CN116203989A - Multi-unmanned aerial vehicle cooperative target searching method and system based on particle swarm optimization - Google Patents
Multi-unmanned aerial vehicle cooperative target searching method and system based on particle swarm optimization Download PDFInfo
- Publication number
- CN116203989A CN116203989A CN202310279162.2A CN202310279162A CN116203989A CN 116203989 A CN116203989 A CN 116203989A CN 202310279162 A CN202310279162 A CN 202310279162A CN 116203989 A CN116203989 A CN 116203989A
- Authority
- CN
- China
- Prior art keywords
- unmanned aerial
- aerial vehicle
- particle swarm
- obstacle
- swarm algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
- G05D1/106—Change initiated in response to external conditions, e.g. avoidance of elevated terrain or of no-fly zones
Abstract
The invention belongs to the technical field of group intelligence and multi-agent target search, and provides a multi-unmanned aerial vehicle collaborative target search method and system based on a particle swarm algorithm. According to the invention, the position of each unmanned aerial vehicle at the next moment is calculated through a particle swarm algorithm, the surrounding environment is detected in real time in the process that the unmanned aerial vehicle actually goes to the target point at the next moment, and re-planning is performed in time when an obstacle is detected so as to realize the functions of avoiding the obstacle and avoiding the collision between the unmanned aerial vehicles. Introducing an energy factor attenuated along with time into a particle swarm algorithm to simulate the law of the decline of searching capacity along with time, introducing a maximum speed constraint, and controlling the speed of the unmanned aerial vehicle in the calculated direction to be not more than the maximum speed constraint after the magnitude and the direction of the speed at the next moment are calculated.
Description
Technical Field
The invention belongs to the technical field of group intelligence and multi-agent target search, and particularly relates to a multi-unmanned aerial vehicle collaborative target search method and system based on a particle swarm algorithm.
Background
In recent years, considering that the particle swarm algorithm (Particle Swarm Optimization, PSO) has the advantages of simple parameters, high convergence speed, low computational complexity, small data transmission capacity and the like, the particle swarm algorithm is widely applied to multi-unmanned aerial vehicle target search; specifically, each unmanned aerial vehicle is abstracted into particles in a particle group, and the positions of the particles are encoded, so that the unmanned aerial vehicle can gradually move to the global optimum, and then the target position is found.
The inventor discovers that the current multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm algorithm does not completely have unmanned aerial vehicle characteristics, and most of researches are based on ideal condition assumptions. The method comprises the following steps: unmanned aerial vehicles are regarded as particles, no obstacle exists in the environment, and collision among the unmanned aerial vehicles is not considered; in a few algorithms considering obstacle avoidance, collision is avoided only by changing the speed direction, emergency braking or changing the flying height when the existence of an obstacle is detected, but in practical application, the obstacle in the environment is ubiquitous, and the obstacle avoidance under the environment with multiple obstacles is difficult to realize by changing the speed direction, emergency braking or changing the flying height and the like, so that the problem of collision among machines cannot be solved, and the searching success rate is seriously influenced. In addition, the speed of the unmanned aerial vehicle cannot tend to be infinite in consideration of the actual characteristics of the unmanned aerial vehicle, and the drastic speed change does not accord with the dynamics rule of the unmanned aerial vehicle; in practical application, the severe speed change causes more energy loss and simultaneously causes accident occurrence phenomenon.
Disclosure of Invention
In order to solve the problems, the invention provides a multi-unmanned aerial vehicle cooperative target searching method and system based on a particle swarm algorithm, which takes the actual characteristics of unmanned aerial vehicles into consideration, introduces an autonomous obstacle avoidance algorithm to optimize the actual running track of the unmanned aerial vehicle, and solves the obstacle avoidance problem and the problem of collision among the unmanned aerial vehicles in a multi-obstacle environment; in addition, the energy limitation of the unmanned aerial vehicle is considered, and an objective rule that the searching capability of the unmanned aerial vehicle gradually decreases along with the time is simulated by adding an energy factor attenuated along with the time into a particle swarm algorithm, so that a reliable basis is provided for the research of the energy loss process.
In order to achieve the above object, the present invention is realized by the following technical scheme:
in a first aspect, the present invention provides a method for searching a collaborative target of multiple unmanned aerial vehicles based on a particle swarm algorithm, including:
acquiring position information of a target to be searched;
controlling a plurality of unmanned aerial vehicles to move towards the position of the target to be searched for according to a preset particle swarm algorithm;
the method comprises the steps of introducing an energy factor which decays along with the time to simulate the law of the decline of searching capacity along with the time into a particle swarm algorithm, introducing a maximum speed constraint, and controlling the speed of the unmanned aerial vehicle in the calculated direction to be not more than the maximum speed constraint after the size and the direction of the speed at the next moment are calculated.
Further, calculating the target position from the current moment to the next moment of each unmanned aerial vehicle through a particle swarm algorithm, and after determining the guiding path from the current moment position to the next moment target position, generating an unmanned aerial vehicle control instruction pushing away the obstacle at all the positions with the obstacle on the guiding path; and when the obstacle on the guiding path is determined, other unmanned aerial vehicles which are in a preset range around the current controlled unmanned aerial vehicle and have higher priority than the current controlled unmanned aerial vehicle are regarded as the obstacle.
Further, inertial weight parameters, individual acceleration factor parameters and social acceleration factor parameters are added in the particle swarm algorithm, and the weights of the parameters are dynamically adjusted at each control stage.
Further, in the beginning stage of the search task, the inertia weight value is the largest; as the search task is executed, the inertia weight is reduced, and the social acceleration factor is increased; after the social acceleration factor reaches a preset value, the individual acceleration factor gradually rises.
Further, the energy factor at the next moment is equal to the product of the energy factor at the current moment and the energy attenuation rate, and the energy attenuation rate simulates the energy attenuation speed in the unmanned aerial vehicle flight process.
Further, each unmanned aerial vehicle is respectively set as a single particle in a particle swarm algorithm; the actual position where the drone is located is set as the particle position.
Further, the control process from the current moment to the next moment is divided into a search phase and an optimization phase; in the searching stage, obtaining a guide path without considering obstacles through an A-algorithm; in the optimizing stage, the surrounding environment of the unmanned aerial vehicle is detected, the track existing in the obstacle is compared with the collision-free guiding path generated in the searching stage, and the gradient information is utilized to pull out the track in the obstacle from the obstacle.
In a second aspect, the present invention further provides a multi-unmanned aerial vehicle collaborative target search system based on a particle swarm algorithm, including:
a data acquisition module configured to: acquiring position information of a target to be searched;
a search control module configured to: controlling a plurality of unmanned aerial vehicles to move towards the position of the target to be searched for according to a preset particle swarm algorithm;
the method comprises the steps of introducing an energy factor which decays along with the time to simulate the law of the decline of searching capacity along with the time into a particle swarm algorithm, introducing a maximum speed constraint, and controlling the speed of the unmanned aerial vehicle in the calculated direction to be not more than the maximum speed constraint after the size and the direction of the speed at the next moment are calculated.
In a third aspect, the present invention also provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm algorithm of the first aspect.
In a fourth aspect, the present invention further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor implements the steps of the multi-unmanned plane collaborative target searching method based on the particle swarm algorithm according to the first aspect when executing the program.
Compared with the prior art, the invention has the beneficial effects that:
1. according to the method, an energy factor which decays along with the time is introduced into a particle swarm algorithm to simulate the law of the decline of searching capacity along with the time, and a maximum speed constraint is introduced, and after the size and the direction of the speed at the next moment are calculated, the speed of the unmanned aerial vehicle in the calculated direction is controlled to not exceed the maximum speed constraint; on the basis of adhering to the objective rule that the energy level of the unmanned aerial vehicle is reduced and the searching capability is weakened along with the time in the task execution process, the speed can reach the maximum, the safety problem caused by exceeding the maximum speed constraint is avoided, and the problems of multiple energy loss and accident occurrence caused by severe speed change are solved;
2. according to the invention, the target position of each unmanned aerial vehicle from the current moment to the next moment is calculated through a particle swarm algorithm, after the guiding path from the current moment position to the next moment target position is determined, an unmanned aerial vehicle control instruction for pushing away the obstacle is generated at all the positions with the obstacle on the guiding path, and obstacle avoidance under a multi-obstacle environment can be realized through the unmanned aerial vehicle control instruction for pushing away the obstacle; in addition, other unmanned aerial vehicles which are in a preset range around the current controlled unmanned aerial vehicle and have higher priority than the current controlled unmanned aerial vehicle are regarded as obstacles, the unmanned aerial vehicle with higher priority is preferentially ensured to fly according to the guiding path, the unmanned aerial vehicle with lower control priority executes the obstacle avoidance task on the guiding path, the flight stability of the unmanned aerial vehicle with higher priority in the multiple unmanned aerial vehicles is ensured, the control stability and the searching capability of the multiple unmanned aerial vehicles are ensured, and the problem of collision among the unmanned aerial vehicles is solved;
3. according to the invention, a self-adaptive strategy is introduced into a particle swarm algorithm, and the problems of local optimum control are avoided and the searching capability is improved by dynamically adjusting the weights of all parameters of inertia weight, individual acceleration factors and social acceleration factors in all stages; specifically, in the beginning stage of the search task, inertia weight takes the dominant role to obtain stronger global searching capability; as the search task is executed, the inertia weight starts to decrease, the social acceleration factor starts to increase, and the attraction of the global optimum to the particles starts to increase, so that the particles can get rid of the current local optimum; after the social acceleration factor reaches a preset value, the individual acceleration factor starts to gradually rise, and the attraction of the individual optimal to the particles rises, so that the unmanned aerial vehicle can realize aggregation near the target position in the later period.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments and are incorporated in and constitute a part of this specification, illustrate and explain the embodiments and together with the description serve to explain the embodiments.
FIG. 1 is a flow chart of embodiment 1 of the present invention;
FIG. 2 is a schematic diagram of the basic particle swarm algorithm according to embodiment 2 of the present invention;
FIG. 3 is a simulated implementation scenario diagram of embodiment 2 of the present invention;
FIG. 4 is a schematic diagram of the control after 1s in example 2 of the present invention;
FIG. 5 is a simulation diagram of the control 30s according to embodiment 2 of the present invention;
FIG. 6 is a simulation diagram of the control 60s according to embodiment 2 of the present invention;
FIG. 7 is a simulation diagram of the control 90s according to embodiment 2 of the present invention;
fig. 8 is a schematic diagram of collision avoidance according to embodiment 2 of the present invention.
Detailed Description
The invention will be further described with reference to the drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present application. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
Example 1:
when the unmanned aerial vehicle is controlled to search a cooperative target based on a particle algorithm, collision is avoided only by changing the speed direction, emergency braking or changing the flying height under the condition that the obstacle exists, however, in practical application, the obstacle in the environment is ubiquitous, obstacle avoidance under the environment with multiple obstacles is difficult to realize by changing the speed direction, emergency braking or changing the flying height and other modes, and in the traditional control method, the problem of collision among unmanned aerial vehicles is not considered, the problem of collision among the unmanned aerial vehicles in the flying process cannot be solved, and the searching success rate is seriously influenced.
Aiming at the obstacle avoidance problem and the inter-machine conflict problem in the multi-obstacle environment existing in the control process of the multi-unmanned aerial vehicle; the embodiment provides a multi-unmanned aerial vehicle collaborative target searching method based on a particle swarm algorithm, which comprises the following steps:
acquiring position information of a target to be searched; it may be understood that the target to be searched may be the same target to be searched by a plurality of aircrafts, or may be different targets to be searched determined for different one or more aircrafts respectively;
controlling a plurality of unmanned aerial vehicles to move towards the position of the target to be searched for according to a preset particle swarm algorithm;
calculating the target position from the current moment to the next moment of each unmanned aerial vehicle through a particle swarm algorithm, and generating an unmanned aerial vehicle control instruction pushing away the obstacle at all the positions with the obstacle on the guide path after determining the guide path from the current moment position to the next moment target position; and when the obstacle on the guiding path is determined, other unmanned aerial vehicles which are in a preset range around the current controlled unmanned aerial vehicle and have higher priority than the current controlled unmanned aerial vehicle are regarded as the obstacle.
Specifically, the target position from the current moment to the next moment of each unmanned aerial vehicle is calculated through a particle swarm algorithm, after a guiding path from the current moment position to the next moment target position is determined, unmanned aerial vehicle control instructions for pushing away the obstacle are generated at all the positions with the obstacle on the guiding path, and obstacle avoidance under a multi-obstacle environment can be realized through the unmanned aerial vehicle control instructions for pushing away the obstacle; and in the preset range around the current unmanned aerial vehicle to be controlled, other unmanned aerial vehicles with higher priority than the current unmanned aerial vehicle to be controlled are regarded as the obstacle, the unmanned aerial vehicle with higher priority is preferentially guaranteed to fly according to the guiding path, the unmanned aerial vehicle with lower control priority executes the obstacle avoidance task on the guiding path, the flight stability of the unmanned aerial vehicle with higher priority in the multiple unmanned aerial vehicles is guaranteed, the control stability and the searching capability of the multiple unmanned aerial vehicles are guaranteed, and the problem of collision among the unmanned aerial vehicles is solved.
In order to explain the multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm algorithm in the embodiment, the steps of the method may include:
s1, modeling a search environment and a plurality of controlled unmanned aerial vehicles;
s2, initializing parameters in a particle swarm algorithm, and initializing the positions and speeds of all unmanned aerial vehicles;
s3, when each controlled unmanned aerial vehicle flies to the target position at the next moment at the current moment, an obstacle avoidance algorithm is applied to avoid obstacles and avoid the controlled unmanned aerial vehicle from colliding with other unmanned aerial vehicles;
s4, the unmanned aerial vehicle reaches a target position at the current stage at a certain moment and calculates the position fitness; and carrying out group information sharing among unmanned aerial vehicles and calculating the target position at the next moment.
In step S1, each unmanned aerial vehicle may be set as a single particle in the particle swarm algorithm, respectively; the actual position where the drone is located is set as the particle position.
In step S3, introducing a collision avoidance mechanism between the autonomous obstacle avoidance and the unmanned aerial vehicle into the execution process of the particle swarm algorithm, and when the unmanned aerial vehicle calculates that the unmanned aerial vehicle goes to the target position x (t+1) and goes to the target position x at the next moment through the particle swarm algorithm, changing a control mode that the unmanned aerial vehicle goes to along a straight line when controlled by the conventional particle swarm algorithm; the method in the embodiment includes triggering a path planning mechanism, wherein the whole planning process from the current moment to the next moment is divided into a front-end searching process and a rear-end optimizing process, and in the front-end searching stage, each unmanned aerial vehicle searches a target position from the current moment to the next moment through dynamics A respectively without considering a guiding path of an obstacle; and in the rear-end track optimization link, checking whether the track generated in the previous stage collides with the obstacle, and generating a force for pushing the track away from the obstacle based on the gradient to be used as an unmanned plane control instruction for pushing away from the obstacle. Furthermore, if the generated trajectory violates the kinetic constraint due to unreasonable time allocation, performing a time reassignment procedure; in order to save limited on-board resources, a back horizon strategy is introduced into the planning process, so that path planning is only performed within the perception range of the unmanned aerial vehicle, and the environment beyond the detection range of the unmanned aerial vehicle sensor is not considered. In addition, the embodiment also considers the collision avoidance among machines, and introduces a priority strategy into the unmanned aerial vehicle. Priority it is understood that when multiple drones are planning to pass through a confined space at the same time, the order priority of each drone to execute the planning is determined. Unmanned aerial vehicles with higher priority can be planned and pass through preferentially, and unmanned aerial vehicles with low priority treat high priority unmanned aerial vehicles in a certain sensing range as barriers and implement obstacle avoidance.
In step S4, inertial weights are introduced into the particle swarm algorithm to eliminate the maximum velocity V for each dimension of the particle max Is not limited to the above-mentioned requirements. The velocity and position formulas of the particles are expressed as:
υ id ＝ω*υ id +c 1 *rand()*(p best -x id )+c 2 *rand()*(g best -x id )
x id ＝x id +υ id
wherein ω represents an inertial weight; p is p best Representing the optimal position that each particle has found at present; g best Representing the optimal position found so far for the whole population of particles; c 1 And c 2 Referred to as individual acceleration factor and social acceleration factor, respectively, representing the pulling of particles toward p best Or g best Weights of random acceleration terms of (a); upsilon (v) id Representing the speed of the ith particle at the d-th iteration; x is x id Indicating where the i-th particle is located at the d-th iteration.
The self-adaptive strategy is introduced into the particle swarm algorithm in real time, and inertia weight omega and individual acceleration factor c are dynamically adjusted in each stage 1 And social acceleration factor c 2 The weights of the parameters are used for avoiding sinking into local optimum, so that the searching capability is improved. At least one searching mode in the present embodiment is for a single target search, at which time, therefore, locally optimal trapping should be avoided; specifically, when the search task is just started, the inertial weight is given a maximum value and should take a dominant role to obtain a stronger global search capability; as the search task is executed, the inertia weight begins to decrease, the social acceleration factor begins to increase, and the attraction of the global optimum to the particles begins to increase, thereby causing the particles toThe current local optimum can be eliminated; after the social acceleration factor reaches the preset optimal value, the individual acceleration factor starts to gradually rise, and the attraction of the individual optimal to the particles rises, so that the unmanned aerial vehicle can realize aggregation near the target position in the later period of the experiment.
In this embodiment, energy and speed constraints may be added to the particle swarm algorithm; in terms of energy, considering that in practical application, the battery endurance of a currently commonly used unmanned aerial vehicle is extremely limited, namely, the energy level inevitably drops along with the time in the actual flight process, energy constraint is introduced into a traditional particle swarm algorithm, the rule of the drop of the searching capability along with the time is simulated by introducing an energy factor which decays along with the time, and optionally, the energy factor at the next moment is equal to the product of the energy factor at the current moment and the energy decay rate which simulates the energy decay speed in the unmanned aerial vehicle flight process; in terms of speed, considering that the speed of the unmanned aerial vehicle in actual flight has an upper limit and cannot be arbitrarily trended to infinity, a constraint of maximum speed is introduced for this phenomenon, and when the magnitude of the speed vector at the next moment calculated by the particle swarm algorithm exceeds the maximum safe speed that can be reached by the unmanned aerial vehicle, the direction of the vector is maintained, but the magnitude is limited within the upper limit.
Specifically, introducing an energy factor attenuated along with time into a particle swarm algorithm to simulate the law of the decline of searching capacity along with time, introducing a maximum speed constraint, and after calculating the size and direction of the speed at the next moment, controlling the speed of the unmanned aerial vehicle in the calculated direction not to exceed the maximum speed constraint; on the basis of observing the objective rule that the energy level of the unmanned aerial vehicle is reduced and the searching capability is weakened along with the time in the task execution process, the speed can reach the maximum, the safety problem caused by exceeding the maximum speed constraint is avoided, and the problems of multiple energy loss and accident occurrence caused by severe speed change are solved.
In the embodiment of the invention, the particle swarm algorithm is applied to the multi-unmanned aerial vehicle target search task, each unmanned aerial vehicle in the unmanned aerial vehicle cluster is regarded as a single particle in the particle swarm, and the positions of the particles are encoded, so that the particles can gradually move towards the global optimum along with the progress of iteration, and the target positions can be effectively found. The self-adaptive strategy is introduced into the particle swarm algorithm, and the weight of each parameter is dynamically adjusted in the whole searching process, so that the unmanned aerial vehicle has stronger global searching capability at the initial stage of executing the searching task, can effectively avoid sinking into a local minimum value in the searching process, and can realize rapid convergence to a target position at the final searching stage. Considering the characteristics of the unmanned aerial vehicle in actual flight, such as energy limitation of the unmanned aerial vehicle in the flight process, an energy factor attenuated along with time is introduced into a traditional particle swarm algorithm, and the objective rule of energy level decline and search capability weakening of the unmanned aerial vehicle along with time in the task execution process is respected. In addition, a maximum speed limit is added to the unmanned aerial vehicle, so that the speed of the unmanned aerial vehicle is limited within a safe speed, and the unmanned aerial vehicle does not tend to be infinite. Considering the possibility of collision between environmental obstacles and machines in the actual task execution process, optimizing the actual running track of the unmanned aerial vehicle by introducing an autonomous obstacle avoidance algorithm into a traditional particle swarm algorithm, and introducing a priority strategy to enable the unmanned aerial vehicle with high priority to be regarded as an obstacle by low priority and to implement obstacle avoidance so as to realize the collision avoidance function under the complex environment of the obstacle; the practical application value of the multi-unmanned aerial vehicle collaborative particle swarm search algorithm is remarkably improved.
Example 2:
in order to further explain and verify the multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm optimization in embodiment 1, the embodiment is explained by a specific simulation example, specifically:
as shown in fig. 3, alternatively, the search task is performed in a rectangular space of 12m×13 m. In order to ensure the universality of the proposed method, the number, the size and the position of the barriers which are relatively random can be distributed in the environment at the beginning of each search task, and the barriers can be circular, cylindrical or other shaped barriers; in one embodiment, the positioning of the circular obstacle can allow the unmanned aerial vehicle to pass through or fly over, the cylindrical obstacle can allow the unmanned aerial vehicle to bypass from the left side and the right side, and the target position is randomly selected in the search space and cannot collide with the obstacle.
The unmanned aerial vehicle can select arbitrary quantity a plurality of, and in one embodiment selects 4 unmanned aerial vehicles, and specifically optional distributes 4 unmanned aerial vehicles that the model is all the same, the configuration side by side at the scene edge of waiting to search, and unmanned aerial vehicle interval 1 meter each other, and every unmanned aerial vehicle all possesses hardware modules such as binocular camera, depth camera, GPS and communication module, and each module keeps the performance good in the flight. In the whole searching process, 4 unmanned aerial vehicles can communicate through wireless communication and other modes, and information exchange is achieved. The specific implementation steps are as follows:
s1, modeling an environment to be searched, and optionally, for a given rectangular search area, dividing the rectangular search area based on square grids, wherein the actual area of each grid is set to be 1m multiplied by 1m. The targets are distributed in the environment to be searched according to probability. Modeling unmanned aerial vehicles, optionally regarding a group of unmanned aerial vehicles as a group of particles, each unmanned aerial vehicle being modeled as a single particle in the group of particles, the actual location of the unmanned aerial vehicle being taken as the particle location. And regarding the sensor return value on the unmanned aerial vehicle as an environment fitness value f. The specific formula is as follows:
area to be searched:
S：{x：0≤x≤x max }，{y：0≤y≤y max }
let d be the drone sensor return value:
f (x, y) =d search targets are found (x, y) ∈s such that
S2, initializing particle swarm parameters, and optionally: particle number n=4, maximum velocity v max =2m/s, the highest iteration number is set to 10000 times, the termination condition is set to reach the highest iteration number or 20 iterations without generating new global optima, c 1 ＝c 2 ＝2，R 1 And R is 2 Set to be between 0 and 1The energy attenuation rate lambda is set to 0.95 by the random number of the rate value. And respectively initializing the speed and the position of the 4 unmanned aerial vehicles.
In the specific implementation process, the initialization parameter assignment used for the particle swarm at the beginning can be based on reference documents, experimental parameters or experience parameters, but the set initialization parameters often cannot achieve the optimal effect. In the debugging process, a certain specific parameter can be properly adjusted while other parameter values are kept unchanged, meanwhile, whether the searching effect is improved or not is observed, if the effect is improved, the modification is kept, and each parameter is adjusted in the mode until the overall searching effect is optimal.
S3, the 4 unmanned aerial vehicles respectively go to the corresponding positions; the whole process can be subdivided into two parts, front-end search and back-end optimization. In the front-end search stage, a guide path which does not consider obstacles is obtained through an A-algorithm. In the back-end optimization stage, the unmanned aerial vehicle detects the surrounding environment by using a sensor, compares the track existing in the obstacle with the collision-free guide path generated in the search stage, and pulls out the track in the obstacle from the obstacle by using gradient information so as to generate and optimize a plurality of tracks of a plurality of threads, and meanwhile, executes the track with the lowest cost. The problem of non-linear optimization of a uniform B-spline with Q control points is given by:
wherein J is a penalty function of the required optimization, and the penalty J is weighted r Composition, wherein r= { s, c, d, t }, represents smoothness, collision, dynamic feasibility and terminal progress, respectively; lambda represents the corresponding weight J r The penalty function, which is the required optimization, consists of penalties. Since the particle swarm search is a multi-machine search, collision avoidance between unmanned aerial vehicles is also required to be considered, in order to avoid collision between unmanned aerial vehicles, an unmanned aerial vehicle priority mechanism can be introduced, and unmanned aerial vehicles with low priorities need to consider trajectories with high priorities as obstacles. In order to save resources, only unmanned aerial vehicles within a certain range are consideredThe impact of a high priority drone is not considered if its distance from it is greater than a threshold. Thus, the above optimization problem becomes J':
wherein lambda is w A conflict weight corresponding to the weighted conflict penalty term; j (J) ω Lambda is the weighted conflict penalty term ω Is the corresponding penalty weight.
When the dynamics of a certain track is not feasible, the time allocated to the track is increased through time redistribution so as to increase the dynamics feasibility of the track. In addition, in order to save limited on-board resources of the unmanned aerial vehicle, a back horizon policy is introduced into the planning process, so that path planning is only performed within the perception range of the unmanned aerial vehicle.
S4, the unmanned aerial vehicle completes obstacle avoidance and reaches the position of the corresponding particle designated previously, the environment is detected by using a sensor carried by the unmanned aerial vehicle, whether a target exists in the current area is judged, and the target is taken as the adaptability of the particle at the position. The broadcasting network is assumed to exist between unmanned aerial vehicles, so that group information sharing is performed between unmanned aerial vehicles, and the speed and the position of particles corresponding to each unmanned aerial vehicle are updated through a particle swarm algorithm, wherein the formula is as follows:
υ id ＝ε*(ω*υ id +c 1 *rand()*(p best -x id )+c 2 *rand()*(g best -x id ))
x id ＝x id +υ′ id
wherein epsilon is an energy factor; omega is the inertial weight; p is p best Representing the optimal position that each particle has found at present; g best Representing the optimal position found so far for the whole population of particles; c 1 And c 2 Respectively referred to as individual acceleration factors and social acceleration factors; representing pulling the particles towards p best Or g best Weights of random acceleration terms of (a); upsilon (v) id Representing the speed of the ith particle at the d-th iteration; x is x id Representation ofThe position of the ith particle at the d-th iteration; v' id Is the actual execution speed of the unmanned aerial vehicle under the speed constraint.
The embodiment introduces an adaptive strategy into the particle swarm algorithm, and dynamically adjusts the inertia weight omega and the individual acceleration factor c 1 And social acceleration factor c 2 The weights of the parameters are used for improving the searching capability and avoiding sinking into local optimum. When a search task just starts, the inertia weight takes the dominant position, and the value is 1, so that stronger global searching capability is obtained; along with the execution of the search task, the inertia weight is reduced from the maximum value, the social acceleration factor is increased from 0, and the attraction of the global optimum to the particles is enhanced so as to prevent the unmanned aerial vehicle from falling into the local optimum; after the social acceleration factor reaches the optimal value, the individual acceleration factor starts to increase from 0, at which point the individual's optimal attraction to the particles increases. Considering the energy limitation of the unmanned aerial vehicle, the objective rule that the searching capability of the unmanned aerial vehicle gradually decreases along with the time is simulated by adding an energy constraint which decays along with the time into a particle swarm algorithm. Wherein the energy factor epsilon follows the following decay law:
ε(t+1)＝λ*ε(t)
the lambda is an energy attenuation rate and is used for simulating the energy attenuation speed in the flight process of the unmanned aerial vehicle, and the lambda value can be 0.95.
In addition, taking into account the fact that the speed cannot reach infinity during the actual flight, a speed constraint is introduced into the unmanned aerial vehicle such that the speed of the unmanned aerial vehicle remains within the upper limit range, i.e. does not exceed the maximum speed V, while maintaining the direction of the speed vector max . The speed constraint is as follows:
and S5, if the termination condition is reached in advance or the iteration number is larger than the set highest iteration number, executing the step S6, otherwise, returning to the step S3 to carry out the next iteration.
S6, ending the search task.
To verify the performance of the proposed algorithm, simulation experiments were performed on a 2.30GHz Intel i7-11800H processor with 16.0GB memory and 512GB SSD, the simulation environment was Linux Ubuntu18.04, and in addition, rviz was used to visualize the entire search process.
The simulation results of the present invention are shown in fig. 4, 5, 6 and 7. According to the experimental results, when an experiment starts, 4 unmanned aerial vehicles fly basically according to a linear track under the premise of not encountering an obstacle due to the obvious influence of inertia weight, and the searching range can be expanded to the whole searching space in a short time; with the advancement of experiments, the unmanned aerial vehicle starts to generate autonomous steering behaviors under the condition of no obstacle interference, which means that the unmanned aerial vehicle enters a second stage, and the attractive force of the globally optimal position starts to be enhanced; in the final stage of the experiment, after 90s of the search task, the unmanned aerial vehicle can be observed to be attracted by the local optimal position and can be basically gathered near the sought target position; when the termination condition is reached, the simulation experiment is ended. In the whole process, the actual rule that the energy level of the unmanned aerial vehicle is reduced and the searching capability is gradually weakened along with the time is well simulated by observing that the searching step length of the unmanned aerial vehicle in each iteration is gradually reduced.
The simulation process well shows the collision prevention and collision avoidance functions of the designed mechanism while the basic function of searching is realized. During the flight, it can be observed that the relative target position of each unmanned aerial vehicle varies frequently as the iteration proceeds, and once a new target is calculated, a new trajectory is quickly planned and generated and executed. Once the generated path overlaps with the detected obstacle, the re-planning mechanism of the algorithm is triggered in a very short time. It is further observed that the re-planning mechanism may be triggered multiple times en route to the target until a smoother collision-free trajectory is generated to the target location of the current iteration. In the aspect of collision avoidance, as shown in fig. 8, when a plurality of unmanned aerial vehicles perform a search task while traversing a relatively dense obstacle gap, an inter-machine collision avoidance mechanism is triggered, and orderly passing among the plurality of unmanned aerial vehicles according to a priority order can be achieved.
In order to exclude contingency, the proposed particle swarm search mechanism is proved to have universality, the embodiment repeatedly performs the above experiment 100 times, each experiment reaches the termination condition or the collision between the unmanned aerial vehicle and the obstacle, the unmanned aerial vehicle is converged near the target position in a collision-free manner in the whole process, and finally the conclusion that the success rate of the algorithm is higher than 90% is obtained.
The simulation experiment shows that the method provided by the invention has a good searching effect when being applied to the unmanned aerial vehicle, can realize positioning to the target position in a short time, verifies that the introduced obstacle avoidance and collision avoidance functions are also very suitable for task execution of the group in a complex environment, and remarkably improves the practicability of the particle swarm algorithm.
Example 3:
the embodiment provides a multi-unmanned aerial vehicle cooperative target search system based on a particle swarm algorithm, which comprises the following steps:
a data acquisition module configured to: acquiring position information of a target to be searched;
a search control module configured to: controlling a plurality of unmanned aerial vehicles to move towards the position of the target to be searched for according to a preset particle swarm algorithm;
the method comprises the steps of introducing an energy factor which decays along with the time to simulate the law of the decline of searching capacity along with the time into a particle swarm algorithm, introducing a maximum speed constraint, and controlling the speed of the unmanned aerial vehicle in the calculated direction to be not more than the maximum speed constraint after the size and the direction of the speed at the next moment are calculated.
The working method of the system is the same as the multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm algorithm in embodiment 1, and is not described here again.
Example 4:
the present embodiment provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the multi-unmanned aerial vehicle cooperative target search method based on the particle swarm algorithm described in embodiment 1.
Example 5:
the embodiment provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm algorithm described in embodiment 1 when executing the program.
The above description is only a preferred embodiment of the present embodiment, and is not intended to limit the present embodiment, and various modifications and variations can be made to the present embodiment by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment should be included in the protection scope of the present embodiment.
Claims (10)
1. The multi-unmanned aerial vehicle cooperative target searching method based on the particle swarm optimization is characterized by comprising the following steps of:
acquiring position information of a target to be searched;
controlling a plurality of unmanned aerial vehicles to move towards the position of the target to be searched for according to a preset particle swarm algorithm;
the method comprises the steps of introducing an energy factor which decays along with the time to simulate the law of the decline of searching capacity along with the time into a particle swarm algorithm, introducing a maximum speed constraint, and controlling the speed of the unmanned aerial vehicle in the calculated direction to be not more than the maximum speed constraint after the size and the direction of the speed at the next moment are calculated.
2. The multi-unmanned aerial vehicle cooperative target searching method based on the particle swarm algorithm according to claim 1, wherein the target position from the current moment to the next moment of each unmanned aerial vehicle is calculated through the particle swarm algorithm, and after a guiding path from the current moment position to the next moment target position is determined, unmanned aerial vehicle control instructions for pushing away the obstacle are generated at all the positions with the obstacle on the guiding path; and when the obstacle on the guiding path is determined, other unmanned aerial vehicles which are in a preset range around the current controlled unmanned aerial vehicle and have higher priority than the current controlled unmanned aerial vehicle are regarded as the obstacle.
3. The multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm optimization according to claim 1, wherein an inertia weight parameter, an individual acceleration factor parameter and a social acceleration factor parameter are added in the particle swarm optimization, and the weights of the parameters are dynamically adjusted at each control stage.
4. The multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm optimization according to claim 3, wherein the inertial weight value is the largest at the beginning of the searching task; as the search task is executed, the inertia weight is reduced, and the social acceleration factor is increased; after the social acceleration factor reaches a preset value, the individual acceleration factor gradually rises.
5. The multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm optimization according to claim 1, wherein the energy factor at the next moment is equal to the product of the energy factor at the current moment and the energy attenuation rate, and the energy attenuation rate simulates the energy attenuation speed in the unmanned aerial vehicle flight process.
6. The multi-unmanned aerial vehicle collaborative target searching method based on a particle swarm algorithm according to claim 1, wherein each unmanned aerial vehicle is respectively set as a single particle in the particle swarm algorithm; the actual position where the drone is located is set as the particle position.
7. The multi-unmanned aerial vehicle collaborative target searching method based on the particle swarm algorithm according to claim 1, wherein the control process from the current time to the next time is divided into a searching stage and an optimizing stage; during the search phase, through A * The algorithm obtains a guiding path without considering the obstacle; in the optimizing stage, the surrounding environment of the unmanned aerial vehicle is detected, the track existing in the obstacle is compared with the collision-free guiding path generated in the searching stage, and the gradient information is utilized to pull out the track in the obstacle from the obstacle.
8. Particle swarm algorithm-based multi-unmanned aerial vehicle cooperative target search system is characterized by comprising:
a data acquisition module configured to: acquiring position information of a target to be searched;
a search control module configured to: controlling a plurality of unmanned aerial vehicles to move towards the position of the target to be searched for according to a preset particle swarm algorithm;
the method comprises the steps of introducing an energy factor which decays along with the time to simulate the law of the decline of searching capacity along with the time into a particle swarm algorithm, introducing a maximum speed constraint, and controlling the speed of the unmanned aerial vehicle in the calculated direction to be not more than the maximum speed constraint after the size and the direction of the speed at the next moment are calculated.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the particle swarm algorithm based multi-unmanned cooperative target search method according to any of claims 1 to 7.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the particle swarm algorithm-based multi-unmanned cooperative target search method according to any of claims 1-7 when executing the program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310279162.2A CN116203989A (en) | 2023-03-17 | 2023-03-17 | Multi-unmanned aerial vehicle cooperative target searching method and system based on particle swarm optimization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310279162.2A CN116203989A (en) | 2023-03-17 | 2023-03-17 | Multi-unmanned aerial vehicle cooperative target searching method and system based on particle swarm optimization |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116203989A true CN116203989A (en) | 2023-06-02 |
Family
ID=86514739
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310279162.2A Pending CN116203989A (en) | 2023-03-17 | 2023-03-17 | Multi-unmanned aerial vehicle cooperative target searching method and system based on particle swarm optimization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116203989A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117474432A (en) * | 2023-12-27 | 2024-01-30 | 运易通科技有限公司 | Unmanned logistics distribution method and system and unmanned aerial vehicle |
-
2023
- 2023-03-17 CN CN202310279162.2A patent/CN116203989A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117474432A (en) * | 2023-12-27 | 2024-01-30 | 运易通科技有限公司 | Unmanned logistics distribution method and system and unmanned aerial vehicle |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108594853B (en) | Unmanned aerial vehicle formation control method | |
| CN108549407B (en) | Control algorithm for multi-unmanned aerial vehicle cooperative formation obstacle avoidance | |
| CN108121358B (en) | Unmanned aerial vehicle control method | |
| CN113589842B (en) | Unmanned cluster task cooperation method based on multi-agent reinforcement learning | |
| CN109254588B (en) | Unmanned aerial vehicle cluster cooperative reconnaissance method based on cross variation pigeon swarm optimization | |
| CN110320930B (en) | Reliable transformation method for formation of multiple unmanned aerial vehicles based on Voronoi diagram | |
| Li et al. | Path planning of unmanned aerial vehicle based on improved gravitational search algorithm | |
| CN114413906B (en) | Three-dimensional trajectory planning method based on improved particle swarm optimization algorithm | |
| CN103744290B (en) | A kind of multiple no-manned plane formation layering target assignment method | |
| CN108153328B (en) | Multi-missile collaborative track planning method based on segmented Bezier curve | |
| CN112000131B (en) | Unmanned aerial vehicle cluster path planning method and system based on artificial potential field method | |
| CN109521794A (en) | A kind of multiple no-manned plane routeing and dynamic obstacle avoidance method | |
| CN112180967B (en) | Multi-unmanned aerial vehicle cooperative countermeasure decision-making method based on evaluation-execution architecture | |
| CN113157000B (en) | Flight formation cooperative obstacle avoidance self-adaptive control method based on virtual structure and artificial potential field | |
| CN116203989A (en) | Multi-unmanned aerial vehicle cooperative target searching method and system based on particle swarm optimization | |
| CN114397911A (en) | Unmanned aerial vehicle cluster confrontation decision-making method based on multiple intelligent agents | |
| CN111859541A (en) | PMADDPG multi-unmanned aerial vehicle task decision method based on transfer learning improvement | |
| CN114492805A (en) | Air combat maneuver decision design method based on fuzzy reasoning | |
| Zhao et al. | Four-dimensional trajectory generation for UAVs based on multi-agent Q learning | |
| CN114089776B (en) | Unmanned aerial vehicle obstacle avoidance method based on deep reinforcement learning | |
| CN113741186A (en) | Double-machine air combat decision method based on near-end strategy optimization | |
| CN111797966B (en) | Multi-machine collaborative global target distribution method based on improved flock algorithm | |
| CN111399534A (en) | Method and system for capturing aerial medium-high speed moving targets by multiple unmanned aerial vehicles | |
| CN112651486A (en) | Method for improving convergence rate of MADDPG algorithm and application thereof | |
| CN114721423B (en) | Distribution method for multi-unmanned aerial vehicle to cooperatively reach preset target in consideration of collision avoidance constraint |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |