Disclosure of Invention
In view of the above, the invention provides a method and a system for detecting a radiation environment based on a mobile robot, which are used for solving the problem of low accuracy of a path planning mode of the mobile robot in a radiation emergency treatment scene.
The invention discloses a radiation environment detection method based on a mobile robot, which is used for planning a radiation detection path of the mobile robot, wherein environment sensing equipment and a radiation detector are carried on the mobile robot, and the method comprises the following steps:
constructing a two-dimensional environment map of the target area through known barrier information;
acquiring coordinates of a starting point and an ending point, and the number of path points to be optimized;
constructing obstacle avoidance constraints based on known obstacle information, and establishing an objective function of global path planning by taking the shortest path length as an objective;
and solving an objective function of global path planning by using the path points as parameters through an improved social network searching algorithm to obtain a global optimal path.
On the basis of the above technical solution, preferably, the objective function of constructing obstacle avoidance constraints based on known obstacle information and establishing global path planning with the shortest path length as the objective specifically includes:
let the number of path points to be optimized beZThe coordinates of the starting point and the end point are respectively%x 0 ,y 0 )、(x Z+1 ,y Z+1 ) The coordinate set of the path points is {x z ,y z )|z=1,2,...,Z};
Objective function for global path planningFThe expression of (2) is:
wherein ,ωin order to penalize the coefficients,Q(t) As a penalty function of the obstacle avoidance constraint,t=1,2,…,T,Tas a total number of obstacles,r t is the firsttRadius of obstacle [ ]a t ,b t ) Is the firsttThe coordinates of the circle centers of the individual obstacles,His a safe distance.
On the basis of the above technical solution, preferably, the solving the objective function of the global path plan by the improved social network search algorithm, to obtain the global optimal path specifically includes:
initializing a population of a social network searching algorithm in a solution space of the path points, and taking the population as a team of a collaborative searching algorithm;
calculating the fitness of each individual by taking the minimum objective function as a fitness function, and selecting M individuals with smaller fitness from the fitness functions as leaders;
randomly selecting any one of behavior modes of imitation, dialogue, dispute or innovation according to a certain probability to update the population position; the team communication thought of the collaborative search algorithm is introduced into the dialogue behavior mode to update the population position;
repeating the above processes and performing iterative operation until the iteration stop condition is met, and outputting the optimal solution.
On the basis of the above technical solution, preferably, the group location updating by introducing the team communication idea of the collaborative search algorithm in the dialogue behavioral mode specifically includes:
wherein ,
respectively representing knowledge of one individual randomly selected from M leaders, collective knowledge of the M leaders and collective knowledge of all individuals, R represents interaction effect of the individual i and the randomly selected individual j,
the position of the d-th dimension data for the j-th individual at the kth iteration;
the position of the ith dimension of the ith individual at the kth, k+1 iterations,/for each individual>
Generating uniformly distributed random numbers within the range of 0-1, ">
A historical optimal solution for a randomly selected one of the M leaders,
bto obtain a new data set from {1,2, …,
Mrandomly selecting a sequence number; />
For the position of the mth individual, m=1, 2, …, M, +.>
For the optimal solution of the ith individual in the kth iteration, i=1, 2, …, N is the population number; alpha is,
βFor the purpose of weight adjustment of the coefficients,
f i 、
f j the fitness of individuals i, j, respectively.
On the basis of the technical scheme, preferably, the group location updating process of entering any one of the behavior modes of imitation, dialogue, dispute or innovation is randomly selected with a certain probability, and a grouping learning mechanism based on capability evaluation is introduced into the behavior mode of dispute to update the location:
wherein ,
for learning ability of individual i, updating by random mode, L is preset ability threshold, P is standard normal distribution with mean value of 0 and variance of 1, and->
Is the position mean of M leaders, and lambda is a constant with a value of 1 or 2.
On the basis of the above technical solution, preferably, the random selection of a certain probability is performed in a process of performing population location update in any one of a simulation, a dialogue, a dispute or an innovation, and in the innovation behavior mode, a location update formula is as follows:
wherein
μ (0,1),/>
For compliance parameter +.>
Is a distribution of the Lewy of (C).
On the basis of the above technical solution, preferably, the method further includes:
judging whether an unknown obstacle affecting the current global optimal path occurs or not through environment sensing equipment in the process that the mobile robot moves according to the global optimal path, if so, acquiring unknown obstacle information, updating a two-dimensional environment map of a target area, and carrying out global path planning again by taking the current position as a starting point; the environment sensing device comprises a camera and a laser radar.
In a second aspect of the present invention, a radiation environment detection system based on a mobile robot is disclosed, the system comprising:
the environment map building module: a two-dimensional environment map for constructing a target area by known obstacle information;
an objective function establishment module: the method is used for acquiring coordinates of a starting point and an ending point and the number of path points to be optimized; constructing obstacle avoidance constraints based on known obstacle information, and establishing an objective function of global path planning by taking the shortest path length as an objective;
and a path planning module: the method comprises the steps of using path points as parameters, and solving an objective function of global path planning through an improved social network searching algorithm to obtain a global optimal path;
and a path adjustment module: the method comprises the steps that in the process that a mobile robot moves according to a global optimal path, whether an unknown obstacle affecting the current global optimal path appears or not is judged through environment sensing equipment, if yes, information of the unknown obstacle is obtained, a two-dimensional environment map of a target area is updated, and global path planning is conducted again with the current position as a starting point; the environment sensing device comprises a camera and a laser radar.
In a third aspect of the present invention, an electronic device is disclosed, comprising: at least one processor, at least one memory, a communication interface, and a bus;
the processor, the memory and the communication interface complete communication with each other through the bus;
the memory stores program instructions executable by the processor which the processor invokes to implement the method according to the first aspect of the invention.
In a fourth aspect of the invention, a computer-readable storage medium is disclosed, storing computer instructions that cause a computer to implement the method according to the first aspect of the invention.
Compared with the prior art, the invention has the following beneficial effects:
1) According to the invention, a mobile robot is carried with a radiation detector for radiation environment detection, in the process of path planning, obstacle avoidance constraint is constructed based on known obstacle information, a global path planning objective function is established by taking the shortest path length as a target, a social network searching algorithm is improved through a collaborative searching algorithm, a hybrid intelligent optimization algorithm is formed for solving the objective function, the solving speed is ensured, and meanwhile, the solving quality is improved, so that the radiation detection efficiency is improved;
2) According to the invention, the team communication thought of the collaborative search algorithm is introduced into the dialogue behavioral mode of the social network search algorithm to update the population position, so that the information exchange among individuals is enhanced, and the grouping learning mechanism based on the capability evaluation is introduced into the dispute behavioral mode to update the position, thereby forming an improved social network search algorithm, reducing the defects of poor stability and easy sinking into a local optimal solution caused by strong randomness of the original social network search algorithm, and improving the robustness of the algorithm.
3) In the process that the mobile robot advances according to the global optimal path, the environment sensing equipment judges whether an unknown obstacle affecting the current global optimal path appears or not, and global path planning is conducted again by taking the current position as a starting point, so that the method is suitable for radiation detection under the condition that the environment changes or part of the environment is unknown, and the robustness and the adaptability of the mobile robot radiation detection are improved.
Detailed Description
The following description of the embodiments of the present invention will clearly and fully describe the technical aspects of the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.
The invention detects the radiation environment based on the mobile robot carrying environment sensing equipment and the radiation detector, and can be suitable for rapid radiation detection bottoming under the radiation emergency treatment scene. The robot path planning is an NP difficult problem in nature, and the method aims at the path planning problem of the mobile robot, establishes an objective function of global path planning by taking the shortest path length as a target, solves a global optimal path through an improved social network searching algorithm, and improves the path optimization speed. In addition, in the process that the mobile robot advances according to the global optimal path, unknown obstacles influencing the current global optimal path can be perceived in real time through the environment perception device, global path planning is conducted again by taking the current position as a starting point, and the robustness and the adaptability of the mobile robot radiation detection are improved.
Referring to fig. 1, the present invention provides a radiation environment detection method based on a mobile robot, the method includes:
s1, constructing a two-dimensional environment map of the target area through known obstacle information.
The target area is generally safe during the operation of the nuclear facility or during the daily environmental quality monitoring, at this time, the obstacle information in the environment is known, and when an emergency such as radiation leakage occurs, the environment of the target area may be changed, for example, an additional obstacle may occur, but basically, some changes occur on the basis of the original environment, so long as serious and destructive damage does not occur, and prior information such as the known obstacle information in the surrounding environment can be utilized.
The two-dimensional environment map of the target area is constructed by using known obstacle information through a grid method, and the information such as the central position, the occupied range and the like of the circular obstacle is marked in the two-dimensional environment map by abstracting the circular obstacle by taking the central coordinate of the obstacle as the center and the maximum distance of the edge of the obstacle from the central coordinate as the radius obstacle because the common obstacle is not necessarily in a regular shape.
S2, acquiring coordinates of a starting point and an ending point, and optimizing the number of the path points.
When the range of the area to be detected is not too large, directly taking the whole area to be detected as a target area, determining the starting point coordinates and the end point coordinates of the target area, and planning the path of the mobile robot in the target area. When the range of the area to be detected is too large, the area to be detected can be divided to form a plurality of target areas, a starting point and an end point are determined for each target area, mobile robot path planning is respectively carried out on each target area, and then path planning results of adjacent target areas are connected in series to form a complete robot radiation detection path.
In the target area, the number of path points to be optimized is required to be set, the number of the path points determines the dimension of path optimization, and the higher the dimension is, the higher the accuracy of path optimization is.
In the embodiment of the invention, the number of the path points needing to be optimized is set asZThe coordinates of the path points are%x 1 ,y 1 ),(x 2 ,y 2 ),...,(x Z ,y Z ) The coordinate set of the path points is {x z ,y z )|z=1,2,...,ZThe coordinates of the starting point and the end point are respectively%x 0 ,y 0 )、(x Z+1 ,y Z+1 ) The global path planning is to plan a safe path which starts from the starting point, passes through each path point to reach the ending point and does not collide with the obstacle.
S3, constructing obstacle avoidance constraints based on known obstacle information, and building an objective function of global path planning by taking the shortest path length as an objective.
Is provided withTAs the total number of obstacles in the target area,r t is the firsttRadius of obstacle [ ]a t ,b t ) Is the firsttThe coordinates of the circle centers of the individual obstacles,t=1,2,…,T。
since the mobile robot requires a certain safe working space, a safe distance is required to be set in order to avoid the obstacle during the working processHI.e. the minimum distance of the waypoint to the edge of the obstacle.
Establishing an objective function of global path planning by taking shortest path length under obstacle avoidance constraint as a target, wherein the objective function isFThe expression of (2) is:
wherein ,ωin order to penalize the coefficients,Q(t) Is the firsttPenalty function of obstacle avoidance constraints for individual obstacles.
And S4, solving an objective function of global path planning by using the path points as parameters through an improved social network searching algorithm to obtain a global optimal path.
Social network search algorithms (Social Network Search, SNS) mainly simulate the behavior of users when expressing opinion: imitates, dialogues, disputes and innovations, has the advantage of fast convergence speed. The algorithm enters a simulation, dialogue, dispute or innovation behavior mode through a random selection mode, in each behavior mode, opinion exchange is basically carried out in a random selection mode, and although strong randomness can search in a large range in a search space, due to strong randomness, the algorithm can miss an optimal solution to be trapped into local optimization so as to cause premature convergence.
The step S42 specifically includes the following sub-steps:
s421, initializing a population of a social network searching algorithm in a solution space of the path points, and taking the population as a team of the collaborative searching algorithm.
In the invention, a two-dimensional environment map is created by a grid method in the step S1, grids where obstacles are located are removed from the two-dimensional environment map, a feasible path set formed by the remaining grids is a solution space of path points, a population is initialized in the solution space, each individual in the population represents a path, the total number Z of the path points is the same as the vector dimension D of the individual in the population, the number N of the population is set, and the population is used as a team of a collaborative search algorithm. And initializing other initial parameters of the social network searching algorithm, the maximum iteration number K and the like.
S422, calculating fitness of each individual by taking an objective function of global path planning as a fitness function, and selecting M individuals with smaller fitness as leaders.
I.e. the fitness function is F, the smaller the fitness function value, the higher the quality of the representation solution.
And (5) carrying out ascending order on the fitness values of the individuals, and taking the first M individuals as leaders.
S423, randomly selecting any one of behavior modes of imitation, dialogue, dispute or innovation with a certain probability to update the population position.
Specifically, generating a random number r within a range of 0-1, and entering an imitation behavior mode if r is less than or equal to 0.25; entering a dialogue behavior mode when r is more than 0.25 and less than or equal to 0.5; if r is less than or equal to 0.75 and is 0.5, entering a dispute behavior mode, and if r is more than 0.75, entering an innovation behavior mode.
In the simulated behavior mode, a conversation is randomly selected
The object is imitated by adopting the following position updating mode:
wherein ,
the position of the ith dimension of the ith individual at the kth, k+1 iterations,/for each individual>
D = 1,2, …, D for the position of the D-th dimension data of the j-th individual at the kth iteration; n is population number; i=1, 2, …, N, k=1, 2, …, K.
In the dialogue behavioral model, team communication ideas of a collaborative search algorithm are introduced to update individual positions:
wherein ,
each representing knowledge of one individual randomly selected from M leaders, collective knowledge of all individuals, R representing interaction effect of individual i with randomly selected individual j>
The d-th dimension data for the j-th individual is at
kThe position at the next iteration;
generating uniformly distributed random numbers within the range of 0-1, ">
A historical optimal solution for an individual randomly selected from the M leaders, b is an individual serial number randomly selected from {1,2, …, M }; />
For the position of the mth individual, m=1, 2, …, M, +.>
Representing the optimal position mean of the M leaders; />
For the optimal solution of the ith individual at the kth iteration,/for the kth iteration>
Representing a position mean of the population; alpha, & alpha>
For the purpose of weight adjustment of the coefficients,
f i 、
f j the fitness of individuals i, j, respectively.
According to the invention, by introducing the team communication thought of the collaborative search algorithm in the dialogue behavior mode, the information exchange among groups is enhanced, the optimal solution is conveniently and quickly found, and the problems that the information from only one individual exchange generation is limited and is easy to fall into the local optimal solution are avoided.
In the dispute behavior mode, a grouping learning mechanism based on capability evaluation is introduced for individual position updating:
wherein ,
for learning ability of individual i, L.sub.L.is updated in a random manner>
(0, 1) is a preset capacity threshold, P is a standard normal distribution with mean 0 and variance 1,/for>
Is the position mean value of M leaders, and lambda is a constant with the value of 1 or 2.
According to the method, through a grouping learning mechanism based on capability evaluation, a group with strong learning capability is focused on improving the position goodness of the user, the user continues to search for a better solution near the user's position, a group with weak learning capability is focused on breaking through the limitation of the user, and the user searches for a position with possibly higher adaptability in a larger search range, so that the balance between local exploration and global development is considered.
In the innovative behavior mode, a Lewy flight strategy is introduced for position update:
wherein
μ (0,1),/>
For compliance parameter +.>
Is a distribution of the Lewy of (C).
In the innovative behavior mode, the invention replaces the original linear random generation strategy by the Lewy flight strategy, and the Lewy flight is also a random walk strategy, but can simulate animal foraging behavior to a certain extent, and experimental research shows that the searching efficiency is higher.
S424, repeating the processes of the steps S422-S423, and performing iterative operation until the iteration stop condition is met, and outputting the optimal solution as a global optimal path.
In the process of carrying out radiation environment detection path planning by adopting the mobile robot to carry the radiation detector, the invention constructs obstacle avoidance constraint based on known obstacle information, establishes an objective function of global path planning by taking the shortest route length as a target, improves a social network searching algorithm by a collaborative searching algorithm, forms a hybrid intelligent optimization algorithm to solve the objective function, and improves the quality of solution while ensuring the solving speed, thereby improving the efficiency of radiation detection.
S5, in the process that the mobile robot moves according to the global optimal path, two-dimensional environment map updating and global path planning adjustment are carried out through the environment sensing equipment.
And (4) carrying an environment sensing device and an environment detector by the mobile robot to detect radiation according to the globally optimal path solved in the step (S4), wherein the environment sensing device comprises a camera and a laser radar.
In the process of detecting radiation along the global optimal path, the mobile robot senses the surrounding environment in real time through the environment sensing equipment, and when a certain obstacle is sensed, whether the obstacle is a known obstacle is judged. If the target area is the unknown obstacle, judging whether the unknown obstacle affecting the current global optimal path appears, if so, acquiring the information of the unknown obstacle, updating the two-dimensional environment map of the target area, and re-executing the steps S2-S4 to carry out global path planning by taking the current position as a starting point.
The mobile robot of the invention can carry out dynamic path adjustment based on the environmental perception result at any time in the running process according to the planned route, can be suitable for radiation detection under the condition that the environment is changed or the condition that part of the environment is unknown, and can improve the robustness and adaptability of the radiation detection of the mobile robot
Corresponding to the embodiment of the method, the invention also discloses a radiation environment detection system based on the mobile robot, which comprises the following steps:
the environment map building module: a two-dimensional environment map for constructing a target area by known obstacle information;
an objective function establishment module: the method is used for acquiring coordinates of a starting point and an ending point and the number of path points to be optimized; constructing obstacle avoidance constraints based on known obstacle information, and establishing an objective function of global path planning by taking the shortest path length as an objective;
and a path planning module: the method comprises the steps of using path points as parameters, and solving an objective function of global path planning through an improved social network searching algorithm to obtain a global optimal path;
and a path adjustment module: the method comprises the steps that in the process that a mobile robot moves according to a global optimal path, whether an unknown obstacle affecting the current global optimal path appears or not is judged through environment sensing equipment, if yes, information of the unknown obstacle is obtained, a two-dimensional environment map of a target area is updated, and global path planning is conducted again with the current position as a starting point; the environment sensing device comprises a camera and a laser radar.
The system embodiments and the method embodiments are in one-to-one correspondence, and the brief description of the system embodiments is just to refer to the method embodiments.
The invention also discloses an electronic device, comprising: at least one processor, at least one memory, a communication interface, and a bus; the processor, the memory and the communication interface complete communication with each other through the bus; the memory stores program instructions executable by the processor that the processor invokes to implement the aforementioned methods of the present invention.
The invention also discloses a computer readable storage medium storing computer instructions for causing a computer to implement all or part of the steps of the methods of the embodiments of the invention. The storage medium includes: a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The system embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, i.e., may be distributed over a plurality of network elements. One of ordinary skill in the art may select some or all of the modules according to actual needs without performing any inventive effort to achieve the objectives of the present embodiment.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.