Multi-UAV Cooperative Search Method, Device, Equipment and Medium for Dynamic Targets

Abstract

This application relates to the fields of artificial intelligence and machine learning technologies, and particularly relates to a multi-unmanned vehicle collaborative search method, device, equipment and medium for dynamic targets. Among them, the method includes: obtaining the current position information, actual attitude information of multiple unmanned vehicles, and relative position information between unmanned vehicles; calculating the information entropy distribution of the dynamic target appearing at each position in the search area according to the current position information and actual attitude information; generating the optimal search path for each unmanned vehicle according to the information entropy distribution, obstacle position information in the search area, and relative position information between unmanned vehicles, and controlling multiple unmanned vehicles to execute the target search action according to the optimal search path, realizing the efficient collaborative search of the multi-vehicle system for dynamic targets in a certain search area. Thus, the problems in the related technologies that do not consider obstacle avoidance during the search process, the planned path is relatively simple, and it is not applicable to the application scenarios of ground search are solved.

Description

Technical Field

[0001] The present application relates to the technical fields of artificial intelligence and machine learning, and particularly relates to a multi-unmanned vehicle collaborative search method, device, equipment and medium for dynamic targets. Background Art

[0002] Collaborative search is a typical task of a robot cluster in various application scenarios. Using a robot cluster to perform a search task can, on the one hand, more thoroughly achieve full coverage reconnaissance of the task area and obtain environmental information of the entire area; on the other hand, it can more efficiently search for a specified target. Compared with a single robot completing a search task, the collaborative search of a robot cluster is more complex, and it gives rise to many problems not encountered in single-robot search, including dynamic information fusion, topology control, search task allocation, collaborative path planning, etc.

[0003] Currently, the most widely studied and relatively maturely applied search robot is an unmanned aerial vehicle (UAV). However, the aerial search using UAVs also has certain limitations. In outdoor areas with more obstructions and most indoor search tasks, the ground search using unmanned vehicles has greater advantages than aerial search. However, the search environment faced by ground search is more complex than aerial search, which poses more challenges to the problem of multi-unmanned vehicle collaborative search: ① During the search process, the vehicle must consider the passability of the driving environment, resulting in a large number of obstacle avoidance problems; ② During the ground search process, the detection signals of sensors are often blocked by surrounding obstacles, and in addition, the detection of the sensors themselves also has blind spots; ③ For vehicles with the most commonly used Ackermann steering chassis structure currently, the restrictions during their movement are more than those of UAVs; ④ The sensor coverage range of unmanned vehicles is relatively smaller than that of UAVs, and the search efficiency is relatively low. Therefore, the movement of the target during the search process cannot be ignored.

[0004] A multi-UAV collaborative search method under no-information conditions is proposed in the related technology. This method conducts search with UAVs as the object, does not consider the obstacle avoidance problem during the search process, and the planned path is relatively simple, and it is not applicable to the application scenario of multi-unmanned vehicle ground search. Although this method considers the situation of the target moving during the search process, it fails to give a specific method for predicting the target movement, and only examines the performance of the proposed search method with a dynamic target. Summary of the Invention

[0005] The present application provides a multi-unmanned vehicle collaborative search method, device, equipment and medium for dynamic targets to solve the problems that the related technology does not consider the obstacle avoidance problem during the search process, the planned path is relatively simple, and it is not applicable to the application scenario of ground search, etc.

[0006] An embodiment of the first aspect of the present application provides a multi-unmanned vehicle collaborative search method for dynamic targets, including the following steps: obtaining the current position information, actual attitude information of multiple unmanned vehicles, and relative position information between unmanned vehicles; calculating the information entropy distribution of the dynamic target appearing at each position in the search area according to the current position information and actual attitude information; generating the optimal search path for each unmanned vehicle according to the information entropy distribution, obstacle position information in the search area, and relative position information between unmanned vehicles, and controlling the multiple unmanned vehicles to perform the target search action according to the optimal search path.

[0007] Optionally, in an embodiment of the present application, the calculating the information entropy distribution of the dynamic target appearing at each position in the search area includes: calculating the change in entropy distribution during the adiabatic free expansion process of an ideal gas in the search area; generating the information entropy distribution of the dynamic target appearing at each position in the search area according to the change in entropy distribution.

[0008] Optionally, in an embodiment of the present application, the calculating the information entropy distribution of the dynamic target appearing at each position in the search area according to the current position information and the actual attitude information includes: calculating the change in the number of first gas molecules at each position in the search area after the free expansion of the gas in the previous time step; based on the change in the number of first gas molecules, calculating the change in the number of second gas molecules at each position in the search area after the search by the multiple unmanned vehicles in the previous time step according to the current position information and the actual attitude information; calculating the information entropy distribution of the dynamic target appearing at each position in the search area according to the number of second gas molecules and the ideal gas entropy calculation formula.

[0009] Optionally, in an embodiment of the present application, based on the change in the number of first gas molecules, calculating the change in the number of second gas molecules at each position in the search area after the search by the multiple unmanned vehicles in the previous time step according to the current position information and the actual attitude information includes: according to the current position information and actual attitude information of the multiple unmanned vehicles in the previous time step, detecting whether the allowed passage area in the search area in the previous time step is within the search range of any unmanned vehicle, and if so, changing the number of gas molecules in the allowed passage area to zero.

[0010] Optionally, in an embodiment of the present application, generating an optimal search path for each unmanned vehicle according to the information entropy distribution, the obstacle position information in the search area, and the relative position information between the unmanned vehicles includes: determining multiple candidate search paths in the search area according to the current position information and the actual attitude information of each unmanned vehicle; calculating the search benefits of the unmanned vehicle after multiple time steps on each candidate search path according to the information entropy distribution and the obstacle position information in the search area; calculating the search overlap range of the unmanned vehicle according to the relative position information between the unmanned vehicles, and when the search overlap range is less than or equal to a preset range threshold, using the candidate search path corresponding to the maximum search benefit as the optimal search path.

[0011] Optionally, in an embodiment of the present application, calculating the search overlap range of the unmanned vehicle according to the relative position information between the unmanned vehicles includes: when the search overlap range is greater than the preset range threshold, taking the maximization of the search benefit as the goal, and using a multi-objective optimization algorithm to optimize the candidate search paths of the multiple unmanned vehicles to generate an optimal search path for each unmanned vehicle.

[0012] An embodiment of the second aspect of the present application provides a multi-unmanned vehicle collaborative search device for dynamic targets, including: an acquisition module, configured to acquire the current position information, the actual attitude information of multiple unmanned vehicles, and the relative position information between the unmanned vehicles; a generation module, configured to calculate the information entropy distribution of the dynamic target at each position in the search area according to the current position information and the actual attitude information; a search module, configured to generate an optimal search path for each unmanned vehicle according to the information entropy distribution, the obstacle position information in the search area, and the relative position information between the unmanned vehicles, and control the multiple unmanned vehicles to perform the target search action according to the optimal search path.

[0013] Optionally, in an embodiment of the present application, the generation module is further configured to calculate the change in the entropy distribution during the adiabatic free expansion process of an ideal gas in the search area; and generate the information entropy distribution of the dynamic target at each position in the search area according to the change in the entropy distribution.

[0014] Optionally, in an embodiment of the present application, the generation module includes: a first calculation unit, configured to calculate the change in the number of first gas molecules at each position in the search area after the free expansion of the gas in the previous time step; a second calculation unit, based on the change in the number of first gas molecules, calculates the change in the number of second gas molecules at each position in the search area after the search of the multiple unmanned vehicles in the previous time step according to the current position information and the actual attitude information.

[0015] Optionally, in an embodiment of the present application, the second calculation unit is further configured to detect whether the passable area of the search area in the previous time step is within the search range of any unmanned vehicle according to the current position information and actual attitude information of the multiple unmanned vehicles in the previous time step. If so, the number of gas molecules in the passable area becomes zero.

[0016] Optionally, in an embodiment of the present application, the search module includes: a determination unit configured to determine multiple candidate search paths in the search area according to the current position information and actual attitude information of each unmanned vehicle; a third calculation unit configured to calculate the search benefits of the unmanned vehicle after multiple time steps on each candidate search path according to the information entropy distribution and the obstacle position information in the search area; a fourth calculation unit configured to calculate the search overlap range of the unmanned vehicles according to the relative position information between the unmanned vehicles, and when the search overlap range is less than or equal to a preset range threshold, use the candidate search path corresponding to the maximum search benefit as the optimal search path.

[0017] Optionally, in an embodiment of the present application, when the search overlap range is greater than the preset range threshold, the fourth calculation unit is further configured to optimize the candidate search paths of the multiple unmanned vehicles by using a multi-objective optimization algorithm with the goal of maximizing the search benefit, and generate the optimal search path for each unmanned vehicle.

[0018] An embodiment of the third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to perform the multi-unmanned vehicle collaborative search method for dynamic targets as described in the above embodiments.

[0019] An embodiment of the fourth aspect of the present application provides a computer-readable storage medium, on which a computer program is stored, and the program is executed by a processor to perform the multi-unmanned vehicle collaborative search method for dynamic targets as described in the above embodiments.

[0020] Therefore, the present application has at least the following beneficial effects:

[0021] Based on the real-time position and attitude information of each unmanned vehicle, the information entropy distribution of the current dynamic target appearing at various positions within the entire search area is calculated and updated. Each unmanned vehicle calculates and plans its own driving path for the next period of time according to the information entropy distribution of the target appearing within the local search area around the vehicle itself, the distribution of nearby obstacles, and the positions of nearby other unmanned vehicles. Each unmanned vehicle performs path tracking according to the currently planned path, achieving the efficient collaborative search of the multi-vehicle system for dynamic targets within a certain search area. Thus, the problems in the related technology that do not consider obstacle avoidance during the search process, the planned path is relatively simple, and it is not applicable to the application scenarios of multi-ground search are solved.

[0022] Additional aspects and advantages of the present application will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and / or additional aspects and advantages of the present application will become apparent and easy to understand from the following description of the embodiments in conjunction with the drawings, wherein:

[0024] Figure 1 is a flowchart of a multi-unmanned vehicle collaborative search method for a dynamic target according to an embodiment of the present application;

[0025] Figure 2 is a schematic diagram of calculating the change in the number of molecules in each grid during the gas expansion process according to an embodiment of the present application;

[0026] Figure 3 is a schematic diagram of removing part of the search vehicle within the theoretical maximum detection range due to occlusion according to an embodiment of the present application;

[0027] Figure 4 is a multi-unmanned vehicle collaborative search architecture diagram according to an embodiment of the present application;

[0028] Figure 5 is a schematic diagram of the search area and the distribution of internal obstacles according to an embodiment of the present application;

[0029] Figure 6 is a schematic diagram of the candidate paths of the search vehicle within the next N time steps according to an embodiment of the present application;

[0030] Figure 7 is a schematic diagram of the path planning results of two vehicles without a target when the obstacle distribution before the search is known according to an embodiment of the present application;

[0031] Figure 8 is a schematic diagram of the path planning results of two vehicles without a target when the obstacle distribution before the search is unknown according to an embodiment of the present application;

[0032] Figure 9 Schematic diagram of the cooperative search process of two vehicles when there are random moving targets according to an embodiment of the present application;

[0033] Figure 10 Example diagram of a multi-unmanned vehicle cooperative search device for dynamic targets according to an embodiment of the present application;

[0034] Figure 11 Schematic diagram of the structure of an electronic device provided by an embodiment of the application.

[0035] Explanation of reference numerals: Acquisition module - 100, Generation module - 200, Search module - 300, Memory - 1101, Processor - 1102, Communication interface - 1103. Detailed implementation manners

[0036] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present application, but should not be construed as limiting the present application.

[0037] The multi-unmanned vehicle cooperative search method, device, electronic device and storage medium for dynamic targets according to the embodiments of the present application will be described below with reference to the accompanying drawings. In view of the problems mentioned in the above background art, the present application provides a multi-unmanned vehicle cooperative search method for dynamic targets. In this method, the information entropy distribution of the current dynamic target at each position in the entire search area is calculated and updated according to the real-time position and attitude information of each unmanned vehicle. Each unmanned vehicle calculates and plans its own driving path in the next period of time according to the information entropy distribution of the target appearance in the local search area around the vehicle itself, the distribution of nearby obstacles, and the positions of nearby other unmanned vehicles. Each unmanned vehicle follows the planned path for trace driving, realizing the efficient cooperative search of the multi-vehicle system for dynamic targets in a certain search area. Thus, the problems that the related technology does not consider the obstacle avoidance problem in the search process, the planned path is relatively simple, and it is not suitable for the application scenarios of multi-ground search are solved.

[0038] Figure 1 Flow schematic diagram of a multi-unmanned vehicle cooperative search method for dynamic targets provided by an embodiment of the present application.

[0039] Based on the existing multi-UAV cooperative search method, the present application proposes a method for simulating and predicting the information entropy distribution of a dynamic target at each position in the search area by using the entropy distribution of the free expansion process of an ideal gas in view of the characteristics of ground search, and thus constructs a multi-unmanned vehicle cooperative search method for dynamic targets, providing a theoretical basis and practical reference for improving the efficiency of ground unmanned cooperative search.

[0040] As Figure 1 shown, the multi-unmanned vehicle collaborative search method for dynamic targets includes the following steps:

[0041] In step S101, obtain the current position information, actual attitude information of multiple unmanned vehicles, and the relative position information between the unmanned vehicles.

[0042] To calculate the information entropy distribution of the dynamic target at each position in the search area and reasonably plan a reasonable vehicle search path, the embodiments of the present application need to obtain the current position information, actual attitude information of multiple unmanned vehicles, and the relative position information between the unmanned vehicles.

[0043] In step S102, calculate the information entropy distribution of the dynamic target at each position in the search area according to the current position information and the actual attitude information.

[0044] For the convenience of calculation, the embodiments of the present application assume that the search area is a two-dimensional closed area with a limited area, and the obstacles in the area are simplified into two-dimensional closed figures. The search area needs to be pre-discretized into grids, and the map is divided into a large number of identical square grids, so that the boundaries of the complex-shaped map or the boundaries of obstacles can be accurately depicted. Correspondingly, the grids are also divided into two types: one is the grid representing obstacles, and the other is the grid representing the area where the vehicle is allowed to pass.

[0045] Discretization processing also needs to be performed on the time scale, and it is considered that the smallest time unit is a time step. During the search process, the search information fusion result in the past time step will be updated every time a time step passes, and the search path for the same duration in the future will be planned every time a fixed number of time steps pass until all targets are found.

[0046] Optionally, in an embodiment of the present application, calculating the information entropy distribution of the dynamic target at each position in the search area includes: calculating the change in entropy distribution during the adiabatic free expansion process of an ideal gas in the search area; generating the information entropy distribution of the dynamic target at each position in the search area according to the change in entropy distribution.

[0047] The embodiments of the present application use the change in entropy distribution with time during the adiabatic free expansion process of an ideal gas to calculate and simulate the change in the information entropy size of the dynamic target appearing in each grid with time.

[0048] Optionally, in an embodiment of the present application, calculating the information entropy distribution of the dynamic target at each position in the search area according to the current position information and the actual attitude information includes: calculating the change in the number of first gas molecules at each position in the search area after the free expansion of the gas in the previous time step; based on the change in the number of first gas molecules, calculating the change in the number of second gas molecules at each position in the search area after the search by multiple unmanned vehicles in the previous time step according to the current position information and the actual attitude information; and calculating the information entropy distribution of the dynamic target at each position in the search area according to the number of second gas molecules and the ideal gas entropy calculation formula.

[0049] Optionally, in an embodiment of the present application, based on the change in the number of first gas molecules, calculating the change in the number of second gas molecules at each position in the search area after the search by multiple unmanned vehicles in the previous time step according to the current position information and the actual attitude information includes: according to the current position information and the actual attitude information of multiple unmanned vehicles in the previous time step, detecting whether the passable area in the search area in the previous time step is within the search range of any unmanned vehicle. If so, the number of gas molecules in the passable area becomes zero.

[0050] In an embodiment of the present application, for the case where the obstacle distribution before the search is known, it is considered that at the start of the search, the ideal gas fills the entire passable area isothermally and uniformly, and the number of gas molecules in all grids representing the passable area is N0. Thereafter, each update of the information entropy distribution {S i,j (k)} includes the following steps:

[0051] Step 11: Calculate the change in the number of gas molecules in each grid caused by the free expansion of the gas in the past time step. As Figure 2 shown, according to Equation (1), the current gas molecule number distribution {N i,j (k)} can be calculated from the gas molecule number distribution {N i,j (k - 1)} obtained from the previous update. Among them, the grids representing obstacles are not affected, and they together with the boundary of the search task area form the closed boundary of the free expansion of the ideal gas. In Equation (1), if the adjacent grid represents the ideal gas boundary, the summation term is removed. In Equation (1), Diff is an adjustable parameter related to the prior dynamic target motion ability;

[0052]

[0053] Step 12: Calculate the change in the number of gas molecules in each grid caused by the detection of vehicles in the past time step. According to the position and attitude information of each vehicle sent in the past time step, determine whether each grid representing the allowed passage area was covered by the detection range of any vehicle in the past time step. If it was covered, it is considered that the grid representing the allowed passage area was evacuated to a vacuum and the number of gas molecules became zero. The number of gas molecules in the grid representing the obstacle is always zero;

[0054] Step 13: For the number of gas molecules {N i,j (k)} in each grid obtained through the above calculation formula, calculate the information entropy magnitude {S i,j (k)} of the current target appearing in each grid according to the calculation formula of the ideal gas entropy (Equation (2)). In Equation (2), s0 is an adjustable parameter, equivalent to the zero-point molecular entropy;

[0055] S i，j = N i,j s0 - N i，j lnN i，j (2)

[0056] Step 14: Send the information entropy distribution, obstacle distribution, and the information of the numbers of other nearby vehicles near each vehicle to the corresponding vehicle.

[0057] In some embodiments, for the case where the obstacle distribution before the search is unknown, it is considered that at the start of the search, all grids in the search area represent allowed passage areas, and the ideal gas fills the entire search task area isothermally and uniformly, and the number of gas molecules in each grid in the area is N0. The update of the information entropy distribution after each time step includes the following steps:

[0058] Step 21: The same as Step 11 in the case where the obstacle distribution before the search is known;

[0059] Step 22: Calculate the change in the number of gas molecules in each grid caused by the detection of vehicles in the past time step. According to the position and attitude information of each vehicle sent in the past time step, determine whether each grid representing the allowed passage area was covered by the detection range of any vehicle in the past time step. If it was covered, it is considered that the grid representing the allowed passage area was evacuated to a vacuum and the number of gas molecules became zero. In addition, the search information fusion module needs to obtain the specific environmental perception information of each vehicle. If a grid that previously represented an allowed passage area is detected and determined to be an obstacle, its attribute is converted to a grid representing an obstacle. The number of gas molecules in the grid representing the obstacle is always zero;

[0060] Step 23: The same as step 13 when the obstacle distribution before search is known;

[0061] Step 24: The same as step 14 when the obstacle distribution before search is known.

[0062] In step S103, according to the information entropy distribution, the obstacle position information in the search area, and the relative position information between the unmanned vehicles, an optimal search path for each unmanned vehicle is generated, and multiple unmanned vehicles are controlled to perform the target search action according to the optimal search path.

[0063] It can be understood that in the embodiments of the present application, according to the information entropy distribution of the dynamic target appearing at each position in the search area, the obstacle position information in the search area, and the relative position information between the unmanned vehicles, an optimal search path for each unmanned vehicle is generated, and multiple unmanned vehicles are controlled to perform target search according to the planned optimal path. By performing real-time and effective prediction on the movement of the dynamic target, it provides a directional guidance for the search of the vehicles, which helps to improve the search efficiency of multiple unmanned vehicles for the dynamic target.

[0064] Optionally, in an embodiment of the present application, generating an optimal search path for each unmanned vehicle according to the information entropy distribution, the obstacle position information in the search area, and the relative position information between the unmanned vehicles includes: determining multiple candidate search paths in the search area according to the current position information and the actual attitude information of each unmanned vehicle; calculating the search benefits of the unmanned vehicle after multiple time steps on each candidate search path according to the information entropy distribution and the obstacle position information in the search area; calculating the search overlap range of the unmanned vehicle according to the relative position information between the unmanned vehicles, and when the search overlap range is less than or equal to the preset range threshold, using the candidate search path corresponding to the maximum search benefit as the optimal search path.

[0065] It can be understood that in most cases, the search path planning is independently completed by each search vehicle. Only when there is an overlap in the search tasks between the vehicles, the relevant vehicles perform joint path planning through communication. The planned path should ensure that the vehicle can avoid obstacles while maximizing the search efficiency.

[0066] Suppose there are n v search vehicles in the multi-vehicle system, numbered 1, 2,..., n v respectively. The entire cluster shares a unified coordinate system. The position coordinates (x, y) and the body attitude angle of the r-th vehicle after k time steps are respectively set as Furthermore, the position state of the entire system after k time steps is set as:

[0067] x(k) = {x r (k), r = 1, 2,..., n v} (3)

[0068] Specifically, in the embodiments of the present application, after every N time steps, the planning of the search path of the r-th vehicle for the next N time steps first requires setting a number of candidate paths. Currently, k time steps have passed (k is divisible by N), and the candidate paths for the next N time steps are represented by the position and attitude combinations of the vehicle after each time step in the future:

[0069] X r (k) = {x r (k + 1), x r (k + 2), ……, x r (k + N)} (4)

[0070] Secondly, based on the current information entropy distribution {N i,j (k)} and the obstacle distribution, the search gain of the vehicle after each time step in the future on each candidate path is calculated according to Equation (5). {J(x r (k + m)), m = 1, 2, …, N}. In Equation (5), {J(x r (k + m)), m = 1, 2, …, N}(k) is the average information entropy calculated based on the average number of gas molecules obtained by dividing the total number of gas molecules in the entire search task area by the total number of grids in the area, as Figure 3 shown, E(x r (l)) is the set of grids within the effective detection range of the vehicle after l time steps, and E U (x(l)) is the set of grids removed due to occlusion of the sensing signal within the theoretical maximum detection range of the vehicle after l time steps. Block is an adjustable parameter for weighing search efficiency and effective obstacle avoidance:

[0071]

[0072] If there are no other search vehicles near the vehicle, that is, when the possibility of overlap with the search tasks of other vehicles in the next N time steps is small, calculate the total search gain of each candidate path of itself and find the path with the maximum total search gain as the path planning result

[0073]

[0074] Optionally, in an embodiment of the present application, calculating the search overlap range of the unmanned vehicle according to the relative position information between unmanned vehicles includes: when the search overlap range is greater than a preset range threshold, using a multi-objective optimization algorithm to optimize the candidate search paths of multiple unmanned vehicles with the goal of maximizing search gain, and generating the optimal search path for each unmanned vehicle.

[0075] It is understandable that if there are other search vehicles near the current vehicle, that is, when there is a high possibility of overlap with the search tasks of other vehicles in the next N time steps, the involved vehicles find, through communication, the combination with the maximum total search benefit among the combinations of their respective candidate paths as the path planning result for multiple vehicles. Specifically, existing multi-objective optimization methods can be used to achieve this. In the next N time steps, the vehicles will track along the planned paths and simultaneously send their real-time position and attitude information.

[0076] It should be noted that the specific optimization method adopted in the embodiments of this application is the Nash optimal iteration method. The iterative calculation process of search path planning takes a certain amount of time. If the calculation starts only at the starting moment of the planned path, then the planning calculation process will inevitably cause a delay in the vehicle path tracking module receiving the planned path. Therefore, path planning in the embodiments should start one or two time steps in advance, and thus it is necessary to predict the positions of each vehicle and the information entropy distribution of the search area at the starting moment of each planned path.

[0077] As Figure 4 shown, the architecture of the multi-unmanned vehicle collaborative search method for dynamic targets in the embodiments of this application is presented. The execution steps of the embodiments of this application can be implemented through multiple modules, specifically including a search information fusion module, multiple path planning modules with the same number as the unmanned search vehicles, and a tracking driving module.

[0078] The search information fusion process is completed in the only information fusion module in the multi-vehicle system. The information fusion uniformly maintains the information entropy density distribution of the targets appearing at various positions in the search area. After the discretization of the search area, the information entropy density distribution is the information entropy size within each grid. Path planning is mostly completed independently by each search vehicle. Only when there is an overlap in the search tasks among the vehicles, the relevant vehicles conduct joint path planning through communication. The planned paths should ensure that the search efficiency is maximized while also ensuring that the vehicles can avoid obstacles during driving. Tracking driving means that multiple unmanned vehicles track along the planned paths and simultaneously send their real-time position and attitude information.

[0079] The following will combine the accompanying drawings and specific embodiments to elaborate in detail on the multi-unmanned vehicle collaborative search method for dynamic targets of this application. For the convenience of understanding, the embodiments of this application classify the search path planning and tracking driving into one stage.

[0080] In the embodiments of this application, the search area is a square area of 300 meters × 300 meters. As Figure 5As shown. It is divided into 300×300 grids of 1 meter×1 meter. Taking the grid center at the lower leftmost corner of the square area as the origin, a right-handed plane rectangular coordinate system is established with the positive direction to the right parallel to the lower base of the square as the x-axis, with the unit of meter. It is considered that the vehicle attitude angle is 0° when the vehicle head points to the positive direction of the x-axis. It is considered that the distribution of obstacles is completely known before the start of the search, and the specific distribution is as Figure 5 shown in the black part in Figure 5 and the rest of the passable part is as v shown in the gray part in Figure 4 . The detection range of each vehicle when there is no obstacle blocking the detection signal is a sector with an included angle of 60° and a radius of 30 meters in front of the vehicle.

[0081] 1) Search information fusion stage

[0082] The search information fusion process is completed in the only information fusion module in the multi-vehicle system. At the start of the search, i.e., k = 0, the ideal gas isothermally and uniformly fills the entire passable area, and the number of gas molecules in all grids representing the passable area is N0. Thereafter, the update of the information entropy distribution {S i,j} at each time step includes the following steps:

[0083] Input: The positions and attitudes of the two vehicles at the current moment (k)

[0084] Output: The information entropy distribution, obstacle distribution near the two vehicles, and whether there is another vehicle within a straight-line distance of 30 meters.

[0085] Step 1: Combine Figure 2 , and calculate the current gas molecule number distribution {N i,j (k)} from the gas molecule number distribution {N i,j (k - 1)} obtained from the previous update according to the following formula. In the formula, if it involves adjacent grids representing obstacles or the search area boundary, then this term of the summation is removed;

[0086]

[0087] Step 2: According to the current positions and attitudes of the two vehicles sent Change the number of gas molecules in the grids within the detection range of each search vehicle at the current moment to zero;

[0088] Step 3: According to the following formula from the gas molecule number distribution {N i,j(k)} Calculate the information entropy distribution of the current target in each grid {S i,j (k)};

[0089] S i，j = N i，j s0 - N i，j ln N i，j

[0090] Step 4: Send the information entropy distribution, obstacle distribution, and whether there is another vehicle within a straight-line distance of 30 meters near the two vehicles to the corresponding vehicles.

[0091] Empirical studies have shown that when the coefficient Diff exceeds about 0.1, the entire simulation process is difficult to converge. To further increase the diffusion rate, it is only possible to iteratively execute Equation (1) multiple times between each time step interval, which will result in a multiple increase in the computational time cost. Especially when the search area is segmented more finely, the update calculation of the entire gas free expansion process is very slow. To increase the diffusion rate on the premise of reasonable computational cost, in the embodiment, adjacent 3×3 grids are used as a finite element unit, and it is considered that the number of molecules in all 3 2 grids within this unit is the same. And in the embodiment, a strategy of gradually increasing the gas diffusion rate from zero is adopted.

[0092] 2) Search path planning and tracking driving stage

[0093] The search path planning process is mostly independently completed by the two vehicles. Only when the straight-line distance between the two vehicles is less than 30 meters, the two vehicles perform joint path planning through communication. As Figure 6 shown, it is considered that the movement of the search vehicle within each time step may be three cases: uniform left turn, uniform right turn, or uniform straight ahead with a 25 front wheel angle, and the driving speed is always 4 m / s. After every N = 3 time steps, the planning of the search paths of the two vehicles in the next 3 time steps includes the following steps:

[0094] Input: The information entropy distribution, obstacle distribution, and whether there is another vehicle within a straight-line distance of 30 meters near the two vehicles at the current moment (k);

[0095] Output: The search paths of the two vehicles in the next 3 time steps and the real-time positions and postures during the actual tracking driving process of the two vehicles at the corresponding times thereafter.

[0096] Step 1: List 3 3 candidate paths that the two vehicles may take in the next 3 time steps, in the form of X r (k) = {x r (k + 1), x r (k + 2), x r (k + 3)};

[0097] Step 2: Based on the current information entropy distribution {Ni,j(k)} and the obstacle distribution, calculate the search gain {J(x r (k+m)) of the vehicle after each time step within the next 3 time steps for each candidate path according to the following formula;

[0098]

[0099] Step 3: If the straight-line distance between the two vehicles is greater than 30 meters, the two vehicles calculate the total search gain of each of their candidate paths respectively according to the following formula, and find the path with the maximum total search gain as the path planning result

[0100]

[0101] Step 4: If the straight-line distance between the two vehicles is less than 30 meters, find the combination with the maximum total search gain in the combination of their respective candidate paths through communication as the path planning result for the multi-vehicle. The specific optimization method adopted in the embodiment is the Nash optimal iteration method;

[0102] Step 5: The trajectory tracking modules of the two vehicles perform trajectory tracking according to the planned path within the next 3 time steps, and at the same time send their own real-time position and attitude information to the search information fusion module.

[0103] Figure 7 and Figure 8 shows the situation where there is no target in the search area. The complete search paths of the two vehicles that have been planned currently are obtained every 50 time steps. The gray background in the figure reflects the distribution of the number density of ideal gas molecules in the search area at this time. The darker the color, the higher the number density.

[0104] Figure 9 shows the situation where there is a target moving randomly in the area. The complete search paths of the two vehicles that have been planned currently are obtained every 50 time steps until the target is found.

[0105] A multi-unmanned vehicle collaborative search method for dynamic targets proposed according to an embodiment of the present application calculates and updates the information entropy distribution of the current dynamic target at various positions in the entire search area based on the real-time positions and attitude information of each unmanned vehicle. Each unmanned vehicle calculates and plans its own driving path for the next period of time according to the information entropy distribution of the target appearance in the local search area around the vehicle itself, the distribution of nearby obstacles, and the positions of nearby other unmanned vehicles. Each unmanned vehicle performs path tracking according to the currently planned path, realizing the efficient collaborative search of the multi-vehicle system for dynamic targets in a certain search area. Thus, the problems in the related technology that do not consider the obstacle avoidance problem during the search process, the planned path is relatively simple, and it is not applicable to the application scenarios of multi-ground search are solved.

[0106] Next, a multi-unmanned vehicle collaborative search device for dynamic targets proposed according to an embodiment of the present application is described with reference to the accompanying drawings.

[0107] Figure 10 It is a block diagram of a multi-unmanned vehicle collaborative search device for dynamic targets according to an embodiment of the present application.

[0108] As Figure 10 shown, the multi-unmanned vehicle collaborative search device 10 for dynamic targets includes: an acquisition module 100, a generation module 200, and a search module 300.

[0109] Among them, the acquisition module 100 is used to acquire the current position information, actual attitude information of multiple unmanned vehicles, and the relative position information between unmanned vehicles. The generation module 200 is used to calculate the information entropy distribution of the dynamic target at each position in the search area according to the current position information and actual attitude information. The search module 300 is used to generate the optimal search path for each unmanned vehicle according to the information entropy distribution, the obstacle position information in the search area, and the relative position information between unmanned vehicles, and control multiple unmanned vehicles to perform target search actions according to the optimal search path.

[0110] Optionally, in an embodiment of the present application, the generation module 200 is further used to calculate the change in entropy distribution during the adiabatic free expansion process of an ideal gas in the search area; generate the information entropy distribution of the dynamic target at each position in the search area according to the change in entropy distribution.

[0111] Optionally, in an embodiment of the present application, the generation module 200 includes: a first calculation unit for calculating the change in the number of first gas molecules at each position in the search area after the free expansion of the gas in the previous time step; a second calculation unit for calculating the change in the number of second gas molecules at each position in the search area after the search by multiple unmanned vehicles in the previous time step based on the change in the number of first gas molecules, according to the current position information and actual attitude information.

[0112] Optionally, in an embodiment of the present application, the second calculation unit is further configured to detect whether the passable area of the search area in the previous time step is within the search range of any unmanned vehicle according to the current position information and actual attitude information of multiple unmanned vehicles in the previous time step. If so, the number of gas molecules in the passable area becomes zero.

[0113] Optionally, in an embodiment of the present application, the search module 300 includes: a determination unit configured to determine multiple candidate search paths within the search area according to the current position information and actual attitude information of each unmanned vehicle; a third calculation unit configured to calculate the search benefits of the unmanned vehicle after multiple time steps on each candidate search path according to the information entropy distribution and the obstacle position information within the search area; and a fourth calculation unit configured to calculate the search overlap range of the unmanned vehicles according to the relative position information between the unmanned vehicles. When the search overlap range is less than or equal to the preset range threshold, the candidate search path corresponding to the maximum search benefit is used as the optimal search path.

[0114] Optionally, in an embodiment of the present application, the fourth calculation unit is further configured to, when the search overlap range is greater than the preset range threshold, optimize the candidate search paths of multiple unmanned vehicles by using a multi-objective optimization algorithm with the goal of maximizing the search benefit, and generate the optimal search path for each unmanned vehicle.

[0115] It should be noted that the foregoing explanation of the embodiments of the multi-unmanned vehicle cooperative search method for a dynamic target also applies to the multi-unmanned vehicle cooperative search method device for a dynamic target in this embodiment, and will not be elaborated here.

[0116] A multi-unmanned vehicle cooperative search method device for a dynamic target proposed according to an embodiment of the present application calculates and updates the information entropy distribution of the current dynamic target at various positions within the entire search area based on the real-time positions and attitude information of each unmanned vehicle. Each unmanned vehicle calculates and plans its own driving path for the next period of time according to the information entropy distribution of the target appearance within the local search area around the vehicle itself, the distribution of nearby obstacles, and the positions of nearby other unmanned vehicles. Each unmanned vehicle performs path tracking according to the currently planned path, realizing the efficient cooperative search of the multi-vehicle system for the dynamic target within a certain search area. Thus, the problems in the related art that do not consider obstacle avoidance during the search process, the planned path is relatively simple, and it is not applicable to the application scenarios of multi-ground search are solved.

[0117] Figure 11 The structural schematic diagram of the electronic device provided for the embodiment of the present application. The electronic device may include:

[0118] A memory 1101, a processor 1102, and a computer program stored on the memory 1101 and executable on the processor 1102.

[0119] When the processor 1102 executes a program, it implements the multi-unmanned vehicle collaborative search method for dynamic targets provided in the above embodiments.

[0120] Furthermore, the electronic device further includes:

[0121] A communication interface 1103 for communication between the memory 1101 and the processor 1102.

[0122] A memory 1101 for storing a computer program that can run on the processor 1102.

[0123] The memory 1101 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory.

[0124] If the memory 1101, the processor 1102, and the communication interface 1103 are implemented independently, the communication interface 1103, the memory 1101, and the processor 1102 can be interconnected through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, Figure 11 only a thick line is shown in the figure, but it does not mean that there is only one bus or one type of bus.

[0125] Optionally, in a specific implementation, if the memory 1101, the processor 1102, and the communication interface 1103 are integrated on a chip, the memory 1101, the processor 1102, and the communication interface 1103 can communicate with each other through an internal interface.

[0126] The processor 1102 may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0127] This embodiment also provides a computer-readable storage medium, on which a computer program is stored, and the program is characterized in that when it is executed by a processor, it implements the above multi-unmanned vehicle collaborative search method for dynamic targets.

[0128] In the description of this specification, the descriptions with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0129] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0130] Any process or method description in the flowchart or described in other ways herein can be understood to represent a module, segment, or part of code including one or more N executable instructions for implementing a customized logic function or process, and the scope of the preferred embodiments of the present application includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in a reverse order according to the involved functions, rather than in the order shown or discussed, which should be understood by those skilled in the art of the embodiments of the present application.

[0131] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0132] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of implementing the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

Claims

1. A multi-unmanned vehicle collaborative search method for dynamic targets, characterized in that Including the following steps: Obtain the current position information, actual attitude information of multiple unmanned vehicles, and relative position information between the unmanned vehicles; Calculate the information entropy distribution of the dynamic target appearing at each position within the search area based on the current position information and the actual attitude information; Generate the optimal search path for each unmanned vehicle according to the information entropy distribution, the obstacle position information within the search area, and the relative position information between the unmanned vehicles, and control the multiple unmanned vehicles to perform the target search action according to the optimal search path; The calculating the information entropy distribution of the dynamic target appearing at each position within the search area based on the current position information and the actual attitude information includes: calculating the change in the number of first gas molecules at each position within the search area after the free expansion of the gas in the previous time step; based on the change in the number of first gas molecules, calculating the change in the number of second gas molecules at each position within the search area after the search by the multiple unmanned vehicles in the previous time step according to the current position information and the actual attitude information; calculating the information entropy distribution of the dynamic target appearing at each position within the search area according to the number of second gas molecules and the ideal gas entropy calculation formula; The calculating the change in the number of second gas molecules at each position within the search area after the search by the multiple unmanned vehicles in the previous time step based on the change in the number of first gas molecules and according to the current position information and the actual attitude information includes: according to the current position information and the actual attitude information of the multiple unmanned vehicles in the previous time step, detecting whether the allowable passage area within the search area in the previous time step is within the search range of any unmanned vehicle, and if so, changing the number of gas molecules in the allowable passage area to zero.

2. The method according to claim 1, wherein The calculating the information entropy distribution of the dynamic target appearing at each position within the search area includes: Calculating the change in the entropy distribution during the adiabatic free expansion process of the ideal gas within the search area; Generating the information entropy distribution of the dynamic target appearing at each position within the search area according to the change in the entropy distribution; 3. The method according to claim 1, wherein Generating the optimal search path for each unmanned vehicle according to the information entropy distribution, the obstacle position information within the search area, and the relative position information between the unmanned vehicles includes: Determining multiple candidate search paths within the search area according to the current position information and the actual attitude information of each unmanned vehicle; Calculating the search gain of the unmanned vehicle after multiple time steps on each candidate search path according to the information entropy distribution and the obstacle position information within the search area; Calculating the search overlap range of the unmanned vehicle according to the relative position information between the unmanned vehicles, and when the search overlap range is less than or equal to the preset range threshold, taking the candidate search path corresponding to the maximum search gain as the optimal search path.

4. The method according to claim 3, wherein It also includes: When the search overlap range is greater than the preset range threshold, aiming at maximizing the search gain, using a multi-objective optimization algorithm to optimize the candidate search paths of the multiple unmanned vehicles to generate the optimal search path for each unmanned vehicle.

5. A multi-unmanned vehicle collaborative search device for dynamic targets, characterized in that, Including: An acquisition module for acquiring the current position information, actual attitude information of multiple unmanned vehicles, and relative position information between the unmanned vehicles; A generation module, configured to calculate the information entropy distribution of the dynamic target at each position within the search area based on the current position information and the actual attitude information; A search module, configured to generate an optimal search path for each unmanned vehicle based on the information entropy distribution, the obstacle position information within the search area, and the relative position information between the unmanned vehicles, and control the multiple unmanned vehicles to perform the target search action according to the optimal search path; The generation module includes: a first calculation unit, configured to calculate the change in the number of first gas molecules at each position within the search area after the free expansion of the gas in the previous time step; A second calculation unit, based on the change in the number of first gas molecules, calculates the change in the number of second gas molecules at each position within the search area after the search by the multiple unmanned vehicles in the previous time step according to the current position information and the actual attitude information; The second calculation unit is further configured to detect whether the allowable passage area within the search area in the previous time step is within the search range of any unmanned vehicle according to the current position information and the actual attitude information of the multiple unmanned vehicles in the previous time step. If so, the number of gas molecules in the allowable passage area becomes zero.

6. The device according to claim 5, wherein The generation module is further configured to calculate the change in the entropy distribution during the adiabatic free expansion process of the ideal gas within the search area; generate the information entropy distribution of the dynamic target at each position within the search area according to the change in the entropy distribution.

7. The device according to claim 5, characterized in that, The search module includes: A determination unit, configured to determine multiple candidate search paths within the search area according to the current position information and the actual attitude information of each unmanned vehicle; A third calculation unit, configured to calculate the search gain of the unmanned vehicle after multiple time steps on each candidate search path according to the information entropy distribution and the obstacle position information within the search area; A fourth calculation unit, configured to calculate the search overlap range of the unmanned vehicles according to the relative position information between the unmanned vehicles. When the search overlap range is less than or equal to a preset range threshold, the candidate search path corresponding to the maximum search gain is used as the optimal search path.

8. The device according to claim 7, characterized in that, The fourth calculation unit is further configured to, when the search overlap range is greater than the preset range threshold, optimize the candidate search paths of the multiple unmanned vehicles by using a multi-objective optimization algorithm with the goal of maximizing the search gain, and generate the optimal search path for each unmanned vehicle.

9. An electronic device, characterized in that, Including: A memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement the multi-unmanned vehicle collaborative search method for the dynamic target as described in any one of claims 1-4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to be used to implement the multi-unmanned vehicle collaborative search method for the dynamic target as described in any one of claims 1-4.

Images