Unmanned aerial vehicle cooperative route planning system

Abstract

The invention discloses an unmanned aerial vehicle cooperative route planning system. The system belongs to the field of intelligent traffic and intelligent route planning and comprises a central control module, a sensing system, an algorithm module, an obstacle avoidance system, a positioning system, a power supply system, an analog simulation module, a communication and data module, a task planning module and a user interface. The central control module serves as the core of the system, is responsible for coordinating the sensing system, the algorithm module and the obstacle avoidance system, and realizes data fusion and provides an optimal route through data integration and screening. The sensing system is provided with temperature, humidity and wind speed sensors so as to monitor and adapt to external weather conditions in real time; the algorithm module adopts a DFS algorithm and a BFS algorithm, the two graph search algorithms can plan paths and explore problems, and support is provided for collaborative inspection of the unmanned aerial vehicle in a complex environment. Through the system, the DFS algorithm and the BFS algorithm can cooperate with a plurality of unmanned aerial vehicles.

Description

Technical Field

[0001] The invention belongs to the field of intelligent transportation and intelligent route planning, and relates to a drone collaborative route planning system. Background Art

[0002] At present, the application of drone technology has become an important force in promoting the digital construction and intelligent management of infrastructure. The application scope of drone technology is wide, covering multiple key areas such as power grid, water conservancy, transportation, agriculture and energy. Among them, the strategy of combining depth-first search (DFS) and breadth-first search (BFS) can improve efficiency in some search scenarios, but it also introduces additional complexity. This combination requires developers to manage two search queues or stacks at the same time, which not only increases space consumption, but also may make the implementation and maintenance of the algorithm more difficult, especially when dealing with large-scale graphs or complex spaces. In certain cases, such as when the target node is close to the starting point or hidden in the deep structure of the graph, it may be more efficient to use BFS or DFS algorithms alone. In addition, not all problems are suitable for this combination strategy, and some may require specific search algorithms. In a dynamically changing environment, it is a challenge to keep the synchronous progress of bidirectional searches, and it becomes more complicated to determine the appropriate termination conditions. For developers who are not familiar with DFS and BFS algorithms, understanding and implementing this combination algorithm undoubtedly increases the technical difficulty. Therefore, although the search strategy combining DFS and BFS algorithms has its advantages, the complexity and applicability it brings need to be carefully weighed in practical applications. Summary of the invention

[0003] In view of the above problems, the purpose of the present invention is to propose a UAV collaborative route planning system, which uses sensors, navigation systems, DFS and BFS algorithms to calculate and plan the most reasonable UAV patrol route, so as to maximize the patrol efficiency, minimize energy consumption, and reduce personnel safety risks.

[0004] The technical solution of the present invention is: the core advantage of the cooperative route planning system for unmanned aerial vehicles described in the present invention lies in its highly integrated multifunctional module design; the system includes a central control module to achieve comprehensive scheduling, a sensor system for environmental perception, a route planning calculation module responsible for providing accurate path calculation, an obstacle avoidance system to ensure flight safety, a positioning system to ensure navigation accuracy, power supply, and a simulation module for pre-testing, a communication and data module responsible for efficient information exchange, a task planning module responsible for reasonable task allocation, and a user interface that provides an intuitive and easy-to-operate interactive platform;

[0005] The power supply module uses solar power generation technology to provide continuous power for the drone, and maximizes the efficiency of power utilization through power calculation and distribution mechanism. The simulation module conducts comprehensive analysis on humidity, temperature, wind speed and other data collected by the sensor module, calculates the real-time weather conditions of the inspected area, and transmits it to the central control module; at the same time, the module also transmits the real-time road condition information collected by the obstacle avoidance system to the central control module; the optimal path calculated by the path planning algorithm module will also be sent to the central control module, which will then transmit the data to the simulation module for data fusion, thereby simulating a flight path that is both efficient and low-risk;

[0006] In the path planning calculation module, the system specially sets up depth-first search (DFS) and breadth-first search (BFS) algorithms, which play a vital role in the wide application of multi-UAV collaborative inspections; these algorithms can quickly plan efficient flight routes, save time and costs, and significantly improve the safety and efficiency of inspection work; in complex environments, the multi-UAV collaborative path planning problem is a challenging task, which involves path planning for multiple UAVs when performing tasks, ensuring the efficiency, safety of task execution and smooth collaborative cooperation.

[0007] Furthermore, the system can use the DFS-BFS intelligent optimization algorithm to accurately calculate and evaluate the coordinated route planning of drones based on the distribution locations of multiple working drones and the distribution locations of drones in standby state, combined with the location information of the equipment to be inspected, the location information of the equipment being or having been inspected, the flight channel image data and its flight distance data; the specific steps of this process are as follows:

[0008] Step (1): Establish a route judgment function:

[0009] S=w1L min +w2T min +w3E min +w4O min +w5F eff +w6S saf

[0010] This formula is the total function of the optimal path;

[0011] Where w1, w2, w3, w4, w5, w6 are weight coefficients, L min is the path length, T min is the flight time, E min is the energy consumption, O min is the number of obstacles, F eff is the flight efficiency, S saf is the safety score;

[0012] Specifically:

[0013] (1.1),

[0014] Where L is the path length, (x i ,y i ) device location coordinates, the path length is defined as the straight-line distance from the starting point to the end point or the length of the actual flight path;

[0015] (1.2),

[0016] Flight time is defined as the minimum time required for a drone to travel from its starting point to its destination;

[0017] In the formula, v i is the speed of the drone on the i-th path, (x i ,y i ) device location coordinates;

[0018] (1.3),

[0019] Energy consumption is defined as the total energy consumed by the drone during flight.

[0020] Where k1, k2, k3 are coefficients related to flight distance, speed and basic energy consumption, e i represents the energy consumption of the i-th path;

[0021] (1.4),

[0022] The number of obstacles is defined as the total number of obstacles on the path,

[0023] In the formula, o i is the number of obstacles on the i-th path;

[0024] (1.5),

[0025] Flight efficiency is the ratio of path length to flight time.

[0026] In the formula, the sensitivity of efficiency is adjusted by θ1, F eff It is flight efficiency;

[0027] (1.6),

[0028] The safety score is the ratio of energy consumption to the number of obstacles.

[0029] In the formula, the sensitivity of security is adjusted by θ2, S saf is the safety score;

[0030] Step (2): Assume there are n drones, and the state of the jth drone is represented by x j (j=1,2,...,n) means that if when t→∞, there is The system is said to have reached consistency;

[0031] The matrix model communication topology is used to maintain the consistency of multiple drones; the operation is as follows:

[0032] Graph G = (V, ε);

[0033] Nodes V = {v1,v2,v3,v4,...,v i};

[0034] Edges ε={(v i ,v j ):i≠j;v i ,v j ∈V};

[0035] v i It is called v j If (v i ,v j )∈ε;

[0036] Weight, In-degree, Adjacency matrix, Spanning tree;

[0037]

[0038] In order to make the states of each drone of multiple drones in the whole system tend to the same value, a control system control protocol is used;

[0039]

[0040] In the formula, is the latest sampling time of drone j, u i represents the control input acting on the ith agent (e.g., drone), k i is the control gain, a ij is an element in the adjacency matrix, indicating the degree of influence of agent j on agent i, x j (t) and x i (t) are the states of agents j and i at time t, k is the control gain, and denote the sampling time of agents j and i respectively, where and is the state transfer matrix, which describes the change of state from the sampling time to the current time;

[0041] Step (3): Through these formulas, the system can comprehensively evaluate the coordinated routes of drones to ensure that when performing tasks, not only efficiency and cost are considered, but also flight safety and environmental complexity are taken into account;

[0042] First, the environmental conditions within a certain range are collected through sensors. In the coordinated route planning of UAVs, sensor system algorithms and data fusion are key technologies, which together ensure that UAVs can accurately perceive the environment and make correct decisions; specifically:

[0043] (3.1) Use the BFS algorithm to perform an initial breadth search within a certain range to find all nearby device nodes. The BFS algorithm searches all nodes layer by layer to find the shortest path;

[0044] The BFS algorithm visits all adjacent nodes of the current vertex before further traversing the vertices in the graph; its operation steps are as follows:

[0045] (3.1.1) First, select a vertex as the starting node and color it gray, and the rest of the nodes white;

[0046] (3.1.2) Put the starting node into the queue;

[0047] (3.1.3) Select a vertex from the head of the queue and find all the nodes adjacent to it. Put the found adjacent nodes at the end of the queue. Color the visited nodes black and the unvisited nodes white. If the vertex is gray, it means it has been found and put into the queue. If the vertex is white, it has not been found yet.

[0048] (3.1.4) Process the next node in the queue in the same way. Basically, the vertices that are out of the queue become black, the ones in the queue are gray, and the ones that have not yet entered the queue are white. Specifically:

[0049] a. Initial state, starting from vertex 1, queue = {1};

[0050] b. Visit the adjacent vertex of 1, 1 is dequeued and turned black, 2 and 3 are enqueued, and the queue = {2, 3,};

[0051] c. Visit the adjacent node of 2, 2 is dequeued, 4 is enqueued, and the queue = {3, 4};

[0052] d. Visit the adjacent node of 3, 3 is dequeued, queue = {4};

[0053] e. Visit the adjacent node of 4, 4 is dequeued, queue = {empty};

[0054] If the nodes of the graph represent points in two-dimensional or three-dimensional space, the distance between nodes is calculated using the Euclidean distance; that is:

[0055] L Manhattan (P i ,P j )=|x i -x j ∣+∣y i -y j |

[0056]

[0057] Where, L Euclidean is the Euclidean straight line distance between two points, L Manhattan is the Manhattan straight-line distance between two points, (x i ,y i ) device location coordinates;

[0058] (3.2), and then use the DFS algorithm to deeply explore each neighboring node to find the possible optimal path, P i and P j represents a point (or node) in two-dimensional or three-dimensional space, (x i ,y i ) device location coordinates, L is the path length;

[0059] After visiting a vertex during the search process, the DFS algorithm needs to recursively visit all the unvisited adjacent vertices of this vertex;

[0060] Initially, all nodes are white. Select one as the starting vertex and traverse it as follows:

[0061] (3.2.1), select the starting vertex and paint it gray, indicating that it has not been visited yet;

[0062] (3.2.2) Select one of the adjacent vertices of the vertex and continue this process (i.e., find the adjacent nodes of the adjacent nodes) until a vertex has no adjacent nodes. Then, paint it black to indicate that it has been visited.

[0063] (3.2.3) Go back to the vertex on the previous layer of the blackened vertex, and then find the remaining adjacent nodes of the vertex on the previous layer, and continue the above operation. If all adjacent nodes have been visited, then blacken yourself and go back to the previous layer; continue the above operation on the previous layer until all vertices have been visited; the specific steps are as follows:

[0064] a. Initial state, starting from vertex 1;

[0065] b. After visiting vertices 1, 2, and 3 in sequence, it ends at vertex 3;

[0066] c. Backtrack from vertex 3 to vertex 2, continue to visit vertex 5, and end at vertex 5;

[0067] d. Backtrack from vertex 5 to vertex 2 and terminate at the vertex;

[0068] e. Backtrack from vertex 2 to vertex 1 and end at vertex 1;

[0069] f. Start visiting from vertex 4 and end at vertex 4;

[0070] (3.2.4) Start depth search from vertex 1; the specific steps are as follows:

[0071] a. Initial state, starting from vertex 1;

[0072] b. After visiting vertices 1, 2, and 3 in sequence, it ends at vertex 3;

[0073] c. Backtrack from vertex 3 to vertex 2, continue to visit vertex 5, and end at vertex 5;

[0074] d. Backtrack from vertex 5 to vertex 2 and end at vertex 2;

[0075] e. Backtrack from vertex 2 to vertex 1 and end at vertex 1;

[0076] f. Start visiting from vertex 4 and end at vertex 4;

[0077] Step (4): In the breadth-first search (BFS) algorithm, the performance of the algorithm is improved by introducing a mathematical function, as shown in the following formula:

[0078] C(u)=C(Parent(u))+d(Parent(u),u)

[0079] Optimal Path = argmin Path(u) C(u) for all u∈goal nodes

[0080] Where C(u) is the cost to reach node u, is the cost to reach the parent node, d(Parent(u),u) is the distance from the parent node to the node; where Optimal Path is the path with the least cost, C(u) is the cost of the path, and Optimal Path is used to describe the goal of the path-finding algorithm;

[0081] Step (5): Use DFS and BFS algorithms to search from the starting point and the end point at the same time, which can reduce the search space and may find a solution faster;

[0082] First initialize the bidirectional search:

[0083] Q start ={s},Q end ={t}

[0084] In the formula, Q start As a starting point, Q end for the end point;

[0085] Step (6): Use the BFS algorithm to mark the levels:

[0086] L start (u)=level of node u from start in BFS tree

[0087] L end (u)=level of node from end in BFS tree

[0088] Where, L start (u) and L end (u) represents the level of the BFA algorithm tree of node u starting from the starting point and the end point, respectively;

[0089] Step (7): Use the DFS algorithm to perform recursive exploration:

[0090]

[0091] Iterate from the starting point and the end point. The iteration process is as follows:

[0092]

[0093] In the formula, Q start is the queue used when searching from the starting point (source node), Q end The function of is to store the nodes to be explored starting from the end point. The dequeue() operation is a standard operation in breadth-first search. start Indicates the node currently being processed in a breadth-first search (BFS) or depth-first search (DFS) starting from the starting point, u end Indicates the node currently being processed during the search process starting from the end point, V start Usually represents the set of nodes that have been visited when searching from the starting point, V end Represents the set of nodes that have been visited when searching from the end point. enqueue() is used to add a node to the end of the queue, which means that the node is marked as pending and will be taken out for exploration at some point in the future. start (u) and C end(u) represents the cost from the starting point to the end point u and from node u to the end point respectively;

[0094] Step (8): Optimal Path = arg min u∈Qmeet (C start (u)+C end (u))

[0095] Select the best path from them;

[0096] In the formula, Q meet It is the set of nodes encountered during the search of the starting point and the end point;

[0097] Step (9): When there are common nodes in the search queues of the starting point and the end point, the search terminates; because this means that a path from the starting point to the end point has been found, as shown in the following formula:

[0098]

[0099] Where: Q start As a starting point, Q end for the end point;

[0100] The above algorithm selects the optimal path.

[0101] Furthermore, after completing the data integration, the system can list a variety of route planning schemes and select the best path through the optimal decision-making algorithm to ensure that the equipment that needs to be inspected is efficiently inspected.

[0102] The beneficial effects of the present invention are as follows: 1. The DFS algorithm can deeply explore each possible path until the target is found or it cannot continue, which helps to quickly locate the target in a complex environment, and the BFS algorithm can systematically explore all neighboring nodes to ensure that no possible path is missed, which is particularly important for finding the shortest path; 2. Combining the DFS and BFS algorithms can balance the depth and breadth of the path exploration, and improve the accuracy and efficiency of path planning. Through the deep exploration of the DFS algorithm and the breadth coverage of the BFS algorithm, the advantages and disadvantages of different paths can be more comprehensively evaluated, so as to select the optimal path; 3. Optimized path planning can reduce the flight distance and time of the drone, thereby reducing energy consumption and operating costs. Through accurate path planning, unnecessary flight and energy waste can be reduced, and energy utilization efficiency can be improved; 4. Through the collaborative work of the DFS and BFS algorithms, the drone can better avoid obstacles and threat areas and improve flight safety. The algorithm can help the drone plan a safe path in a complex environment to avoid collisions with obstacles or other drones; 5. In a multi-drone system, the combination of the DFS and BFS algorithms can improve the collaborative efficiency between drones and ensure the collaborative execution of tasks. Through optimized path planning, drones can share information, coordinate task allocation, and plan paths more effectively; 6. The path planning algorithm that combines DFS and BFS algorithms can quickly adapt to changes in task requirements and improve the flexibility of task execution. The algorithm can quickly respond to new task requirements, adjust path planning in a timely manner, and ensure timely completion of tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] Figure 1 It is a flow chart of UAV collaborative route planning in the present invention;

[0104] Figure 2 It is a flowchart of algorithm optimization in an embodiment of the present invention. DETAILED DESCRIPTION

[0105] The specific technical scheme of the present invention is further described in detail below with reference to specific examples.

[0106] As shown in the figure, this embodiment provides a UAV collaborative route planning system according to the present invention, and its core advantage lies in its highly integrated multifunctional module design; the system includes a central control module to achieve comprehensive scheduling, a sensor system for environmental perception, a route planning calculation module responsible for providing accurate path calculation, an obstacle avoidance system to ensure flight safety, a positioning system to ensure navigation accuracy, power supply, a simulation module for pre-testing, a communication and data module responsible for efficient information exchange, a task planning module responsible for reasonable task allocation, and a user interface that provides an intuitive and easy-to-operate interactive platform;

[0107] The power supply module uses solar power generation technology to provide continuous power for the drone, and maximizes the efficiency of power utilization through power calculation and distribution mechanism. The simulation module conducts comprehensive analysis on humidity, temperature, wind speed and other data collected by the sensor module, calculates the real-time weather conditions of the inspected area, and transmits it to the central control module; at the same time, the module also transmits the real-time road condition information collected by the obstacle avoidance system to the central control module; the optimal path calculated by the path planning algorithm module will also be sent to the central control module, which will then transmit the data to the simulation module for data fusion, thereby simulating a flight path that is both efficient and low-risk;

[0108] In the path planning calculation module, the system specially sets up depth-first search (DFS) and breadth-first search (BFS) algorithms, which play a vital role in the wide application of multi-UAV collaborative inspections; these algorithms can quickly plan efficient flight routes, save time and costs, and significantly improve the safety and efficiency of inspection work; in complex environments, the multi-UAV collaborative path planning problem is a challenging task, which involves path planning for multiple UAVs when performing tasks, ensuring the efficiency, safety of task execution and smooth collaborative cooperation.

[0109] Furthermore, the system can use the DFS-BFS intelligent optimization algorithm to accurately calculate and evaluate the coordinated route planning of drones based on the distribution locations of multiple working drones and the distribution locations of drones in standby state, combined with the location information of the equipment to be inspected, the location information of the equipment being or having been inspected, the flight channel image data and its flight distance data; the specific steps of this process are as follows:

[0110] Step (1): Establish a route judgment function:

[0111] S=w1L min +w2T min +w3E min +w4O min +w5F eff +w6S saf

[0112] This formula is the total function of the optimal path;

[0113] Where w1, w2, w3, w4, w5, w6 are weight coefficients, L min is the path length, T min is the flight time, E min is the energy consumption, O min is the number of obstacles, F eff is the flight efficiency, S saf is the safety score;

[0114] Specifically:

[0115] (1.1),

[0116] Where L is the path length, (x i ,y i ) device location coordinates, the path length is defined as the straight-line distance from the starting point to the end point or the length of the actual flight path;

[0117] (1.2),

[0118] Flight time is defined as the minimum time required for a drone to travel from its starting point to its destination;

[0119] In the formula, v i is the speed of the drone on the i-th path, (x i ,y i ) device location coordinates;

[0120] (1.3),

[0121] Energy consumption is defined as the total energy consumed by the drone during flight.

[0122] Where k1, k2, k3 are coefficients related to flight distance, speed and basic energy consumption, e i represents the energy consumption of the i-th path;

[0123] (1.4),

[0124] The number of obstacles is defined as the total number of obstacles on the path,

[0125] In the formula, o i is the number of obstacles on the i-th path;

[0126] (1.5),

[0127] Flight efficiency is the ratio of path length to flight time.

[0128] In the formula, the sensitivity of efficiency is adjusted by θ1, F eff It is flight efficiency;

[0129] (1.6),

[0130] The safety score is the ratio of energy consumption to the number of obstacles.

[0131] In the formula, the sensitivity of security is adjusted by θ2, S saf is the safety score;

[0132] Step (2): Assume there are n drones, and the state of the jth drone is represented by x j (j=1,2,...,n) means that if when t→∞, there is The system is said to have reached consistency;

[0133] The matrix model communication topology is used to maintain the consistency of multiple drones; the operation is as follows:

[0134] Graph G = (V, ε);

[0135] Nodes V = {v1,v2,v3,v4,...,v i};

[0136] Edges ε={(v i ,v j ):i≠j;v i ,v j ∈V};

[0137] v i It is called v j If (v i ,v j )∈ε;

[0138] Weight, In-degree, Adjacency matrix, Spanning tree;

[0139]

[0140] In order to make the states of each drone of multiple drones in the whole system tend to the same value, a control system control protocol is used;

[0141]

[0142] In the formula, is the latest sampling time of drone j, u i represents the control input acting on the ith agent (e.g., drone), k i is the control gain, a ij is an element in the adjacency matrix, indicating the degree of influence of agent j on agent i, x j (t) and x i (t) are the states of agents j and i at time t, k is the control gain, and denote the sampling time of agents j and i respectively, where and is the state transfer matrix, which describes the change of state from the sampling time to the current time;

[0143] Step (3): Through these formulas, the system can comprehensively evaluate the coordinated routes of drones to ensure that when performing tasks, not only efficiency and cost are considered, but also flight safety and environmental complexity are taken into account;

[0144] First, the environmental conditions within a certain range are collected through sensors. In the coordinated route planning of UAVs, sensor system algorithms and data fusion are key technologies, which together ensure that UAVs can accurately perceive the environment and make correct decisions; specifically:

[0145] (3.1) Use the BFS algorithm to perform an initial breadth search within a certain range to find all nearby device nodes. The BFS algorithm searches all nodes layer by layer to find the shortest path;

[0146] The BFS algorithm visits all adjacent nodes of the current vertex before further traversing the vertices in the graph; its operation steps are as follows:

[0147] (3.1.1) First, select a vertex as the starting node and color it gray, and the rest of the nodes white;

[0148] (3.1.2) Put the starting node into the queue;

[0149] (3.1.3) Select a vertex from the head of the queue and find all the nodes adjacent to it. Put the found adjacent nodes at the end of the queue. Color the visited nodes black and the unvisited nodes white. If the vertex is gray, it means it has been found and put into the queue. If the vertex is white, it has not been found yet.

[0150] (3.1.4) Process the next node in the queue in the same way. Basically, the vertices that are out of the queue become black, the ones in the queue are gray, and the ones that have not yet entered the queue are white. Specifically:

[0151] a. Initial state, starting from vertex 1, queue = {1};

[0152] b. Visit the adjacent vertex of 1, 1 is dequeued and turned black, 2 and 3 are enqueued, and the queue = {2, 3,};

[0153] c. Visit the adjacent node of 2, 2 is dequeued, 4 is enqueued, and the queue = {3, 4};

[0154] d. Visit the adjacent node of 3, 3 is dequeued, queue = {4};

[0155] e. Visit the adjacent node of 4, 4 is dequeued, queue = {empty};

[0156] If the nodes of the graph represent points in two-dimensional or three-dimensional space, the distance between nodes is calculated using the Euclidean distance; that is:

[0157] L Manhattan (P i ,P j )=|x i -x j ∣+∣y i -y j |

[0158]

[0159] Where, L Euclidean is the Euclidean straight line distance between two points, L Manhattan is the Manhattan straight-line distance between two points, (x i ,y i ) device location coordinates;

[0160] (3.2), and then use the DFS algorithm to deeply explore each neighboring node to find the possible optimal path, P i and P j represents a point (or node) in two-dimensional or three-dimensional space, (x i ,y i ) device location coordinates, L is the path length;

[0161] After visiting a vertex during the search process, the DFS algorithm needs to recursively visit all the unvisited adjacent vertices of this vertex;

[0162] Initially, all nodes are white. Select one as the starting vertex and traverse it as follows:

[0163] (3.2.1), select the starting vertex and paint it gray, indicating that it has not been visited yet;

[0164] (3.2.2) Select one of the adjacent vertices of the vertex and continue this process (i.e., find the adjacent nodes of the adjacent nodes) until a vertex has no adjacent nodes. Then, paint it black to indicate that it has been visited.

[0165] (3.2.3) Go back to the vertex on the previous layer of the blackened vertex, and then find the remaining adjacent nodes of the vertex on the previous layer, and continue the above operation. If all adjacent nodes have been visited, then blacken yourself and go back to the previous layer; continue the above operation on the previous layer until all vertices have been visited; the specific steps are as follows:

[0166] a. Initial state, starting from vertex 1;

[0167] b. After visiting vertices 1, 2, and 3 in sequence, it ends at vertex 3;

[0168] c. Backtrack from vertex 3 to vertex 2, continue to visit vertex 5, and end at vertex 5;

[0169] d. Backtrack from vertex 5 to vertex 2 and terminate at the vertex;

[0170] e. Backtrack from vertex 2 to vertex 1 and end at vertex 1;

[0171] f. Start visiting from vertex 4 and end at vertex 4;

[0172] (3.2.4) Start depth search from vertex 1; the specific steps are as follows:

[0173] a. Initial state, starting from vertex 1;

[0174] b. After visiting vertices 1, 2, and 3 in sequence, it ends at vertex 3;

[0175] c. Backtrack from vertex 3 to vertex 2, continue to visit vertex 5, and end at vertex 5;

[0176] d. Backtrack from vertex 5 to vertex 2 and end at vertex 2;

[0177] e. Backtrack from vertex 2 to vertex 1 and end at vertex 1;

[0178] f. Start visiting from vertex 4 and end at vertex 4;

[0179] Step (4): In the breadth-first search (BFS) algorithm, the performance of the algorithm is improved by introducing a mathematical function, as shown in the following formula:

[0180] C(u)=C(Parent(u))+d(Parent(u),u)

[0181] Optimal Path = argmin Path(u) C(u) for all u∈goal nodes

[0182] Where C(u) is the cost to reach node u, is the cost to reach the parent node, d(Parent(u),u) is the distance from the parent node to the node; where Optimal Path is the path with the least cost, C(u) is the cost of the path, and Optimal Path is used to describe the goal of the path-finding algorithm;

[0183] Step (5): Use DFS and BFS algorithms to search from the starting point and the end point at the same time, which can reduce the search space and may find a solution faster;

[0184] First initialize the bidirectional search:

[0185] Q start ={s},Q end ={t}

[0186] In the formula, Q start As a starting point, Q end for the end point;

[0187] Step (6): Use the BFS algorithm to mark the levels:

[0188] L start (u)=level of node u from start in BFS tree

[0189] L end (u)=level of node from end in BFS tree

[0190] Where, L start (u) and L end (u) represents the level of the BFA algorithm tree of node u starting from the starting point and the end point, respectively;

[0191] Step (7): Use the DFS algorithm to perform recursive exploration:

[0192]

[0193] Iterate from the starting point and the end point. The iteration process is as follows:

[0194]

[0195] In the formula, Q start is the queue used when searching from the starting point (source node), Q end The function of is to store the nodes to be explored starting from the end point. The dequeue() operation is a standard operation in breadth-first search. start Indicates the node currently being processed in a breadth-first search (BFS) or depth-first search (DFS) starting from the starting point, u end Indicates the node currently being processed during the search process starting from the end point, V start Usually represents the set of nodes that have been visited when searching from the starting point, V end Represents the set of nodes that have been visited when searching from the end point. enqueue() is used to add a node to the end of the queue, which means that the node is marked as pending and will be taken out for exploration at some point in the future. start (u) and C end(u) represents the cost from the starting point to the end point u and from node u to the end point respectively;

[0196] Step (8): Optimal Path = arg min u∈Qmeet (C start (u)+C end (u))

[0197] Select the best path from them;

[0198] In the formula, Q meet It is the set of nodes encountered during the search of the starting point and the end point;

[0199] Step (9): When there are common nodes in the search queues of the starting point and the end point, the search terminates; because this means that a path from the starting point to the end point has been found, as shown in the following formula:

[0200]

[0201] Where: Q start As a starting point, Q end for the end point;

[0202] The above algorithm selects the optimal path.

[0203] Calculate the best path, simulate and emulate drone flight, plan and assign tasks, and enable multiple drones within range to perform tasks simultaneously and efficiently

[0204] Furthermore, after completing the data integration, the system can list a variety of route planning schemes and select the best path through the optimal decision-making algorithm to ensure that the equipment that needs to be inspected is efficiently inspected.

[0205] Furthermore, environmental conditions within a certain range are collected through sensors and the data is fed into the central controller. At the same time, the flight route conditions are detected through the obstacle avoidance system and the data is fed into the central controller. At the same time, instructions from users are received, the above data are integrated, the UAV flight is simulated and emulated, and tasks are planned and assigned, so that multiple UAVs within the range can perform tasks efficiently at the same time.

Claims

1. A UAV collaborative route planning system, characterized in that: It includes central control module, power supply system, sensor system, route planning and calculation module, obstacle avoidance system, communication and data exchange module, mission planning module, simulation and emulation module and positioning system; The central control module, power supply system, sensor system, route planning and calculation module, obstacle avoidance system, communication and data exchange module, task planning module, simulation and emulation module and positioning system are connected to each other and then connected to the user interface through the communication and data exchange module.

2. The UAV collaborative route planning system according to claim 1, characterized in that: The central control module is used to achieve comprehensive scheduling; The sensing system is used for environmental perception; The obstacle avoidance system ensures flight safety; The communication and data exchange module is responsible for the efficient exchange of information; The task planning module is responsible for reasonable task allocation; The positioning system ensures the accuracy of navigation; The user interface is an interactive platform; The power supply system supports the long-term operation of the UAV. It adopts solar power generation technology to provide continuous power energy for the UAV, and maximizes the utilization efficiency of electricity through the power calculation and distribution mechanism; The simulation and emulation module is used to conduct a preliminary test, comprehensively analyze the humidity, temperature and wind speed data collected by the sensor module, calculate the real-time weather conditions of the inspected area, and transmit them to the central control module; at the same time, the module also transmits the real-time road condition information collected by the obstacle avoidance system to the central control module; The route planning calculation module is responsible for providing accurate path calculation. The optimal path calculated by the module is sent to the central control module, which then transmits the data to the simulation and emulation module for data fusion, thereby simulating a flight path.

3. The UAV collaborative route planning system according to claim 2, characterized in that: In the route planning calculation module, the system sets DFS and BFS algorithms; The system uses the DFS-BFS intelligent optimization algorithm to accurately calculate and evaluate the coordinated route planning of drones based on the distribution locations of multiple working drones and the distribution locations of drones in standby state, combined with the location information of the equipment to be inspected, the location information of the equipment being or having been inspected, the flight channel image data and its flight distance data; the specific steps of this process are as follows: Step (1): Establish a route judgment function: <h2 style=";text-align:left;direction:ltr">S = w1L<h2 style=";text-align:left;direction:ltr"> min <h2 style=";text-align:left;direction:ltr"> +w2T<h2 style=";text-align:left;direction:ltr"> min <h2 style=";text-align:left;direction:ltr"> +w3E<h2 style=";text-align:left;direction:ltr"> min <h2 style=";text-align:left;direction:ltr"> +w4O<h2 style=";text-align:left;direction:ltr"> min <h2 style=";text-align:left;direction:ltr"> +w5F<h2 style=";text-align:left;direction:ltr"> eff <h2 style=";text-align:left;direction:ltr"> +w6S<h2 style=";text-align:left;direction:ltr"> saf This formula is the total function of the optimal path; Where w1, w2, w3, w4, w5, w6 are weight coefficients, L min is the path length, T min is the flight time, E min is the energy consumption, O min is the number of obstacles, F eff is the flight efficiency, S saf is the safety score; Step (2): Assume there are n drones, and the state of the jth drone is represented by x j (j=1,2,...,n) means that if when t→∞, there is ||x j -x i ||→0, The system is said to have reached consistency; Step (3): The system evaluates the coordinated routes of the UAVs through the formula; Step (4): In the BFS algorithm, the performance of the algorithm is improved by introducing a mathematical function, as shown in the following formula: C(u)=C(Parent(u))+d(Parent(u),u) Optimal Path=argmin Path(u) C(u)for all u∈goal nodes Where C(u) is the cost to reach node u, C(Parent(u)) is the cost to reach the parent node, and d(Parent(u),u) is the distance from the parent node to the node. Optimal Path is the path with the least cost, C(u) is the cost of the path, and Optimal Path is used to describe the goal of the path-finding algorithm. Step (5): Use DFS and BFS algorithms to search from the starting point and the end point at the same time. First initialize the bidirectional search: Q start ={s},Q end ={t} In the formula, Q start As a starting point, Q end for the end point; Step (6): Use the BFS algorithm to mark the levels: L start (u)=level of node u from start in BFS tree L end (u)=level of node from end in BFS tree Where, L start (u) and L end (u) represents the level of the BFA algorithm tree of node u starting from the starting point and the end point, respectively; Step (7): Use the DFS algorithm to perform recursive exploration: Iterate from the starting point and the end point. The iteration process is as follows: you start ←Q start .dequeue() you end ←Q end .dequeue() V start ←V start ∪{u start } V end ←V end ∪{u end } for each neighbor v start of u start do Q start .enqueue(v start ) V start ←V start ∪{v start } for each neighbor v end of u end do Q end .enqueue(v end ) V end ←V end ∪{v end } C start (u)=C start (Parent start (u))+d(Parent start (u),u) C end (u)=C end (Parent end (u))+d(Parent end (u),u) In the formula, Q start is the queue used when searching from the starting point, Q end The function of is to store the nodes to be explored starting from the end point. The dequeue() operation is a standard operation in breadth-first search. start Indicates the node currently being processed in the DFS and BFS algorithms starting from the starting point, u end Indicates the node currently being processed during the search process starting from the end point, V start Usually represents the set of nodes that have been visited when searching from the starting point, V end Represents the set of nodes that have been visited when searching from the end point. enqueue() is used to add a node to the end of the queue, which means that the node is marked as pending and will be taken out for exploration at some point in the future. start (u) and C end (u) represents the cost from the starting point to the end point u and from node u to the end point respectively; Step (8): Optimal Path = arg min u∈Qmeet (C start (u)+C end (u)) Select the best path from them; In the formula, Q meet It is the set of nodes that the starting point and the end point meet during the search process; Step (9): When there are common nodes in the search queues of the starting point and the end point, the search terminates; as follows: Where: Q start As a starting point, Q end for the end point; The above algorithm selects the optimal path.

4. The UAV collaborative route planning system according to claim 3, characterized in that: In step (1), the route determination function is specifically established as follows: (1.1)、 Where L is the path length, (x i ,y i ) device location coordinates, the path length is defined as the straight-line distance from the starting point to the end point or the length of the actual flight path; (1.2)、 Flight time is defined as the minimum time required for a drone to travel from its starting point to its destination; In the formula, v i is the speed of the drone on the i-th path, (x i ,y i ) device location coordinates; (1.3)、 Energy consumption is defined as the total energy consumed by the drone during flight. Where k1, k2, k3 are coefficients related to flight distance, speed and basic energy consumption, e i represents the energy consumption of the i-th path; The number of obstacles is defined as the total number of obstacles on the path, In the formula, o i is the number of obstacles on the i-th path; (1.5)、 Flight efficiency is the ratio of path length to flight time. In the formula, the sensitivity of efficiency is adjusted by θ1, F eff It is flight efficiency; (1.6)、 The safety score is the ratio of energy consumption to the number of obstacles. In the formula, the sensitivity of security is adjusted by θ2, S saf is the safety score.

5. The UAV collaborative route planning system according to claim 3, characterized in that: The step (2) specifically includes: maintaining the consistency of multiple drones through the matrix model communication topology; the operation is as follows: Figure: G = (V, ε), Node: V = {v1,v2,v3,v4,...,v i }, Edge: ε = {(v i , v j ): i ≠ j; v i , v j ∈ V}, v i It is called v j If (v i ,v j )∈ε, Weight, in-degree, adjacency matrix, spanning tree, In order to make the states of each drone of multiple drones in the whole system tend to the same value, a control system control protocol is used; In the formula, is the latest sampling time of drone j, u i represents the control input acting on the i-th agent, k i is the control gain, a ij is an element in the adjacency matrix, indicating the degree of influence of agent j on agent i, x j (t) and x i (t) are the states of agents j and i at time t, k is the control gain, and denote the sampling time of agents j and i respectively, where and It is the state transfer matrix, which describes the change of state from the sampling time to the current time.

6. The UAV collaborative route planning system according to claim 3, characterized in that: In step (3), the specific process of the system comprehensively evaluating the coordinated routes of the drones is: First, the environmental conditions within the specified range are collected through sensors; specifically: (3.1) Use the BFS algorithm to perform an initial breadth search on the specified range to find all nearby device nodes. The BFS algorithm searches all nodes layer by layer to find the shortest path; (3.2), and then use the DFS algorithm to deeply explore each neighboring node to find the optimal path, P i and P j represents a point in two-dimensional or three-dimensional space, (x i ,y i ) device location coordinates, and L is the path length.

7. The UAV collaborative route planning system according to claim 6, characterized in that: In step (3.1), the BFS algorithm visits all adjacent nodes of the current vertex before further traversing the vertices in the graph; its operation steps are as follows: (3.1.1) First, select a vertex as the starting node and color it gray, and the rest of the nodes white; (3.1.2) Put the starting node into the queue; (3.1.3) Select a vertex from the head of the queue and find all the nodes adjacent to it. Put the found adjacent nodes at the end of the queue. Color the visited nodes black and the unvisited nodes white. If the vertex is gray, it means it has been found and put into the queue. If the vertex is white, it has not been found yet. (3.1.4) Process the next node in the queue in the same way; basically, the vertices that have left the queue become black, those in the queue are gray, and those that have not yet entered the queue are white.

8. The UAV collaborative route planning system according to claim 7, characterized in that: In step (3.1), step (3.1.4) specifically refers to: a. Initial state, starting from vertex 1, queue = {1}; b. Visit the adjacent vertex of 1, 1 is dequeued and turned black, 2 and 3 are enqueued, and the queue = {2, 3,}; c. Visit the adjacent node of 2, 2 is dequeued, 4 is enqueued, and the queue = {3, 4}; d. Visit the adjacent node of 3, 3 is dequeued, queue = {4}; e. Visit the adjacent node of 4, 4 is dequeued, queue = {empty}; If the nodes of the graph represent points in two-dimensional or three-dimensional space, the distance between nodes is calculated using the Euclidean distance; that is: L Manhattan (P i ,P j )=|x i -x j ∣+∣y i -y j | Where, L Euclidean is the Euclidean straight line distance between two points, L Manhattan is the Manhattan straight-line distance between two points, (x i ,y i ) device location coordinates.

9. The UAV collaborative route planning system according to claim 6, characterized in that: In step (3.2), after the DFS algorithm visits a vertex during the search process, it needs to recursively visit all the unvisited adjacent vertices of this vertex; Initially, all nodes are white. Select one as the starting vertex and traverse it as follows: (3.2.1), select the starting vertex and paint it gray, indicating that it has not been visited yet; (3.2.2) Select one of the adjacent vertices of the vertex, and continue this process until a vertex has no adjacent nodes. Then paint it black to indicate that it has been visited. (3.2.3) Go back to the previous layer of vertices of the blackened vertex, and then find the remaining adjacent nodes of the previous layer, and continue the above operation. If all adjacent nodes have been visited, then blacken yourself and go back to the previous layer; continue the above operation on the previous layer until all vertices have been visited; (3.2.4) Start depth search from vertex 1.

10. The UAV collaborative route planning system according to claim 9, characterized in that: The specific steps of step (3.2.3) are as follows: a. Initial state, starting from vertex 1; b. After visiting vertices 1, 2, and 3 in sequence, it ends at vertex 3; c. Backtrack from vertex 3 to vertex 2, continue to visit vertex 5, and end at vertex 5; d. Backtrack from vertex 5 to vertex 2 and terminate at the vertex; e. Backtrack from vertex 2 to vertex 1 and end at vertex 1; f. Start visiting from vertex 4 and end at vertex 4; The specific steps of step (3.2.4) are as follows: a. Initial state, starting from vertex 1; b. After visiting vertices 1, 2, and 3 in sequence, it ends at vertex 3; c. Backtrack from vertex 3 to vertex 2, continue to visit vertex 5, and end at vertex 5; d. Backtrack from vertex 5 to vertex 2 and end at vertex 2; e. Backtrack from vertex 2 to vertex 1 and end at vertex 1; f. Start visiting from vertex 4 and end at vertex 4.

Images