Unmanned aerial vehicle global route planning method based on improved quantum particle swarm optimization

Abstract

The invention discloses an unmanned aerial vehicle global route planning method based on an improved quantum particle swarm algorithm. The method comprises the following steps: constructing a grid environment model of a global route of an unmanned aerial vehicle; based on the grid environment model of the global route of the unmanned aerial vehicle, an objective function is established, and the objective function comprises a route length sub-function, a route safety sub-function and a route smoothness sub-function; improving a quantum particle swarm optimization algorithm to obtain an improved quantum particle swarm optimization algorithm; and solving the target function by using an improved quantum particle swarm optimization algorithm to obtain a global optimal route of the unmanned aerial vehicle. The technical problem of low precision of the planned global route of the unmanned aerial vehicle caused by poor search effect, low convergence precision and easy falling into a local optimal solution in a complex environment when the global route of the unmanned aerial vehicle is planned by the quantum particle swarm algorithm in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of UAV route planning, and in particular, to a global route planning method for UAVs based on an improved quantum particle swarm algorithm. Background Technique

[0002] A UAV is an unmanned aerial vehicle that can be remotely controlled over a long distance. It has obvious advantages such as small size, strong mobility, and low cost. Route planning, as a very important research direction in the UAV mission planning problem, is the key prerequisite and core guarantee for the aircraft to successfully complete the established tasks. Considering the current external factors and available resources, the system must formulate a series of task execution sequences and smooth and reliable flyable routes before the flight starts, and guide the aircraft to safely reach the task position under the condition of meeting the range and time requirements, and successfully complete the processing of the specified tasks. Different from civil airliners, the UAV system needs to find an optimal non-conflicting route from the flight starting point to the task target point according to the complex flight environment and different task requirements in the changing air battlefield. The planning result needs to meet the dynamic requirements of the UAV itself and the coordination constraints between multiple UAVs. Therefore, the route planning of UAVs is essentially a complex optimal control problem affected by multiple constraints, multiple variables, and non-linear summation, and has good application potential and research value.

[0003] There are many classification methods for current UAV route planning algorithms. According to their basic ideas, they can be divided into five categories: graph search-based methods, machine learning methods, sampling methods, mathematical model-based methods, and bio-inspired algorithms. The graph search-based method has excellent obstacle avoidance ability in two-dimensional space and high solution efficiency, but it is often difficult to be applied to multi-aircraft cooperation in multi-dimensional space; machine learning technology meets the requirements of high real-time performance, but consumes too many resources and the pressure load on hardware devices is slightly large; the sampling method utilizes probabilistic completeness and is relatively suitable for application in high-dimensional space with high search efficiency; the mathematical model method mainly relies on many very mature software, and the algorithm process is relatively complex. If the scale of the problem model increases, the calculation amount may increase exponentially; the heuristic algorithm is proposed by people analyzing the population behavior among natural species and is usually divided into neural network algorithms and evolutionary algorithms. It has been widely studied and applied in recent years. Its advantages of fast convergence speed and strong versatility are more suitable for route planning under multi-UAV cooperation.

[0004] The results of route planning often have high requirements for accuracy and real-time performance. The choice of planning method has a great impact on the planning results. There is still room for further optimization in some existing algorithms. Many research organizations at home and abroad are continuously and deeply carrying out relevant research and have proposed many new optimization algorithms. For example, an adaptive inertia weight and mutation factor are proposed in the existing literature, and the quantum particle swarm optimization algorithm is improved by using the particle convergence degree, and the convergence performance of the algorithm has been greatly improved.

[0005] Due to the rapid development of artificial intelligence and some emerging technologies, the technology of route planning involves many disciplinary fields and there are still some difficult problems to be solved. For example, the flight environment of unmanned aerial vehicles contains uncertain factors such as weather and obstacles. When planning routes, it not only needs to be restricted by its own performance, but also needs to calculate various comprehensive requirements such as external environmental constraints and task-related couplings. The scale of the planning problem gradually increases, and the complexity of the algorithm increases exponentially. These factors have significantly increased the difficulty of unmanned aerial vehicle route planning in aspects such as model construction and solution. Traditional heuristic optimization algorithms such as the quantum particle swarm optimization algorithm have the advantages of simple iterative update, few adjustable parameters, not being restricted by the problem scale, being easy to implement, and having strong global search performance, and are more suitable for solving the global route planning problem of unmanned aerial vehicles. At the same time, the quantum particle swarm optimization algorithm also has deficiencies such as poor search effect in complex environments, low convergence accuracy, and being prone to falling into local optimal solutions, and still needs to be further improved by researchers. Summary of the Invention

[0006] An embodiment of the present invention provides a method for global route planning of an unmanned aerial vehicle based on an improved quantum particle swarm optimization algorithm, so as to at least solve the technical problem that when the quantum particle swarm optimization algorithm in the prior art is used for global route planning of an unmanned aerial vehicle, the search effect in a complex environment is poor, the convergence accuracy is low, and it is easy to fall into a local optimal solution, resulting in low accuracy of the global route of the planned unmanned aerial vehicle.

[0007] According to one aspect of the embodiment of the present invention, a method for global route planning of an unmanned aerial vehicle based on an improved quantum particle swarm optimization algorithm is provided. The method may include: Step S1, constructing a grid environment model of the global route of the unmanned aerial vehicle; Step S2, establishing an objective function based on the grid environment model of the global route of the unmanned aerial vehicle, where the objective function includes: a route length sub-function, a route safety sub-function, and a route smoothness sub-function; Step S3, improving the quantum particle swarm optimization algorithm to obtain an improved quantum particle swarm optimization algorithm; Step S4, using the improved quantum particle swarm optimization algorithm to solve the objective function to obtain the global optimal route of the unmanned aerial vehicle.

[0008] Optionally, the expression of the route length sub-function is:

[0009] Wherein, is the route length sub - function. Assume that the global route of the UAV has path sub - nodes , , is the path point at the starting position, is the intermediate path point, is the path point at the target position, is the current position of the UAV, is the position of the UAV at the next path point corresponding to the current position.

[0010] Optionally, the expression of the route safety sub - function is:

[0011] where, is the route safety sub - function. When there are obstacle grids among the eight adjacent grids around the grid where the path point is located, is the number of obstacles, and , represents the penalty function of the th path point relative to the th obstacle, is the th path point relative to the th obstacle's distance, is the safety factor, indicating the number of obstacle grids among all adjacent grids around the grid where the path point is located.

[0012] Optionally, the expression of the route smoothness sub - function is:

[0013] where, is the route smoothness sub - function, is the direction angle difference between the vectors of two adjacent route segments, is the number of direction angle differences between the vectors of two adjacent route segments.

[0014] Optionally, the expression of the objective function is:

[0015] where, is the objective function, is the route length sub - function, is the route safety sub - function, is the route smoothness sub - function, is the weight of the route length sub - function, is the weight of the route safety sub - function, is the weight of the route smoothing sub-function, and .

[0016] Optionally, the specific implementation process of step S4 is as follows: Step S41, initialize the particle population size and the maximum number of iterations, and set the current iteration number; Step S42, perform chaotic initialization processing on the velocity vector and position vector of each particle in the search area, and at the same time determine the validity of each particle. If the result is invalid, each particle needs to be initialized again, where each particle is each global route of the UAV; Step S43, record the starting position vector of each particle as its individual optimal position, obtain the optimal value of the particle population position and the evolution degree of each particle using the objective function, and give each particle a weight at the average optimal position according to the natural selection method; Step S44, update the velocity vector and position vector of each particle according to the particle update formula in the QPSO algorithm, and classify the particles that exceed the search boundary after update as invalid particles; Step S45, calculate the objective function value for each particle after updating the velocity vector and position vector, compare the objective function value of each particle with the individual optimal position of each particle. If they are similar, replace the objective function value of each particle with the individual optimal position of each particle, and at the same time update the optimal value of the particle population; Step S46, calculate the particle population dispersion. If it is less than the set value, randomly perturb the particle group, recalculate the objective function value of each particle, compare the objective function value with the local extreme value, and retain the particles whose objective function value is better than the local extreme value, otherwise discard them; Step S47, sort the particle population individuals according to the objective function value of each particle, update the weight of each particle in the average optimal extreme value, and update the evolution degree, individual optimal extreme value, global optimal extreme value of the particle population, and the improved average optimal extreme value of the particle population; Step S48, determine whether the evolution degree of each updated particle or the iteration number of the evolution degree of each updated particle satisfies the condition for terminating the iteration of the algorithm. If it is satisfied, stop the iteration and output the global optimal route of the UAV, otherwise jump back to step S44 for another iterative search.

[0017] Optionally, the expression for performing chaotic initialization processing on the position vector of each particle is:

[0018] where is the th iteration process and the The state variable of the position of the particle is the state variable of the position of the particle in the -th iteration process, and is the bifurcation control parameter.

[0019] Optionally, the expression of the evolutionary degree of each particle is:

[0020] wherein, is the evolutionary degree of the -th particle in the -th iteration process, is the objective function value of the global best position in the -th iteration process, is the objective function value of the best position of the -th particle in the -th iteration process.

[0021] Optionally, the expression of the dispersion degree of the particle population is:

[0022] wherein, is the dispersion degree of the particle population in the -th iteration process, and K is the maximum number of iterations.

[0023] Optionally, the expression of the improved average optimal extreme value of the particle population is:

[0024] wherein, is the improved average optimal extreme value of the particle population in the -th iteration process, is the weight coefficient of the -th particle, is the individual optimal extreme value of the -th particle in the -th iteration process.

[0025] Advantages of the present invention: (1) The Logistic chaos initialization method is adopted to make the initial particles more evenly distributed in the search space, broaden the coverage range of the particles, improve the diversity of the particle population, and effectively avoid the problem of premature convergence; (2) Select elite individuals with better objective function values from each generation of individuals to guide other individuals to learn from themselves. By monitoring the change of the particle evolution degree parameter value during the iteration process, the particles are made to approach the best position in the population, improving the convergence performance and search accuracy of the algorithm; (3) Adjust the congestion degree of particles in the search space by dynamically adjusting the particle dispersion, making the algorithm more inclined to disperse search in the initial stage and helping the particles approach the extreme value point in the later stage, helping the particles jump out of the trap of local optimal solutions; (4) Use the natural selection method of survival of the fittest, and assign linearly decreasing weights to the particles according to the objective function values to distinguish the particles with better and worse positions in the population, improving the convergence speed of the algorithm. Description of the Drawings

[0026] The drawings described herein are used to provide a further understanding of the present invention and form a part of this application. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings: Figure 1 is a flowchart of a global route planning method for an unmanned aerial vehicle based on an improved quantum particle swarm optimization algorithm according to an embodiment of the present invention; Figure 2 is a schematic diagram of particle swarm velocity update according to an embodiment of the present invention; Figure 3 is a schematic diagram of path sub-nodes according to an embodiment of the present invention; Figure 4 is a schematic diagram of route safety according to an embodiment of the present invention; Figure 5 is a schematic diagram of route smoothness according to an embodiment of the present invention; Figure 6 is a schematic diagram of two-dimensional images of four benchmark test functions and convergence curves of different algorithms under each function according to an embodiment of the present invention; Figure 7 is a schematic diagram of path planning results and convergence curves of different algorithms according to an embodiment of the present invention. Specific Embodiments

[0027] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

[0028] It should be noted that the terms "first", "second", etc. in the description, claims and above-mentioned drawings of the present invention are used to distinguish similar objects and are used to describe a specific order or sequence. It should be understood that such terms can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0029] Embodiment 1 According to an embodiment of the present invention, a global route planning method for an unmanned aerial vehicle (UAV) based on an improved quantum particle swarm optimization algorithm is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system including at least one set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that here.

[0030] Figure 1 is a flowchart of a global route planning method for an unmanned aerial vehicle based on an improved quantum particle swarm optimization algorithm according to an embodiment of the present invention, as Figure 1 shown, the method may include the following steps: Step S1, construct a grid environment model of the UAV's global route.

[0031] In the technical solution provided in step S1 of the present invention, since each global route of the UAV is composed of multiple path sub-nodes, therefore, a grid environment model of the UAV's global route is constructed, wherein the grid environment model includes multiple path sub-nodes of each route and obstacle grids.

[0032] Step S2, based on the grid environment model of the UAV's global route, establish an objective function, wherein the objective function includes: a route length sub-function, a route safety sub-function, and a route smoothness sub-function.

[0033] In the technical solution provided in step S102 of the present invention, according to the knowledge of operations research, the route planning problem is reduced to a constrained optimization problem. The objective function is an important measure for evaluating the quality of the planning result. It directly determines the success or failure of route planning and also affects the accuracy and calculation speed of the algorithm. When facing a multi-objective optimization problem, different optimization objectives should be comprehensively weighed according to specific requirements to construct a comprehensive objective function. In a grid environment, it is usually necessary to consider (total route length, route safety, and route smoothness) as evaluation objectives. In order to solve the limitations of a single evaluation index in global route planning, the total route length, route safety, and route smoothness are combined. The weighted coefficient method is used to weigh the roles of the three in planning. By linearly weighting and combining the three sub-functions (route length sub-function, route safety sub-function, and route smoothness sub-function), an optimization problem with multiple objectives is simplified into an optimization problem with a single objective.

[0034] Step S3: Improve the quantum particle swarm optimization algorithm to obtain an improved quantum particle swarm optimization algorithm.

[0035] In the technical solution provided in step S103 of the present invention, the particle swarm optimization algorithm (Particle Swarm Optimization, abbreviated as PSO) is a heuristic search algorithm based on simulating the foraging behavior of birds in nature. Its core principle lies in the optimization of local cognition and overall cognition of gregarious animals; the particle swarm algorithm is initialized as a group of randomly generated particles, that is, a series of random solutions to the problem. Then, the optimal solution is sought through an iterative process. At the initial stage of the algorithm, the velocity and position of each particle are randomly set. As the iteration progresses, the particles will adjust their moving directions by communicating with each other, which is based on their respective positions and the position information obtained from other particles; each time an iteration is performed, the result is evaluated through an objective function value determined by a specific function, enabling the particles to search in the direction of the best-performing particle in the solution space until they finally converge at a point, which is the optimal solution to the problem sought.

[0036] Suppose in the search space of dimension, represents the number of all possible solutions to the problem, that is, the population size. The th particle is a dimensional vector , where , and its flight speed is also a dimensional vector, denoted as ; Each time the particle iterates, it is based on the individual optimal extreme value and the global optimal extreme value to adjust its own speed and position. One is the optimal position currently searched by the th particle, and the other is the optimal position found among all particles. The iterative update rule of particles in the PSO algorithm is as follows:

[0037]

[0038] where, is the speed of the th particle in the th iteration process, is the speed vector of the th particle in the th iteration process, is the individual optimal value of the th particle in the th iteration process, is the position vector of the th particle in the th iteration process, is the global optimal extreme value of the particle population in the th iteration process, and are both acceleration factors, enabling the particles to automatically adjust and learn from other particles in the group. Usually, they are set as constants; and are uniformly distributed random numbers within [0, 1], which can make the particle group show randomness and diversity; represents the current iteration number, is the upper limit of the iteration number, ; is the inertia weight factor, used to balance the global search ability and local detection ability of the particle group, and represent its maximum and minimum values respectively.

[0039] Figure 2 is the schematic diagram of particle swarm velocity update according to the embodiment of the present invention. As can be intuitively shown by Figure 2 , the composition of the particle velocity includes three aspects: the memory term represents the ability of the particle to maintain its previous velocity; the self-cognition term is based on individual experience and tends to approach its personal historical optimal position; the group-cognition term shows the mutual assistance and experience sharing among particles, helping to approach the collective historical optimal position.

[0040] Quantum Particle Swarm Optimization (abbreviated as QPSO) is a swarm intelligence algorithm improved on the basis of the PSO algorithm. In the particle swarm algorithm, the update of particles depends on their previous movement states, and the exploration range is limited and cannot cover the entire potential solution space. The quantum particle swarm algorithm utilizes the principles of quantum physics. Instead of using a definite position and velocity vector to describe the appearance location of particles, it uses a wave function to determine the specific state of particles, making the probability of a particle appearing at a specific position in the search space only related to the amplitude of the wave function. This higher uncertainty helps to overcome the limitation of local optimal solutions and accelerate the search in the global scope. The update formula of particles in the QPSO algorithm is as follows:

[0041]

[0042]

[0043]

[0044] Among them, is the position vector of the th particle in the th iteration process, is the average optimal extreme value of the population in the th iteration process, is the individual optimal value of the th particle in the th iteration process. The average optimal extreme value is introduced to represent the average value of the individual optimal positions in the population; is the contraction and expansion factor, which is used to affect the convergence speed of the algorithm; , are its maximum and minimum values; , are uniformly distributed numerical values on .

[0045] Compared with the PSO algorithm, the QPSO algorithm has the advantages of fewer parameters and strong generality. Particles without speed limitations have stronger global search capabilities. However, even if it can explore within all possible solution ranges, it still faces problems of inaccurate positioning and unsatisfactory convergence effects. The present invention adopts the following measures for improvement.

[0046] (1) Chaos initialization strategy The distributed mass of particles greatly affects the solution efficiency and solution quality of the algorithm. The more uniformly the initial particles are distributed in space, the richer the diversity maintained by the population, the greater the chance of obtaining the optimal solution, and naturally the higher the solution efficiency of the algorithm. Simply increasing the population size can improve the global search performance of the algorithm, but it cannot completely overcome the defect of the algorithm being prone to premature convergence. Therefore, an initialization method introducing a non-linear Logistic chaotic mapping is proposed based on the QPSO algorithm, and the expression is as follows:

[0047] Among them, is the state variable of the position of the -th particle in the -th iteration process, is the state variable of the position of the -th particle in the -th iteration process, , is the bifurcation control parameter,

[0048] (2)Particle evolution degree To improve the convergence performance and search accuracy of the algorithm, drawing on the idea of the genetic algorithm, elite individuals with better objective function values are selected from each generation of individuals to guide other individuals to learn from themselves. A particle evolution degree parameter is proposed to represent the self-evolution speed of a particle compared with the elite particles in the population, and the expression is as follows:

[0049] In the formula, is the evolution degree of the -th particle in the -th iteration process, and are the objective function values of the best position in the population and the best position of a single individual respectively, is the evolution degree of the particle individual during the iteration process, which gradually increases during the iteration, meaning that the particle is accelerating its evolution. When , it indicates that the optimal solution of the algorithm has been found.

[0050] Therefore, the optimal position of the particle individual can be expressed as:

[0051] (3)Particle dispersion When there are too many particles aggregated within a certain planned space range, it may cause the algorithm to fall into local extreme points, reducing the convergence accuracy and efficiency. It is necessary to perturb the overly aggregated particles at appropriate positions to enhance the algorithm's ability to jump out of local extremes. A dynamically adjusted particle dispersion parameter is proposed to monitor the particle swarm and regulate the degree of particle congestion in the space, as follows:

[0052] In the formula, is the particle swarm dispersion during the -th iteration. The log function is selected as the variation function, and

[0053] is its maximum value. At the initial stage of the algorithm, a larger particle dispersion is beneficial to improving the ability of dispersed search, but it is not conducive to global convergence. When the particles gradually aggregate in the later stage of the algorithm, the dispersion decreases, helping the particles approach the extreme points. On the other hand, to reduce the probability of the algorithm being trapped in the local optimum, when the dispersion is less than a certain threshold , the particles are randomly moved within a certain range at a set step size, and the objective function value is calculated. If it is better than the local extreme value, it is retained; otherwise, it is discarded, helping the particles jump out of the local optimal position.

[0054] (4) Natural selection method The added in the QPSO algorithm represents the average value of the individual optimal positions. However, in the population, there are naturally particles with better and worse positions. Considering each particle as equal is unreasonable to a certain extent. To further improve the convergence speed of the algorithm, a natural selection method is designed by referring to the principle of survival of the fittest in nature. In each iteration, the particles are sorted in descending order according to the objective function values of the particle swarm, and weights are assigned to each particle, and these weights linearly decrease with the change of the particles. The improved average optimal extreme value is expressed as:

[0055] In the formula, is is the improved average optimal extreme value of the particle swarm during the -th iteration, is the individual optimal extreme value of the -th particle during the -th iteration, is the weight coefficient of the -th particle, and its value linearly decreases between .

[0056] Step S4: Use the improved quantum particle swarm optimization algorithm to solve the objective function to obtain the global optimal route of the UAV.

[0057] In the technical solution provided in step S104 of the present invention, the objective function is solved by an improved quantum particle swarm optimization algorithm to obtain the global optimal route of the UAV.

[0058] The above method of this embodiment will be further introduced below.

[0059] As an alternative embodiment, in step S2, the expression of the route length sub-function is:

[0060] Wherein, is the route length sub-function. Assuming that the global route of the UAV has path sub-nodes , , is the path point at the starting position, is the intermediate path point, is the path point at the target position, is the current position of the UAV, is the position of the UAV at the next path point corresponding to the current position.

[0061] In this embodiment, Figure 3 is a schematic diagram of path sub-nodes according to an embodiment of the present invention. The flight path of the UAV is represented as a list composed of multiple route points. Each route point corresponds to a path sub-node in the route planning search map, and the node position is the center of the grid where it is located. is the path point at the starting position, is the intermediate path point, is the path point at the target position. The route is judged to be good or bad by minimizing the route length. The route length sub-function is defined as the sum of the distances on each flight segment connected by two consecutive route points, as shown in the expression of the route length sub-function.

[0062] As an alternative embodiment, in step S2, the expression of the route safety sub-function is:

[0063] Wherein, is the route safety sub-function. When there are obstacle grids in the eight adjacent grids around the grid where the path point is located, is the number of obstacles, and , represents the penalty function of the th path point relative to the th obstacle, is the th path point relative to the The distance to an obstacle is the safety factor, indicating the number of obstacle grids among all adjacent grids around the grid where the path point is located.

[0064] In this embodiment Figure 4 is a schematic diagram of airway safety according to an embodiment of the present invention. While the airway length is optimal, it is necessary to ensure the safety of the UAV during flight. The threat cost of obstacles is introduced to guide the UAV to avoid obstacles. The distance between the simulated trajectory and the obstacle is used as an index to measure the safety degree of the trajectory. The closer the distance to the obstacle, the greater the penalty for the airway, reducing the generation probability of this airway.

[0065] Considering that the shortest distance between a single path point and the surrounding obstacles is not sufficient to accurately express the complexity of the surrounding obstacles, the advantage of the grid map is selected, and the number of obstacle grids in the eight adjacent grids around the grid where the path point is located is introduced to help evaluate the safety degree of the airway. The expression of the airway safety sub-function is defined as Figure 4 shown , .

[0066] As an optional embodiment, in step S2, the expression of the airway smoothness sub-function is:

[0067] wherein is the airway smoothness sub-function is the direction angle difference between two adjacent airway segment vectors is the number of direction angle differences between two adjacent airway segment vectors.

[0068] In this embodiment Figure 5 is a schematic diagram of airway smoothness according to an embodiment of the present invention. The airway smoothness sub-function represents the turning cost of the UAV flying along this airway. To facilitate expressing the airway smoothness degree, in the grid map, the minimum turning angle of the UAV is used as the benchmark unit for measuring the smoothness of the flight path; ignoring the attitude angles of the UAV at the initial and target positions, the turning cost can be determined by calculating the sum of the absolute values of the direction angle differences between two successive flight paths.

[0069] As an optional embodiment, in step S2, the expression of the objective function is:

[0070] wherein is the objective function is the airway length sub-function is the airway safety sub-function It is a sub-function for route smoothing, is the weight of the route length sub-function, is the weight of the route safety sub-function, is the weight of the route smoothing sub-function, and .

[0071] In this embodiment, in order to solve the limitations of a single evaluation index in global route planning, the weighted coefficient method is adopted to weigh the roles of the three in replanning. By linearly weighted combining the three sub-functions (route length sub-function, route safety sub-function, and route smoothing sub-function), an optimization problem with multiple objectives is simplified into an optimization problem with a single objective. The objective function is defined as , therefore, the problem of UAV route planning can be expressed as finding , that is, for the UAV from the starting point to the ending point, the path with the minimum cost is the optimal path.

[0072] As an alternative embodiment, the specific implementation process of step S4 is as follows: Step S41, initialize the particle population size and the maximum number of iterations, and set the current iteration number; Step S42, perform chaotic initialization processing on the velocity vector and position vector of the particles in the search area, and at the same time determine the particle validity. If the result is invalid, the particle needs to be initialized again, where the particle is several global routes of the UAV; Step S43, record the starting position vector of the particle as its individual optimal position, use the objective function to obtain the optimal value of the particle population position and the particle individual evolution degree at this time, and give the weight of each particle at the average optimal position according to the natural selection method; Step S44, update the velocity vector and position vector of the particle individual according to the particle update formula in the QPSO algorithm, and classify the particles that exceed the search boundary after update as invalid particles; Step S45, calculate the objective function value for each particle individual with the updated velocity vector and position vector, compare the objective function value of each particle with the individual optimal position of each particle. If they are similar, replace the objective function value of each particle with the individual optimal position of each particle, and at the same time update the optimal value of the particle population; Step S46, calculate the particle population dispersion. If it is less than the set value, randomly perturb the particle group, recalculate the objective function value of each particle, compare the objective function value with the local extreme value, and keep the objective function value if it is better than the local extreme value, otherwise discard it; Step S47: Sort the individuals in the particle swarm according to the objective function values of each particle, update the weights of each particle in the average optimal extreme value, and update the evolution degree of each particle individual, the improved average optimal extreme value, the individual optimal extreme value, and the global optimal extreme value; Step S48: Determine whether the evolution degree of each updated particle individual or the number of iterations for calculating the evolution degree of each updated particle individual satisfies the condition for terminating the iteration of the algorithm. If it is satisfied, stop the iteration and output the global optimal route of the UAV. Otherwise, jump back to Step S44 for another iterative search.

[0073] In this embodiment, meanwhile, to obtain a realistic flyable route planning result, under the constraints of the UAV's own performance and flight space, it is necessary to constrain the effectiveness of the particles to ensure that each feasible solution meets the flight mission requirements. The basic principles are as follows: (1) Assume that the UAV still needs to return to the ground command center along the original route after performing the task. Then the total fuel carried by the UAV should ensure that it can fly to and fro. Therefore, by limiting the route length to below the set maximum distance and ensuring that the route planning does not cross the boundaries of the search area, the search scope is reduced, thus accelerating the convergence process of the algorithm; (2) Due to the physical characteristics constraints of the UAV itself, there must be a maximum limit on the turning angle during flight, and the UAV may deviate slightly from the original route when turning. It is necessary to consider the curvature of the flight route and set a maximum turning radius so that the change in the heading angle of the UAV when flying from one flight segment to the next flight segment does not exceed the maximum value; in the grid environment, the finally obtained path cannot have the situation of up-and-down or left-and-right circuitous connection, that is, the particle individual appears more than once in the same row or the same column in the grid map, ensuring the safety of the route.

[0074] Therefore, the specific process of using the improved quantum particle swarm optimization algorithm to solve the objective function and obtain the global optimal route of the UAV is from Step S41 to Step S48.

[0075] As an optional embodiment, in Step S42, the expression for performing chaotic initialization processing on the position vector of the particle is:

[0076] where is the state variable of the position of the -th iteration process of the -th particle, is the state variable of the position of the -th iteration process of the -th particle, , is the bifurcation control parameter, .

[0077] As an alternative embodiment, in step S43, the expression for the individual particle evolution degree is:

[0078] where is the evolution degree of the -th particle in the -th iteration process, is the objective function value of the global best position in the -th iteration process, is the objective function value of the best position of the -th particle in the -th iteration process.

[0079] As an alternative embodiment, in step S43, the expression for the particle population dispersion degree is:

[0080] where is the particle population dispersion degree in the -th iteration process, and K is the maximum number of iterations.

[0081] As an alternative embodiment, in step S43, the expression for the improved average optimal extreme value is:

[0082] where is the improved average optimal extreme value of the particle population in the -th iteration process, is the weight coefficient of the -th particle, is the individual optimal extreme value of the -th particle in the -th iteration process.

[0083] Experimental part: To verify the effectiveness of the algorithm proposed in the present invention, the convergence performance of the algorithm is experimentally tested on the CEC2005 test set. The two-dimensional images of the four selected benchmark test functions and the convergence curves of different algorithms under each function are shown. Fig. 6 is a schematic diagram of the two-dimensional images of the four benchmark test functions according to the embodiments of the present invention and the convergence curves of different algorithms under each function. Fig. 6(a) is a schematic diagram of the two-dimensional image of the first benchmark test function and the convergence curves of different algorithms under each function; Fig. 6(b) is a schematic diagram of the two-dimensional image of the second benchmark test function and the convergence curves of different algorithms under each function; Fig. 6(c) is a schematic diagram of the two-dimensional image of the third benchmark test function and the convergence curves of different algorithms under each function; Fig. 6(d) is a schematic diagram of the two-dimensional image of the fourth benchmark test function and the convergence curves of different algorithms under each function. As shown in Figs. 6(a), (b), (c) and (d), the standard PSO algorithm is added to the optimization convergence curve graph for comparison; the experimental results can show that the convergence rate of the improved algorithm is not low, and better optimization results can also be obtained, and the defect that the standard PSO algorithm is prone to fall into local extrema is solved to a certain extent.

[0084] The scenario of global path planning is a 20×20 grid map, and there are several static obstacles with different shapes manually set in the map. The task requirement is to find an optimal flight path from the starting point to the ending point for the unmanned aerial vehicle without colliding with the obstacles. The planning results of the algorithm proposed in the present invention and the standard PSO algorithm are shown simultaneously, and the convergence curve of the algorithm is drawn. Fig. 7 is a schematic diagram of the path planning results and the convergence curve of different algorithms according to the embodiments of the present invention. Fig. 7(a) is a schematic diagram of the path planning results of different algorithms according to the embodiments of the present invention, and Fig. 7(b) is a schematic diagram of the convergence curves of different algorithms according to the embodiments of the present invention. As shown in Figs. 7(a) and (b), the experimental results show that after improving the QPSO algorithm according to the proposed method, the optimization accuracy and speed of the algorithm are significantly improved, and it has good stability, effectively improving the deficiency that the standard PSO algorithm is often trapped in local optima.

[0085] In the embodiments of the present invention, a grid environment model of the global route of the unmanned aerial vehicle (UAV) is constructed; based on the grid environment model of the global route of the UAV, an objective function is established, where the objective function includes: a route length sub-function, a route safety sub-function, and a route smoothness sub-function; the quantum particle swarm optimization algorithm is improved to obtain an improved quantum particle swarm optimization algorithm; the improved quantum particle swarm optimization algorithm is used to solve the objective function to obtain the global optimal route of the UAV, solving the technical problem in the prior art that when the quantum particle swarm algorithm is used for global route planning of the UAV, the search effect is poor, the convergence accuracy is low, and it is easy to fall into a local optimal solution in a complex environment, resulting in low accuracy of the planned global route of the UAV. The technical effect is achieved that when performing global route planning for the UAV, a chaotic initialization strategy, particle evolution degree, particle dispersion degree, and the selection method of natural survival of the fittest are proposed to obtain an improved quantum particle swarm optimization algorithm. Through the route length sub-function, route safety sub-function, and route smoothness sub-function, an objective function is obtained. During the process of using the improved quantum particle swarm optimization algorithm to solve the objective function, the improved quantum particle swarm optimization algorithm has a good search effect, high convergence accuracy, and is difficult to fall into a local optimal solution in a complex environment, making the accuracy of the planned global route of the UAV high.

[0086] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments.

[0087] In the above embodiments of the present invention, the descriptions of the respective embodiments have their own emphases. For parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0088] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of units can be a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of units or modules can be in an electrical or other form.

[0089] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0090] In addition, each functional unit in various embodiments of the present invention may be integrated into a first processing unit, may exist separately as individual physical units, or two or more units may be integrated into one unit. The above integrated units may be implemented in the form of hardware or in the form of software functional units.

[0091] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A global path planning method for UAV based on improved quantum particle swarm algorithm, characterized in that: include: Step S1, constructing a grid environment model of the global route of the UAV; Step S2, establishing an objective function based on the grid environment model of the global route of the UAV, wherein the objective function includes: a route length sub-function, a route safety sub-function and a route smoothness sub-function; Step S3, improving the quantum particle swarm optimization algorithm to obtain an improved quantum particle swarm optimization algorithm; Step S4, using the improved quantum particle swarm optimization algorithm to solve the objective function and obtain the global optimal route of the UAV.

2. The method according to claim 1, characterized in that The expression of the route length sub-function is: in, is the path length sub-function, assuming that the global path of the UAV has Path child nodes , , is the waypoint of the starting position, is the middle path point, is the waypoint to the target location, is the current position of the drone, The next path point corresponding to the current position is the location of the drone.

3. The method according to claim 2, characterized in that The expression of the route safety subfunction is: in, is the route safety sub-function. When there are obstacle grids in the adjacent grids in eight directions around the grid where the path point is located, is the number of obstacles, and , Indicates The path point is relative to the The penalty function for obstacles is For the The path point is relative to the The distance of obstacles, is the safety factor, indicating the path point The number of barrier grids among all neighboring grids around the grid.

4. The method according to claim 3, characterized in that The expression of the route smoothness subfunction is: in, is the route smoothness function, is the direction angle difference between two adjacent route segment vectors, is the number of direction angle differences between two adjacent route segment vectors.

5. The method according to claim 4, characterized in that The expression of the objective function is: in, is the objective function, is the route length sub-function, is the route safety sub-function, is the route smoothness function, is the weight of the route length sub-function, is the weight of the route safety sub-function, is the weight of the route smoothness subfunction, and .

6. The method according to claim 4, characterized in that The specific implementation process of step S4 is as follows: Step S41, initializing the particle population size and the maximum number of iterations, and setting the current number of iterations; Step S42, performing chaos initialization processing on the velocity vector and position vector of each particle in the search area, and determining the validity of each particle. If the result is invalid, each particle needs to be initialized again, wherein each particle is each global route of the UAV; Step S43, recording the starting position vector of each particle as its individual optimal position, using the objective function to obtain the optimal value of the particle population position and the evolution degree of each particle at this time, and giving each particle a weight in the average optimal position according to the natural selection method; Step S44, updating the velocity vector and position vector of each particle according to the particle update formula in the QPSO algorithm, and classifying the particles beyond the search boundary after the update as invalid particles; Step S45, calculating the objective function value for each particle after the velocity vector and position vector are updated, comparing the objective function value of each particle with the individual optimal position of each particle, and if they are similar, replacing the objective function value of each particle with the individual optimal position of each particle, and updating the optimal value of the particle population at the same time; Step S46, calculate the particle population dispersion, if it is less than the set value, randomly perturb the particle population, recalculate the objective function value of each particle, compare the objective function value with the local extreme value, retain the particle whose objective function value is better than the local extreme value, otherwise discard it; Step S47, sorting the individuals of the particle population according to the objective function value of each particle, updating the weight of each particle in the average optimal extreme value, and updating the evolution degree of each particle, the individual optimal extreme value, the global optimal extreme value of the particle population and the average optimal extreme value of the particle population after improvement; Step S48, determine whether the evolution degree of each particle after update or the number of iterations of the evolution degree of each particle after update meets the conditions for terminating the algorithm iteration. If so, stop the iteration and output the global optimal route of the drone. Otherwise, jump back to step S44 and iterate the search again.

7. The method according to claim 6, characterized in that The expression for implementing the chaos initialization process on the position vector of each particle is: in, For the In the iteration process The state variable of the particle's position, For the In the iteration process The state variable of the particle's position, , is the fork control parameter, .

8. The method according to claim 6, characterized in that The expression of the evolution degree of each particle is: in, For the In the iteration process The evolution degree of a particle, For the The objective function value of the optimal position of the group during the iteration, For the In the iteration process The objective function value of the optimal position of each particle.

9. The method according to claim 8, characterized in that The particle population discreteness expression is: in, For the The particle population dispersion in the iteration process, K is the maximum number of iterations.

10. The method according to claim 9, characterized in that The expression of the improved average optimal extreme value of the particle population is: in, For the The average optimal extreme value of the particle population after improvement in the iteration process, For the The weight coefficient of each particle, For the In the iteration process The individual optimal extreme value of each particle.

Images