection cycle according to the corrected normalized trace; The grid position reached by the unmanned driving platform is determined within the next detection cycle.

9. The method according to claim 1, characterized in that The method further comprises: Sending the current cycle detection strategy to the unmanned driving platform, wherein the current cycle detection strategy includes path planning, next cycle detection position and detection time; The unmanned driving platform is used to detect the target according to the current cycle detection strategy to obtain the target distance and target angle; Calculating detection target information according to the target distance and the target angle, the detection target information including position information and speed information; Determine the detection strategy for the next cycle based on the detection target information.

10. A mobile target detection device based on an unmanned driving platform, characterized in that: include: The first module is used to calculate the target detection demand change; The second module is used to calculate the maximum detection period according to the target detection demand change; A third module is configured to calculate a target state set based on the current position of the mobile target and the maximum detection period; A fourth module is used to determine the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform according to the target state set; The fifth module is used to perform mobile target detection in the next cycle based on the unmanned driving platform combination, the next detection cycle and the grid position reached by the unmanned driving platform.

Images