Genetic algorithm and sliding window optimization-based air-ground cooperative route planning method, storage medium and equipment

Abstract

The invention discloses an air-ground cooperative flight path planning method based on a genetic algorithm and sliding window optimization, a storage medium and equipment. The method is oriented to a cooperative track planning scene of the mooring unmanned aerial vehicle and the unmanned vehicle, and aims at constraints such as space coupling of the unmanned aerial vehicle and the unmanned vehicle and ground safety collision avoidance of the unmanned vehicle caused by mismatching of navigational speeds of the unmanned aerial vehicle and the unmanned vehicle and limited length and retracting speed of a mooring cable, and aims at targets such as shortest track, shortest task time and optimal track. Based on a genetic algorithm, track planning tasks with multiple constraints and multiple optimization targets under an air-ground collaborative operation system are realized, a real-time track is adjusted and adaptively optimized through a sliding window optimization method, and the problem of collaborative track planning conflicts of a tethered unmanned aerial vehicle and an unmanned vehicle during outdoor inspection of a power system is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to a kind of air-ground collaborative path planning method based on genetic algorithm and sliding window optimization, storage medium and equipment, for the field of tethered unmanned aerial vehicle and unmanned vehicle cooperative operation of outdoor power system inspection, especially in the process of air-ground collaborative path planning. BACKGROUND

[0002] As a new type of unmanned aerial vehicle technology, tethered unmanned aerial vehicle is connected with ground power supply and data transmission system through tethered cable, realizing long-time air stay and operation. Compared with traditional independent flight unmanned aerial vehicle, tethered unmanned aerial vehicle not only has longer flight duration, but also can provide stronger reliability and data transmission capability through connection with ground system. Typical tethered unmanned aerial vehicle system mainly includes unmanned aerial vehicle platform, tethered cable and ground integrated platform, wherein the ground integrated platform includes power supply system and automatic cable winding and unwinding device, which provides power supply and communication link for tethered unmanned aerial vehicle, and ensures that the cable is orderly deployed and recovered when the unmanned aerial vehicle is lifted and lowered, avoiding entanglement.

[0003] In the field of power inspection, tethered unmanned aerial vehicle can hover in high-altitude position such as power line and substation for a long time, monitor equipment operation status in real time, and quickly transmit high-definition images to help staff identify potential fault points. It is especially suitable for power inspection scenes that require uninterrupted lighting, monitoring or data collection, including night maintenance lighting, high-altitude operation safety monitoring, fine autonomous inspection and emergency and special inspection.

[0004] Unmanned vehicle power inspection is an important and mature part of intelligent power grid and digital transformation, which refers to the use of ground mobile robots (usually called unmanned vehicles or ground robots) to perform equipment inspection, data collection and status monitoring of power facilities autonomously or remotely. For outdoor power inspection scenes, unmanned vehicle can form perfect complement with tethered unmanned aerial vehicle inspection, and unmanned vehicle can serve as the ground integrated platform of tethered unmanned aerial vehicle. Both unmanned vehicle and tethered unmanned aerial vehicle have autonomous operation capability, and they can make collaborative decision and planning around task scenes, realize data sharing and collaborative operation, and build an integrated intelligent inspection system of "air + ground". SUMMARY

[0005] The present application provides a kind of air-ground collaborative path planning method based on genetic algorithm and sliding window optimization, computer equipment and storage medium to solve the technical problems of the prior art.

[0006] In order to achieve the above purpose, the present application provides the following technical solutions:

[0007] Firstly, a power inspection air-ground collaborative path planning method based on genetic algorithm and sliding window optimization is provided, which comprises:

[0008] Step 1) constructing a multi-constraint multi-objective optimization model for the cooperative path planning of the tethered UAV and the unmanned vehicle, defining gene coding variables, the gene coding variables including three-dimensional path point coordinates and speed values of the tethered UAV, two-dimensional plane trajectory coordinates of the unmanned vehicle, and cable length information of the tethering winch;

[0009] Step 2) setting constraint conditions of the multi-constraint multi-objective optimization model, including tethered UAV trajectory constraints, speed constraints, unmanned vehicle collision constraints, speed constraints, motion rotation angle constraints, and winch cable length constraints and cable length retraction speed constraints;

[0010] Step 3) setting multi-objective optimization functions of the multi-constraint multi-objective optimization model, including the shortest time for coupling path chains, the minimum energy consumption of the tethered UAV and the unmanned vehicle, the optimal trajectory smoothness, the minimum collision risk of the unmanned vehicle navigation, the minimum number of winch cable length retraction and extension, and the minimum average cable length, using a weight coefficient variation method to normalize the multi-objective, and constructing an individual fitness function;

[0011] Step 4) solving the multi-constraint multi-objective optimization model using an improved genetic algorithm, including generating an initial population based on the constraint conditions, calculating individual fitness, using a tournament selection method for selection operation, using an adaptive multi-point crossover method for crossover operation, using a Gaussian mutation method based on the constraint conditions for mutation operation, and performing sliding window optimization on individual gene coding after each generation evolution, and finally outputting a cooperative path that meets all constraint conditions and has the optimal fitness.

[0012] Optionally, in step 1), the gene coding variables The tethered UAV path coding , the unmanned vehicle path coding and the winch cable length coding are packaged into a unified node, and the position of the tethered UAV on the predetermined path line is taken as a common reference.

[0013] Optionally, in step 2), the unmanned vehicle collision constraint uses a rasterized electronic map, combines obstacles and safety radii into forbidden grid, and instantly counts the number of forbidden grids crossed by the unmanned vehicle trajectory in the gene decoding stage.

[0014] Optionally, in step 4), the adaptive multi-point crossover method selects a maximum number of crossover points ≥8 in the early evolution stage to increase diversity, and decreases to ≤4 in the later stage according to a linear decreasing strategy.

[0015] Optionally, in step 4), the Gaussian mutation performs mutation operation in a hierarchical progressive manner within , , .

[0016] Optionally, step 5) is further included, in actual execution, the environment and the trajectory deviation are monitored in real time, the local trajectory is re-planned and dynamically adjusted by using the sliding window optimization, and the real-time feasibility, safety and optimality of the tethered unmanned aerial vehicle and the unmanned vehicle collaborative trajectory are ensured.

[0017] Optionally, the multi-objective optimization function in step 3) adopts a dynamic weight strategy after normalization: the degree of optimization target deviation is calculated according to a preset fixed weight vector in the offline evolution stage; and the weight is adjusted online according to the real-time deviation and the remaining cable length margin in the online sliding window stage, so that the energy consumption and safety weight are increased with the increase of the real-time deviation.

[0018] Optionally, the sliding window optimization in step 5) is specifically: the actual deviation and the environmental change are monitored, the unexecuted trajectory segment is rolled and cut off according to the node number, the local path of the unmanned vehicle is calculated by using the A star algorithm, and then the winch speed and the cable length are verified whether they are out of limit; if they are out of limit, the unmanned aerial vehicle speed is reduced to be re-planned, otherwise the modified segment is written back to the gene chain and the control instruction is output after the deviation correction amount is added, then the node number is incremented, and the cycle is executed until the task is completed.

[0019] Secondly, a computer readable storage medium is further provided, the storage medium stores a computer program, and the computer program is executed by a processor to realize the above method.

[0020] Finally, a computer device is further provided, which includes a memory, a processor and a computer program stored on the memory and executable on the processor, and the processor realizes the above method when executing the program.

[0021] The beneficial effects of the present application are as follows: the present application is directed to the tethered unmanned aerial vehicle and the unmanned vehicle collaborative trajectory planning scene, the unmanned aerial vehicle and the unmanned vehicle space coupling caused by the mismatch of the unmanned aerial vehicle and the unmanned vehicle speed, the tethered cable length and the winch speed, the unmanned vehicle ground safety collision avoidance constraint, and the like, around the shortest trajectory, the shortest time, the optimal trajectory and the like, the trajectory planning task of multiple constraints and multiple optimization targets under the air-ground collaborative operation system is realized based on the genetic algorithm, and the real-time trajectory is adjusted and adaptively optimized by the sliding window optimization method, and the problem of the tethered unmanned aerial vehicle and the unmanned vehicle collaborative trajectory planning in the outdoor inspection of the power system is solved. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 It is a schematic diagram of the tethered unmanned aerial vehicle and the unmanned vehicle collaborative operation;

[0023] Figure 2 It is a flow chart of the multi-constraint multi-objective genetic algorithm;

[0024] Figure 3 It is a real-time sliding window optimization flow chart. DETAILED DESCRIPTION

[0025] In order for those skilled in the art to better understand the technical solutions of the present application and to implement them, the present application will be described in detail below in conjunction with the drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, and cannot be used to limit the protection scope of the present application.

[0026] As shown in Figure 1 , the tethered unmanned aerial vehicle (T-UAV) A001 needs to sail according to the planned trajectory, and is connected with the unmanned ground vehicle (UGV) A002 through the tether cable A003 to obtain power supply and communication data. The unmanned ground vehicle A002 sails following the tethered unmanned aerial vehicle A001, and needs to avoid ground obstacles to ensure its own safety. Since the position between the unmanned ground vehicle A002 and the tethered unmanned aerial vehicle A001 is constrained by the cable length, the unmanned ground vehicle A002 needs to search for the optimal sailing path in a limited space. In addition, the unmanned ground vehicle A002 can also obtain a larger movement space by adjusting the cable length between the unmanned ground vehicle A002 and the tethered unmanned aerial vehicle A001. Generally, the winch cable length needs to be maintained in the optimal range according to the spatial relationship between the unmanned ground vehicle A002 and the tethered unmanned aerial vehicle A001. A cable length that is too short will limit the maneuverability of the tethered unmanned aerial vehicle A001, and a cable length that is too long will increase the sailing resistance of the tethered unmanned aerial vehicle A001 and the risk of entanglement. The radius R1 represents the circular range radius within which the unmanned ground vehicle A002 can move under the current cable length constraint.

[0027] The global cooperative path planning of the tethered unmanned aerial vehicle and the unmanned ground vehicle in this embodiment can be regarded as a multi-constraint multi-objective optimization problem, which is defined as , indicating two highly coupled sailing trajectory chains of the tethered unmanned aerial vehicle and the unmanned ground vehicle. The path points on the two sailing trajectory chains are coupled with a certain point on the opposite trajectory chain. The distance between the coupled path points anchors the cable length information of the tether winch at the same node. In order to accurately represent the coupled trajectory chain, the node coding variable is defined as follows:

[0028]

[0029] Let represent the coordinate value and the sailing speed value of the path point numbered in the three-dimensional trajectory of the tethered unmanned aerial vehicle, represent the coordinate value of the two-dimensional ground plane trajectory of the unmanned ground vehicle, and couple, that is, the tethered unmanned aerial vehicle and the unmanned ground vehicle must move to this position at the same time, represent the cable length of the tether winch, and , couple. Then the node coding variable can be further described as:

[0030]

[0031] The gene of an individual encodes all the key variables within the whole system directly or indirectly, which is the only instruction to guide the individual behavior, The gene of an individual encodes all the key variables within the whole system directly or indirectly, which is the only instruction to guide the individual behavior, In this paper, Instead of representing the position of the tethered UAV in the spatial coordinate system, it represents the position of the tethered UAV on the predetermined trajectory line, And represents the lateral and longitudinal offsets of the tethered UAV at the th node relative to the predetermined trajectory line, and after using the above gene encoding form, only the value range of and can meet the trajectory constraint of the tethered UAV , and the continuous value provides a reference line for the subsequent population individual crossover process, which is conducive to gene splicing and optimization. The gene chain of an individual with gene encoding nodes can be represented as follows:

[0032]

[0033] To solve the th node, the segment of is represented by , and the subsequent is referred to as .

[0034] Around the three constraint objects of the tethered UAV, unmanned vehicle and winch, the constraint conditions of the optimization problem can be extracted as follows:

[0035] (1) Tethered UAV trajectory constraint : The tethered UAV must sail along the predetermined trajectory, with a certain offset error in the lateral and longitudinal directions, and the trajectory encoding of the tethered UAV , represents the position of the tethered UAV on the predetermined trajectory line, and represents the lateral and longitudinal offsets of the tethered UAV at the th node relative to the predetermined trajectory line, and by limiting the value range of and , the trajectory constraint of the tethered UAV , is used to store the value range of and in the unmanned vehicle trajectory, and requires ​.

[0036] (2) Speed ​​constraints of tethered UAVs The maximum flight speed and maximum acceleration / deceleration capability of the tethered drone must be selected within a limited set range. express The flight speed of the tethered drone at that location, express The acceleration of tethered drones, Indicates the maximum permissible flight speed of the tethered drone. This indicates the maximum permissible acceleration and speed constraint for tethered drones. Used to save the current solution middle and Quantity required .

[0037] (3) Collision constraints for unmanned vehicles The unmanned vehicle's trajectory should not collide with any obstacles and must remain outside the safe distance of all ground obstacles. The boundary of the no-entry zone for ground obstacles includes the obstacle itself and the safe radius. This embodiment uses a grid method to process electronic charts and obstacle targets. Used to save solutions Unmanned vehicle flight path The number of obstacle grids traversed is required. .

[0038] (4) Unmanned vehicle speed constraints The maximum speed and maximum acceleration / deceleration capability of the unmanned vehicle must be selected within a limited set range; define and constrain the speed of the tethered drone. Similarly, requirements .

[0039] (5) Unmanned vehicle rotation angle constraint When searching for a path, autonomous vehicles must consider the maximum turning angle constraint, making sharp turns impossible. The selection of the next point's location must be based on the previous point's location and heading, overlaid with the search viewpoint, within the reachable range. express and The turning angles of the two segments of the unmanned vehicle's flight path. This indicates the maximum permissible threshold for the turning angle of the autonomous vehicle's trajectory. Used for saving The number of nodes and the trajectory turning angle constraints. .

[0040] (6) Winch cable length constraint The winch cable length must accommodate the distance between the tethered drone and the unmanned vehicle, and there is a maximum length limit; representing the node winch cable length at time t, representing the maximum winch cable length available, the minimum winch cable length required for the tethered UAV to maintain stable flight is determined according to the three-dimensional spatial relationship between the tethered UAV and the unmanned vehicle using a parabolic fitting method , to store the number of and , requiring .

[0041] (7) winch cable speed constraint : there is a maximum winch cable speed limit when the winch adjusts the cable length, let represent the average winch cable speed in the segment , let represent the maximum winch movement speed, to store the number of , requiring .

[0042] To achieve better tethered UAV and unmanned vehicle air-ground collaborative path planning, the optimization problem is set up with the following eight optimization objectives: including the shortest time of the coupled path chain , the minimum energy consumption of the tethered UAV , the smoothness of the tethered UAV trajectory , the minimum collision risk objective of the unmanned vehicle , the minimum energy consumption of the unmanned vehicle , the smoothness of the unmanned vehicle trajectory , the minimum number of winch cable length extension and retraction , the minimum average winch cable length , after normalization, the eight optimization objectives are transformed into final objective values using the weight coefficient method , which is mathematically described as follows:

[0043]

[0044] In the above formula, are the weight values of the eight optimization objectives, respectively, indicating the degree of bias towards this optimization objective.

[0045] Therefore, the air-ground global path planning problem of the tethered UAV and the unmanned vehicle can be described as: outputting a coupled path chain that satisfies the constraint condition , and the optimization total objective value is the minimum, that is, the flight time of the double-chain coupled path is as short as possible, the path of the tethered UAV and the unmanned vehicle is as smooth as possible, the energy consumption is as low as possible, the winch action is as few as possible and the cable length is maintained at a reasonable length. The mathematical quantitative representation of the eight optimization objectives is as follows:

[0046] (1) Target with the shortest coupling track chain time : Representing fragments The flight time for an internally tethered drone is as follows:

[0047]

[0048] (2) Tethered drones have the lowest energy consumption In the objective optimization function In this study, the energy consumption of tethered drones is one of the eight optimization objectives included in the horizontal comparison. The calculation is performed using weighted coefficients, and it is not necessary to determine the specific energy consumption value. Therefore, some parameters can be simplified. The energy consumption required for the tethered drone to maintain its balance can be ignored, and only the energy consumption due to the drone's speed can be considered. The energy consumption objective function can then be approximately expressed as:

[0049]

[0050] The drag of a tethered drone can generally be expressed as:

[0051]

[0052] in Indicates air density; Indicates the speed of the tethered drone; This indicates the drag coefficient of a tethered unmanned aerial vehicle (UAV). Representing the frontal cross-sectional area (the projected area perpendicular to the direction of motion), we can obtain:

[0053]

[0054] (3) Smoothness of tethered UAV trajectory All trajectory turns within the tethered drone's trajectory segment The mean is defined as follows:

[0055]

[0056] (4) Minimize the risk of collision during unmanned vehicle navigation Definition method and collision constraints of autonomous vehicles Consistency is a double guarantee for ensuring the safety of unmanned vehicle navigation trajectories.

[0057]

[0058] (5) Minimize energy consumption of unmanned vehicles Definition and Minimum Energy Consumption Target for Tethered Unmanned Aerial Vehicles Similarly, as follows:

[0059]

[0060] (6) Unmanned vehicle trajectory smoothness : Definition and tethered unmanned aerial vehicle trajectory smoothness Similarly, the definition is as follows:

[0061]

[0062] (7) Winch cable length retraction frequency minimum target : This optimization target can avoid frequent adjustment of cable length; represents a segment The average speed of the winch retraction cable in the segment, is used to store The number of times the winch retraction cable is retracted in the segment, is defined as follows:

[0063]

[0064] (8) Winch average cable length minimum target : The winch cable length should be used as needed, and the excessive cable length will increase the tethered unmanned aerial vehicle navigation resistance and the risk of winding, represents the node The winch cable length at the node, is defined as follows:

[0065]

[0066] In summary, the calculation method of each optimization target is as follows:

[0067]

[0068] Therefore, the tethered unmanned aerial vehicle and unmanned vehicle air-ground collaborative trajectory planning problem can be expressed as:

[0069]

[0070] For the optimization problem The corresponding related constraints and optimization targets, this embodiment is based on genetic algorithm, through improved population initialization strategy and heuristic evolution method, and in the actual execution stage of coupling trajectory, the local trajectory is optimized by sliding window method, so as to realize the power inspection air-ground collaborative trajectory planning method based on genetic algorithm and sliding window optimization. Experiments prove that this algorithm can meet the global collaborative trajectory planning requirements of tethered unmanned aerial vehicle and unmanned vehicle, and has the advantages of multi-objective adaptive optimization, fast convergence speed, higher trajectory search success rate and the like.

[0071] (1) Gene coding design:

[0072]

[0073] (2) Individual fitness design:

[0074] Individual fitness function selects the sum of optimization target weights , The smaller it represents the higher the environmental fitness of the individual.

[0075] (3) Multi-constrained multi-objective genetic algorithm for power inspection air-ground collaborative path planning, see Figure 2 :

[0076] Step 001: initialization of power inspection air-ground collaborative system parameters;

[0077] Step 002: based on constraint conditions , generate an initial population;

[0078] Step 003: based on function, individual fitness calculation in population;

[0079] Step 004: selection: select excellent individuals according to individual fitness;

[0080] Step 005: crossover: excellent individuals in the population are crossed to generate offspring;

[0081] Step 006: mutation: mutate the gene coding in the offspring individual with a certain probability;

[0082] Step 007: individual optimization: optimize each offspring gene based on the sliding window optimization operator;

[0083] Step 008: generate a new generation population based on individual screening strategy;

[0084] Step 009: check if the number of evolutionary iterations reaches the set value, if not, return to step 003 and continue the new round of evolution process; if yes, output the optimal individual according to the current state and end the algorithm process.

[0085] In a preferred example: the difficulty of population initialization lies in complex and diverse constraint conditions and optimization targets, and the embodiment gradually completes the population initialization process based on the creative gene coding method combined with the iterative progressive search method. All individuals in the initial population meet all constraint conditions while having high fitness.

[0086] The population initialization process is as follows:

[0087] Step 201: initialize system parameters, tethered unmanned aerial vehicle search starting point coordinates and node serial number, , target population individual number ;

[0088] Step 202: ++, random sampling The probability of winning a lottery value is determined based on a normal distribution, and a random selection algorithm is used to select the winning value.

[0089] Step 203: Select the horizontal movement based on a greedy algorithm and longitudinal movement Priority is given to the location of the endpoint of the trajectory of unmanned vehicles and tethered drones, and the probability of winning is allocated accordingly. Then, the winning value is determined according to the random selection algorithm.

[0090] Step 204: Determine the motion time from the average velocity between the two nodes of the tethered drone. ;

[0091] Step 205: Randomly sample the winch speed based on a normal distribution , ;

[0092] Step 206: Fit the parabola according to the constraint relationship Calculate the maximum reachable radius of the autonomous vehicle (lock up );

[0093] Step 207: Determine the maximum achievable speed of the autonomous vehicle based on its maximum acceleration and deceleration capabilities. and minimum achievable speed Calculate the average velocity and the radius of motion (circles) respectively. and circle );

[0094] Step 208: Select Overlapping areas, obstacle removal areas and their safety radii, generate a set of areas selectable by the unmanned vehicle;

[0095] Step 209: Based on a greedy algorithm, prioritize selecting coordinates close to the tethered drone. If the coordinates are unreachable, then use... The algorithm is connected, and the number of coordinate points of the tethered UAV is expanded according to the number of reversal nodes to obtain... and ;

[0096] Step 210: Iterate through and check whether each individual meets the mandatory constraints; discard those that do not.

[0097] Step 211: (The population has not reached the specified number of individuals) Return to step 202;

[0098] Step 212: Population initialization complete.

[0099] Furthermore, in step 203, the longitudinal and lateral displacements are selected based on a greedy algorithm. and vertical lateral movement , the logic is as follows: define the candidate position The distance from the unmanned vehicle to the tethered unmanned aerial vehicle is The distance from the tethered unmanned aerial vehicle to the end point is Then the intermediate quantity is designed as follows:

[0100]

[0101] The percentage on the predetermined trajectory is represented by , and the candidate position is defined as follows:

[0102]

[0103] Further, in step 206, in order to calculate the maximum movement radius of the unmanned vehicle under the current cable length constraint, a parabolic fitting method is used to approximate the cable length attitude at the maximum movement radius, assuming that the height of the tethered unmanned aerial vehicle relative to the unmanned vehicle is , and the horizontal distance is , then the minimum winch cable length satisfies the following formula:

[0104]

[0105] Based on the above formula, the maximum movement radius of the unmanned vehicle under the current cable length can be calculated in reverse.

[0106] In a preferred example, the preferred example adds an individual optimization link after the selection, crossover, and mutation links of the conventional genetic algorithm, adopts a heuristic evolution method, artificially optimizes the gene coding in part of the individuals, improves the fitness of the individuals, and speeds up the convergence process by guiding the evolution of the population individuals.

[0107] Further:

[0108] (1) The selection link adopts tournament selection method (Tournament Selection), and the selection basis is the fitness of each individual The main idea is to select the best individuals in a batch of individuals through multiple comparisons, as the basis for the next round of evolution. The advantage of tournament selection method is that when dealing with high-dimensional problems, it is more stable and less likely to fall into local optimal solution, and can effectively avoid the premature problem in the selection process, improve the global search ability of the algorithm, and increase the diversity of the population.

[0109] (2) The crossover link selects the adaptive multi-point crossover method, and sets the crossover point number (generally set to ), and the population evolution in the early stage selects the crossover point number , for improving population diversity, and the number of crossovers is selected in the later stage to accelerate convergence of population individuals.

[0110] (3) The mutation link selects a Gaussian mutation method based on constraint conditions. Because constraint conditions are diverse and complex, blind mutation can cause new individuals to not meet the constraint conditions. Therefore, the mutation link must consider the constraint condition restrictions, otherwise it is invalid mutation. To maintain the stability of population individuals, the individual screening rate is set to 0.1, and the selected individuals are randomly selected for mutation. According to the gene mutation rate , it is determined whether each gene encodes mutation or no mutation. The maximum range of gene mutation is limited, generally taking of the mutated value to avoid too large mutation values from random walking. The gene encoding is internally coupled between variables, so it is divided into , , three parts according to the subsystem object, and a hierarchical progressive mutation operation is performed. In addition, to ensure that the mutation operation is controllable under the constraint condition, the gene encoding is mutated, and then the mutation range that meets the constraint condition is calculated according to the constraint condition combined with the mutation value constraint coefficient . Then, a Gaussian mutation is used to select the mutation value. After mutation is completed, a constraint condition check is performed again, and only the mutated individuals that meet the constraint condition are retained.

[0111] (4) The individual optimization link takes local optimization as the goal, relies on an optimization operator to slide a window to traverse the gene sequence, and optimizes the local gene encoding. This can accelerate population evolution and individual convergence. The optimization process is similar to the mutation link, and is divided into , , three parts for batch optimization, including the mooring unmanned aerial vehicle trajectory optimization link, the mooring unmanned aerial vehicle speed optimization link, and the winch cable length optimization link. After local optimization, it is checked whether it meets the constraint condition. If not, it is discarded.

[0112] In a preferred example, the optimization operator is designed as follows:

[0113] (1) Mooring unmanned aerial vehicle trajectory optimization operator:

[0114] Mooring unmanned aerial vehicle trajectory optimization can improve the smoothness of the mooring unmanned aerial vehicle navigation trajectory, make the wild values deviating from the trajectory based on a sliding window operator for adaptive approaching processing, and design a matrix operator , the processing object is , and the mooring unmanned aerial vehicle trajectory optimization operator is defined as follows:

[0115]

[0116]

[0117]

[0118] (2) Tethered UAV speed optimization operator:

[0119] The tethered UAV speed optimization operator is matrix , and the processing object is , and the operation mode is similar to the tethered UAV trajectory optimization operator .

[0120]

[0121]

[0122] (3) Winch cable length optimization operator:

[0123] The winch cable length optimization operator is designed in the same way as the tethered UAV speed optimization operator , and is expressed as follows:

[0124]

[0125]

[0126] The individuals before optimization and the individuals after optimization are both retained in the current population, and the individuals with high fitness are selected by tournament selection to form the next generation population. Due to the characteristics of tournament selection, the two types of individuals theoretically have the possibility to enter the next generation population. This retention mechanism can accelerate the convergence of population individuals to a certain extent, but also retains a certain diversity evolution possibility.

[0127] The foregoing describes the multi-constraint multi-objective genetic algorithm for power inspection air-ground collaborative path planning in detail. The optimal collaborative operation path under the condition of multi-constraint multi-objective can be searched. However, in the actual power inspection air-ground collaborative operation process, the environment will fluctuate and change, and the tethered UAV and the unmanned vehicle cannot accurately execute the predetermined trajectory. The trajectory deviation will be accumulated gradually. Therefore, the following real-time sliding window planning module is designed, as shown in Figure 3 , to re-plan the current local path of the tethered UAV and the unmanned vehicle to ensure that the constraint conditions and the safe collision avoidance requirements are met.

[0128] In a preferred example, the real-time sliding window planning module performs the following steps:

[0129] Step 301: Obtain the gene code of the optimal individual of the multi-constraint multi-objective genetic algorithm;

[0130] Step 302: Obtain the tethered unmanned aerial vehicle, unmanned ship, and winch cable length gene code based on the GCN (gene code number, initially 0);

[0131] Step 303: The trajectory safety monitoring module monitors the environmental changes and the subsequent to-be-executed trajectory in real time, judges whether local correction is needed, if not, directly jumps to step 310 for execution, and if needed, executes step 304;

[0132] Step 304: Intercept the local trajectory segment that needs to be re-planned;

[0133] Step 305: Calculate the time of the local segment ;

[0134] Step 306: Use the A-star algorithm to re-plan the local trajectory of the unmanned ship;

[0135] Step 307: Winch reachability evaluation, check whether the requirements for winch speed and cable length in the re-planned trajectory exceed the maximum speed and maximum cable length of the winch, if yes, it is unreachable, execute step 308, if not, it is reachable, jump to step 309;

[0136] Step 308: Based on the speed adjustment module, reduce the sailing speed of the tethered unmanned aerial vehicle and start local planning again, jump to step 305;

[0137] Step 309: Update the new gene code after local adjustment to the optimal individual to generate a new current optimal individual.

[0138] Step 310: The gene code sequence value extractor obtains the tethered unmanned aerial vehicle, unmanned ship, and winch cable length gene code based on the GCN (gene code number) and delivers it to the deviation correction module;

[0139] Step 311: The offset correction module executes the deviation according to the current positions and speeds of the tethered unmanned aerial vehicle, unmanned vehicle, and winch, calculates the correction amount, and superimposes the gene code obtained in step 310 to generate the final gene code control amount that needs to be executed;

[0140] Step 312: The tethered unmanned aerial vehicle trajectory tracking module, unmanned vehicle trajectory tracking module, and winch cable length following control module execute the gene code control amount;

[0141] Step 313: GCN (gene code number) ++, jump to step 302 for cyclic execution.

[0142] The embodiment of the application also provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the steps of the power inspection air-ground cooperative trajectory planning method.

[0143] The embodiments of the present application also provide a computer device. At the hardware level, the computer device comprises a processor, an internal bus, a network interface, a memory and a nonvolatile memory, and can also comprise other hardware required by business. The processor reads the corresponding computer program from the nonvolatile memory into the memory and then runs, so as to implement the steps of the power inspection air-ground cooperative path planning method.

[0144] Those skilled in the art should understand that the embodiments of the present application can be provided as a method or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes.

[0145] The present application is described with reference to flowcharts and / or block diagrams of the method and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams can be realized by computer program instructions, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing devices to produce a machine, so that the instructions executed by the computer or other programmable data processing devices produce a device implemented in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the function specified by one or more flows and / or blocks.

[0146] These computer program instructions can also be stored in a computer readable memory capable of guiding the computer or other programmable data processing devices to work in a specific way, so that the instructions stored in the computer readable memory produce a manufactured product comprising instruction means, which implements the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the function specified by one or more flows and / or blocks.

[0147] These computer program instructions can also be loaded into the computer or other programmable data processing devices, so that a series of operation steps are performed on the computer or other programmable devices to produce a computer implemented process, so that the instructions executed on the computer or other programmable devices provide steps for implementing the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the function specified by one or more flows and / or blocks.

[0148] It should be noted that the above detailed description of the present application is merely intended to provide a more complete understanding of the present application to those skilled in the art, and is not intended to limit the present application in any way. Therefore, although the present application has been described in detail, those skilled in the art should understand that modifications and equivalent replacements can still be made to the present application; and all technical solutions and improvements that do not depart from the spirit and scope of the present application are encompassed in the protection scope of the patent of the present application. Any reference signs in the claims should not be regarded as limiting the claims involved.

Claims

1. A method for air-ground cooperative trajectory planning based on genetic algorithm and sliding window optimization, characterized in that: Includes the following steps: Step 1) Construct a multi-constraint, multi-objective optimization model for collaborative trajectory planning of tethered UAVs and unmanned vehicles, and define gene-encoded variables, which include the coordinates and speed of the three-dimensional trajectory points of the tethered UAV, the two-dimensional plane trajectory coordinates of the unmanned vehicle, and the cable length information of the tethered cable winch. Step 2) Set the constraints of the multi-constraint multi-objective optimization model, including the trajectory constraints and speed constraints of the tethered UAV, the collision constraints, speed constraints, and motion rotation angle constraints of the unmanned vehicle, as well as the winch cable length constraints and cable deployment / retraction speed constraints. Step 3) Set the multi-objective optimization function of the multi-constraint multi-objective optimization model, including the shortest time for the coupled track chain, the minimum energy consumption of the tethered UAV and the unmanned vehicle, the optimal trajectory smoothness, the minimum collision risk of the unmanned vehicle, the minimum number of winch cable releases and take-ups, and the minimum average cable length. The weight coefficient change method is used to normalize the multi-objectives and construct the individual fitness function. Step 4) Solve the multi-constraint multi-objective optimization model using an improved genetic algorithm, including: generating an initial population based on constraints, calculating individual fitness, performing selection operations using tournament selection, performing crossover operations using an adaptive multi-point crossover method, performing mutation operations using a Gaussian mutation method based on constraints, and performing sliding window optimization on individual gene encoding after each generation of evolution, finally outputting a cooperative trajectory that satisfies all constraints and has the best fitness.

2. The method according to claim 1, characterized in that: Gene-encoded variables in step 1) A layered coupling structure is used to encode the tethered UAV track. Unmanned vehicle track coding Coding with winch cable length Encapsulated as a unified node, and the position of the tethered drone on the predetermined trackline. As a public benchmark.

3. The method according to claim 1, characterized in that: In step 2), the collision constraints of the unmanned vehicle adopt a gridded electronic map, which merges obstacles and safety radii into no-entry grids, and counts the number of no-entry grids crossed by the unmanned vehicle trajectory in real time during the gene decoding stage.

4. The method according to claim 1, characterized in that: The adaptive multi-point crossover method described in step 4) selects the maximum number of crossover points in the early stages of evolution. To increase diversity, a value of ≥8 is used, and then the value is reduced to a linear decreasing value in the later stages. ≤4.

5. The method according to claim 2, characterized in that: The Gaussian mutation described in step 4) is in , , The mutation operation is performed in a hierarchical and progressive manner.

6. The method according to any one of claims 1 to 5, characterized in that: It also includes step 5), which involves real-time monitoring of environmental and trajectory deviations during actual execution, and using sliding window optimization to replan and dynamically adjust local trajectories to ensure the real-time feasibility, safety, and optimality of the collaborative flight path between tethered UAVs and unmanned vehicles.

7. The method according to claim 6, characterized in that: Step 3) The multi-objective optimization function adopts a dynamic weighting strategy after normalization: in the offline evolution stage, the degree of emphasis on the optimization objective is calculated according to a preset fixed weight vector; During the online sliding window phase, the weights are adjusted online based on the real-time deviation and the remaining cable length margin, so that the weights of energy consumption and safety increase as the real-time deviation increases.

8. The method according to claim 6, characterized in that: The sliding window optimization described in step 5) specifically involves: monitoring actual deviations and environmental changes, extracting unexecuted track segments by node number, first calculating the local path of the unmanned vehicle using the A* algorithm, and then verifying whether the winch speed and cable length exceed the limits; if they exceed the limits, the speed of the unmanned vehicle is reduced and replanned; otherwise, the corrected segment is written back to the gene chain and the deviation correction amount is superimposed before outputting control commands. Subsequently, the node number is incremented, and the process is repeated until the task is completed.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the method described in any one of claims 1 to 8.

10. A computer device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in any one of claims 1 to 8.

Images