An ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function

Abstract

The application discloses an ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function, and belongs to the technical field of intelligent optimization and automatic control. The application solves the problems of slow convergence speed, easy falling into local optimum and poor environmental adaptability of the existing ant colony algorithm. The application firstly pre-plans feasible paths, so as to realize non-uniform pheromone initialization according to the pre-planned paths, and improve the directionality and convergence speed of initial search; an obstacle factor is introduced into a state transition probability function, so as to enhance the obstacle avoidance ability of the algorithm in a complex environment; an adaptive coefficient based on normal distribution is designed, so as to dynamically balance global and local search; and an adaptive pheromone evaporation coefficient mechanism is combined, so as to prevent premature convergence. The method is suitable for application fields such as micro operation, mobile robots and intelligent manufacturing, which have high requirements on path precision and real-time performance.

Description

Technical Field

[0001] This invention belongs to the field of intelligent optimization and automatic control technology, specifically relating to an ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function. Background Technology

[0002] Path planning is a core technology for autonomous mobile robots and micro-manipulation platforms to achieve automatic navigation and operation. Ant colony optimization (ACO) algorithms are widely used in path planning problems in complex environments due to their distributed computing, positive feedback mechanism, and strong robustness. However, traditional ACO algorithms suffer from problems in the early stages of application, such as uniform pheromone distribution and strong randomness in ant search directions, leading to low search efficiency, slow convergence speed, and often redundant paths with poor smoothness. Furthermore, in complex environments with dense obstacles, ants are prone to "self-locking" or local optima, making it difficult to plan short, smooth, and collision-free paths.

[0003] While existing improvement methods enhance performance by introducing strategies such as chaotic mapping and hybrid particle swarm optimization, most fail to systematically address issues such as weak initial search guidance, poor environmental adaptability, and the difficulty in balancing global exploration and local development. Especially in high-precision applications like micro-operations, where there are extremely high requirements for path length, smoothness, and real-time planning, traditional improvement methods still struggle to meet these demands.

[0004] Therefore, there is an urgent need to propose an improved ant colony optimization method that can converge quickly, adapt to complex obstacle environments, and plan the optimal smooth path. Summary of the Invention

[0005] This invention addresses the problems of slow convergence, susceptibility to local optima, and poor environmental adaptability in existing ant colony optimization algorithms by proposing an ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function.

[0006] The technical solution adopted by this invention to solve the above-mentioned technical problems is: an ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function, the method specifically including the following steps:

[0007] Step 1: Divide the map model of the micro-operation working environment into grids to obtain the grid map model. Mark the start point, end point, and location of each obstacle in the micro-operation working environment in the grid map model.

[0008] Step 2: Plan an initial path from the starting point to the ending point on the constructed raster map model. ;

[0009] Step 3: Initialize the ant colony algorithm parameters, specifically including the number of ants. Maximum number of iterations Pheromones as heuristic factors Expectation heuristic factor The initial pheromone concentration and initial pheromone volatility coefficient of each edge. ;

[0010] Step 4: Initialize the number of iterations ;

[0011] Step 5: Place each ant at the starting point. Each ant starts from the starting point and selects the next moving node step by step according to the state transition probability function.

[0012] Step Six, in the... After the iteration, according to the... During each iteration, the pheromone concentration is updated along the paths of each ant that successfully reaches the destination.

[0013] Step 7: Determine whether the iteration termination condition is met;

[0014] If the iteration termination condition is not met, proceed to step eight;

[0015] If the iteration termination condition is met, output the optimal path obtained during the entire iteration process;

[0016] Step 8: Determine if the pheromone volatile coefficient update condition is met:

[0017] If the conditions for updating the pheromone evaporation coefficient are met, adjust the pheromone evaporation coefficient and then proceed to step nine; otherwise, proceed directly to step nine.

[0018] Step Nine, Order Return to step five.

[0019] Furthermore, step two employs... algorithm.

[0020] Furthermore, the initial pheromone concentration values ​​of each edge are determined according to the initial path. The calculation is as follows:

[0021] Treat each grid cell as a node, and for any two adjacent nodes... and nodes Connecting nodes and nodes The initial pheromone concentration of the edge is:

[0022]

[0023] in, For connecting nodes and nodes The initial pheromone concentration value of the edge, The preset baseline pheromone concentration, and All are constants greater than 1. For path The length.

[0024] Furthermore, connecting nodes and nodes The edge is The surrounding area refers to the connecting nodes. and nodes Edges and paths The minimum distance is less than the set threshold.

[0025] Furthermore, the state transition probability function is:

[0026]

[0027] in, For the first During the iteration, the first Only ants from the node To the node The state transition probability; Pheromones are the inductive factor; For expected heuristic factors; Weights for obstacle factors; For the first During the next iteration, the nodes are connected. and nodes edge The concentration of pheromones on the surface; For the first During the next iteration, the edge The inspirational information above; For the first The set of possible next nodes for each ant; Indicates about nodes Obstacles.

[0028] Furthermore, the barrier factor for:

[0029]

[0030] in, Represents nodes The number of adjacent grid cells that contain obstacles; For nodes The total number of adjacent grid cells.

[0031] Furthermore, the heuristic information for:

[0032]

[0033] in, For nodes To the node European distance, It is a constant. These are adaptive coefficients.

[0034] Furthermore, the adaptive coefficient for:

[0035]

[0036] in, It is a constant; This represents the maximum number of iterations.

[0037] Furthermore, the iteration stopping condition is:

[0038] Condition (1) The maximum number of iterations set has been reached. ;

[0039] Condition (2) Adaptively adjust the pheromone evaporation coefficient The path cannot be further optimized.

[0040] Furthermore, the specific process of step eight is as follows:

[0041] If the first The iteration continues to the... During each iteration, there was no correlation with the first... If a better path is generated before the nth iteration, then let the nth iteration... The pheromone evaporation coefficient of the next iteration ;

[0042] in, Adjust the step size for the pheromone evaporation coefficient; The set threshold;

[0043] If the first The iteration continues to the... During the iteration, all are relative to the first iteration. If a better path is found before the nth iteration, then let the nth iteration... The pheromone evaporation coefficient of the next iteration ;

[0044] Otherwise, let the first The pheromone evaporation coefficient of the next iteration .

[0045] The beneficial effects of this invention are:

[0046] This invention is based on The algorithm generates initial paths with differentiated pheromone allocation, introducing a non-uniform pheromone initialization strategy that significantly reduces search randomness in the early stages and improves the efficiency of generating feasible solutions. Furthermore, obstacle avoidance factors and adaptive coefficients based on normal distribution are incorporated into the state transition probability function to dynamically balance global exploration and local exploitation capabilities, effectively reducing "self-locking" and local optima problems in path planning. In addition, an adaptive adjustment mechanism for the pheromone evaporation coefficient is designed to ensure rapid convergence even in complex environments, enhancing the adaptability of the proposed method. Attached Figure Description

[0047] Figure 1 This is a flowchart of an ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to the present invention.

[0048] Figure 2 This is a path planning result diagram in a 40×40 grid environment;

[0049] Figure 3 This is a schematic diagram of the path of the first movement in a real-world micro-operation scenario;

[0050] Figure 4 This is a schematic diagram of the path of the second movement in a real-world micro-operation scenario;

[0051] Figure 5 This is a schematic diagram of the path for the third movement in a real-world micro-operation scenario. Detailed Implementation

[0052] Specific implementation method one: Combining Figure 1 This embodiment describes an ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function. The method specifically includes the following steps:

[0053] Step 1: Divide the map model of the micro-operation working environment into grids to obtain the grid map model. Mark the start point, end point, and location of each obstacle in the micro-operation working environment in the grid map model.

[0054] Step Two: Utilize The algorithm plans an initial path from the starting point to the ending point on the constructed raster map model. ;

[0055] Step 3: Initialize the ant colony algorithm parameters, specifically including the number of ants. Maximum number of iterations Pheromones as heuristic factors Expectation heuristic factor The initial pheromone concentration and initial pheromone volatility coefficient of each edge. ;

[0056] Furthermore, the initial pheromone concentration values ​​of each edge are based on the initial path. Calculate, specifically:

[0057] Treat each grid cell as a node, and for any two adjacent nodes... and nodes Connecting nodes and nodes The initial pheromone concentration of the edge is:

[0058]

[0059] in, For connecting nodes and nodes The initial pheromone concentration value of the edge, The preset baseline pheromone concentration, and All are constants greater than 1. For path Length;

[0060] Furthermore, it should be noted that the connection node... and nodes The edge is The surrounding area refers to the connecting nodes. and nodes Edges and paths The minimum distance is less than a set threshold, which can be set based on experience to balance exploration needs and planning efficiency.

[0061] This step is based on The algorithm plans an initial path that initializes the pheromone concentration of the grid cells in the map using a non-uniform method. The ant colony assigns a higher initial pheromone concentration to the grids it passes through, and assigns a relatively higher initial pheromone concentration to the grids around the initial path. This strategy can effectively guide the ant colony's initial search direction, reduce the randomness of the ant colony's initial search, and accelerate the generation of high-quality feasible solutions.

[0062] Step 4: Initialize the number of iterations ;

[0063] Step 5: Clear the historical path record, place each ant at the starting point, and each ant starts from the starting point and selects the next moving node step by step according to the state transition probability function;

[0064] Step five of this invention improves the state transition probability function. The improved state transition probability function integrates pheromone concentration, heuristic distance, obstacle avoidance factor, and coefficients for adaptive adjustment of heuristic information. Specifically, the improved state transition probability function is as follows:

[0065]

[0066] in, For the first During the iteration, the first Only ants from the node To the node The state transition probability; Pheromones are the inductive factor; For expected heuristic factors; The weight of the obstacle factor can be set based on experience; in this invention, it is set to 2. For the first During the next iteration, the nodes are connected. and nodes edge The concentration of pheromones on the surface; For the first During the next iteration, the edge The inspirational information above; For the first The set of possible next nodes for each ant; Indicates about nodes The obstacle factor enables ants to actively avoid areas with dense obstacles when choosing a path, effectively reducing the "deadlock" phenomenon.

[0067]

[0068] in, Represents nodes The number of adjacent grid cells that contain obstacles; For nodes The total number of adjacent grid cells;

[0069]

[0070] in, For nodes To the node Euclidean distance (i.e., edge distance) The length, and by node The center and nodes of the corresponding grid The Euclidean distance of the center of the corresponding grid is used as the edge. (length) It is a constant. The adaptive coefficient decreases as the number of iterations increases, which makes the algorithm value heuristic information in the early stage to quickly approach the target, and weaken its influence in the later stage to conduct a fine search, thus avoiding getting trapped in local optima.

[0071]

[0072] in, It is a constant; This represents the maximum number of iterations.

[0073] Step Six, in the... After the iteration, according to the... During each iteration, the pheromone concentration is updated along the paths of each ant that successfully reaches the destination.

[0074] In this step, path length is used as an indicator to evaluate path quality. When updating pheromone concentration, the pheromone concentration of better paths is increased, while the pheromone concentration of poorer paths is decreased to avoid interference from invalid paths. Upper and lower limits for pheromone concentration are also set. The pheromone concentration needs to be constrained when the updated concentration exceeds the upper or lower limit of the range. For example, when the lower limit of the pheromone concentration is exceeded, the pheromone concentration should be adjusted to... When the pheromone concentration exceeds the upper limit, the pheromone concentration will be adjusted to... This is to prevent premature convergence.

[0075] Step 7: Determine whether the iteration termination condition is met;

[0076] Specifically, the iteration stopping condition is set to condition (1) or condition (2):

[0077] Condition (1) The maximum number of iterations set has been reached. ;

[0078] Condition (2) Adaptively adjust the pheromone evaporation coefficient The path cannot be further optimized; that is, if no better path has been generated in N consecutive iterations after the last adjustment of the pheromone evaporation coefficient, the pheromone evaporation coefficient is reduced by one step, and the iteration continues using the adjusted pheromone evaporation coefficient. If no better path has been generated compared to the previous iterations after the last adjustment of the pheromone evaporation coefficient, the iteration stopping condition is met. Here, the iteration stopping condition can be set according to experience when no better path is generated in several iterations.

[0079] If the iteration termination condition is not met, proceed to step eight;

[0080] If the iteration termination condition is met, the optimal path obtained during the entire iteration process (i.e. the path with the shortest path length) is output, and the path length and search time of the optimal path are obtained.

[0081] Step 8: Determine if the pheromone volatile coefficient update condition is met:

[0082] If the first The iteration continues to the... During each iteration, there was no correlation with the first... If a better path is generated before the nth iteration, then let the nth iteration... The pheromone evaporation coefficient of the next iteration Enhance global search capabilities;

[0083] in, Adjust the step size for the pheromone evaporation coefficient; The threshold value is set (in this invention, it is set to 5).

[0084] If the first The iteration continues to the... During the iteration, all are relative to the first iteration. If a better path is found before the nth iteration, then let the nth iteration... The pheromone evaporation coefficient of the next iteration To accelerate convergence;

[0085] Otherwise, let the first The pheromone evaporation coefficient of the next iteration ;

[0086] By adaptively adjusting the pheromone evaporation coefficient, the algorithm's exploration and development capabilities can be dynamically balanced. Since the pheromone concentration of each edge is updated after each iteration, the pheromone evaporation coefficient can be updated repeatedly, allowing for the exploration of completely different paths.

[0087] Step Nine, Order Return to step five.

[0088] Experimental Section

[0089] like Figure 2 As shown, this invention is in 40 The path planning results in a 40-grid environment, and its simulation process includes:

[0090] Environment Setup: Build 40 The map has 40 grids, with obstacles accounting for approximately 25% of the area. The starting coordinates are (1,40) and the ending coordinates are (40,1). In this environment, the obstacles are more densely distributed, which places higher demands on the algorithm's global search and obstacle avoidance capabilities.

[0091] Set the number of ants Maximum number of iterations Pheromones as heuristic factors Expectation factor Barrier factor weights Initial pheromone concentration ,constant , , , .

[0092] use Algorithm planning initial path The non-uniform pheromone concentration is initialized based on the differential pheromone formula; ants select the next node according to the improved state transition probability function to avoid obstacles; and the pheromone concentration is adjusted according to the adaptive pheromone update mechanism after each iteration.

[0093] The final results show that, under these conditions, the optimal path length planned by the improved ant colony algorithm of this invention is 64.775 μm, and the search time is 42.931 s. Compared with the optimal path length planned by the traditional ant colony algorithm, the optimal path length is reduced by 12.36%, and the search time is shortened by 29.01%.

[0094] like Figures 3 to 5 The diagram shows an application of the method of this invention in a micro-manipulation assembly scenario. Specifically, it includes the following:

[0095] Three microspheres are set as movable manipulators, three different shaped micro-components are set as docking targets, and a pipette is set as a manipulator. Multiple static metal micro-components exist in the environment as obstacles to simulate interference conditions in a real assembly environment.

[0096] A grid-based method was used to construct a micro-manipulation environment map, setting the initial position of the microsphere as the starting point and the docking points of irregular micro-components as the ending points. The proposed method was used to plan a collision-free optimal path. The microsphere assembly was completed by executing path movements through an electrically controlled stepper motor micro-motion platform and a piezoelectric ceramic micro-motion platform. The final results show that the optimal path planned by the improved ant colony algorithm of this invention has a length of 4290.273 μm and a total search time of 124.07 s, which is 43.48% shorter than the optimal path planned by the traditional ant colony algorithm. The search time is reduced by 20.61%. The assembly process is collision-free, and the docking position deviation is within 1 μm, meeting the high efficiency and high precision requirements of micro-assembly tasks.

[0097] The above examples of the present invention are merely illustrative of the computational model and process of the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is impossible to exhaustively list all possible implementations here. Any obvious variations or modifications derived from the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. An ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function, characterized in that, The method specifically includes the following steps: Step 1: Divide the map model of the micro-operation working environment into grids to obtain the grid map model. Mark the start point, end point, and location of each obstacle in the micro-operation working environment in the grid map model. Step 2: Plan an initial path from the starting point to the ending point on the constructed raster map model. ; Step 3: Initialize the ant colony algorithm parameters, specifically including the number of ants. Maximum number of iterations Pheromones as heuristic factors Expectation heuristic factor The initial pheromone concentration and initial pheromone volatility coefficient of each edge. ; Step 4: Initialize the number of iterations ; Step 5: Place each ant at the starting point. Each ant starts from the starting point and selects the next moving node step by step according to the state transition probability function. Step Six, in the... After the iteration, according to the... During each iteration, the pheromone concentration is updated along the paths of each ant that successfully reaches the destination. Step 7: Determine whether the iteration termination condition is met; If the iteration termination condition is not met, proceed to step eight; If the iteration termination condition is met, output the optimal path obtained during the entire iteration process; Step 8: Determine if the pheromone volatile coefficient update condition is met: If the conditions for updating the pheromone evaporation coefficient are met, adjust the pheromone evaporation coefficient and then proceed to step nine; otherwise, proceed directly to step nine. Step Nine, Order Return to step five.

2. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 1, characterized in that, The method used in step two is... algorithm.

3. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 2, characterized in that, The initial pheromone concentration values ​​of each edge are based on the initial path. The calculation is as follows: Treat each grid cell as a node, and for any two adjacent nodes... and nodes Connecting nodes and nodes The initial pheromone concentration of the edge is: in, For connecting nodes and nodes The initial pheromone concentration value of the edge, The preset baseline pheromone concentration, and All are constants greater than 1. For path The length.

4. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 3, characterized in that, Connecting nodes and nodes The edge is The surrounding area refers to the connecting nodes. and nodes Edges and paths The minimum distance is less than the set threshold.

5. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 4, characterized in that, The state transition probability function is: in, For the first During the iteration, the first Only ants from the node To the node The state transition probability; Pheromones are the inductive factor; For expected heuristic factors; Weights for obstacle factors; For the first During the next iteration, the nodes are connected. and nodes edge The concentration of pheromones on the surface; For the first During the next iteration, the edge The inspirational information above; For the first The set of possible next nodes for each ant; Indicates about nodes Obstacles.

6. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 5, characterized in that, The obstacle factor for: in, Represents nodes The number of adjacent grid cells that contain obstacles; For nodes The total number of adjacent grid cells.

7. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 6, characterized in that, The heuristic information for: in, For nodes To the node European distance, It is a constant. These are adaptive coefficients.

8. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 7, characterized in that, The adaptive coefficient for: in, It is a constant; This represents the maximum number of iterations.

9. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 8, characterized in that, The iteration stopping condition is: Condition (1) The maximum number of iterations set has been reached. ; Condition (2) Adaptively adjust the pheromone evaporation coefficient The path cannot be further optimized.

10. The ant colony optimization path planning method based on non-uniform pheromone initialization and adaptive heuristic function according to claim 9, characterized in that, The specific process of step eight is as follows: If the first The iteration continues to the... During each iteration, there was no correlation with the first... If a better path is generated before the nth iteration, then let the nth iteration... The pheromone evaporation coefficient of the next iteration ; in, Adjust the step size for the pheromone evaporation coefficient; The set threshold; If the first The iteration continues to the... During the iteration, all are relative to the first iteration. If a better path is found before the nth iteration, then let the nth iteration... The pheromone evaporation coefficient of the next iteration ; Otherwise, let the first The pheromone evaporation coefficient of the next iteration .

Images