Robot track planning method and system for electric vehicle stator welding

Abstract

The invention discloses a robot trajectory planning method and system for electric vehicle stator welding, and belongs to the field of industrial robots. An optimal welding path is found through an integer programming algorithm; setting a safe distance; calculating a joint angle; possible collision is judged and processed by detecting the geometrical relationship between an obstacle and a robot connecting rod; smoothing the planned welding track by using a quintic spline interpolation method; the speed and the acceleration of the robot are limited by setting the initial speed, the acceleration and the termination condition of an end effector of the robot; introducing an optimization algorithm to carry out iterative optimization on the trajectory, and calculating a joint angle condition number, a global joint motion amount and a robot flexibility index; selecting an optimal position and path through a cost function and a collision detection algorithm; and updating the population and carrying out iterative optimization again. According to the method, the optimal smooth trajectory can be planned offline, the working efficiency of the robot is remarkably improved, and the preliminary planning and debugging time of an engineer is shortened.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial robots, and in particular relates to a robot trajectory planning method and system for stator welding of electric vehicles. Background Art

[0002] The rapid growth of the electric vehicle market is driving increasing performance requirements for motor manufacturing, particularly in welding processes, which place ever-increasing demands on precision and efficiency. Therefore, to meet the growing demands of the electric vehicle market, motor manufacturers are constantly seeking ways to increase motor power and performance. Flat-wire motors, which use flat copper wire inserted into stator slots, offer a greater fill factor than traditional motors, increasing power density and heat load capacity, enhancing motor reliability and stability.

[0003] To overcome the limitations of traditional laser welding technology used for electric vehicle stators, galvanometer-scanning laser welding has been developed. It offers the advantages of fast positioning speed and high welding precision. However, the diverse weld profiles encountered with this technology present new technical challenges for robot motion planning, such as difficulty in trajectory planning and low robot efficiency. Furthermore, galvanometer-scanning laser welding can easily damage the enameled wire film, impacting production efficiency and preventing it from fully utilizing the advantages of fast positioning speed. Summary of the Invention

[0004] To overcome the technical issues of existing galvanometer-scanning laser welding technology, the present invention provides a robot trajectory planning method and system for electric vehicle stator welding. This collaborative trajectory planning method and system, based on the optimization of dynamic and static performance indicators, is primarily applicable to electric vehicle stator laser welding robots.

[0005] The technical solutions adopted by the present invention to solve the technical problems are as follows:

[0006] The beneficial effects of the present invention are:

[0007] The present invention provides a robot trajectory planning method and system for electric vehicle stator welding, which adopts an integer programming algorithm for path planning, and combines quintic spline interpolation and an improved chicken flock optimization algorithm to achieve trajectory optimization. Specifically, it includes initializing the environment and robot model, setting parameters and updating strategies, calculating joint angle condition numbers and motion inverse solutions, etc. Through the method of the present invention, the optimal smooth trajectory can be planned offline, which significantly improves the work efficiency of the robot and shortens the initial planning and debugging time of engineers. Experimental results show that the method of the present invention can effectively suppress the impact of the robot system and obtain an obstacle avoidance trajectory with optimal time and continuous acceleration, which provides a reference value for the further application of engineering machinery automation and industrial robot technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 It is the iterative evolution curve of the algorithm generation value of the present invention.

[0009] Figure 2 is the angular acceleration of the robot end effector joint 1 of the present invention.

[0010] Figure 3 is the angle of the robot end effector joint 1 of the present invention.

[0011] Figure 4 is the angular velocity of the robot end effector joint 1 of the present invention.

[0012] Figure 5 is the angular acceleration of the robot end effector joint 1 of the present invention.

[0013] Figure 6 is the angle of the robot end effector joint 2 of the present invention.

[0014] Figure 7 is the angular velocity of the joint 2 of the robot end effector of the present invention.

[0015] Figure 8 is the angular acceleration of the robot end effector joint 2 of the present invention.

[0016] Figure 9 is the angular acceleration of the robot end effector joint 2 of the present invention.

[0017] Figure 10 is the angle of the robot end effector joint 3 of the present invention.

[0018] Figure 11 is the angular velocity of the joint 3 of the robot end effector of the present invention.

[0019] Figure 12 is the angular acceleration of the joint 3 of the robot end effector of the present invention.

[0020] Figure 13 is the angular acceleration of the joint 3 of the robot end effector of the present invention.

[0021] Figure 14 is the angle of the robot end effector joint 4 of the present invention.

[0022] Figure 15 is the angular velocity of the joint 4 of the robot end effector of the present invention.

[0023] Figure 16 is the angular acceleration of the robot end effector joint 4 of the present invention.

[0024] Figure 17is the angular acceleration of the joint 4 of the robot end effector of the present invention.

[0025] Figure 18 is the angle of the robot end effector joint 5 of the present invention.

[0026] Figure 19 is the angular velocity of the robot end effector joint 5 of the present invention.

[0027] Figure 20 is the angular acceleration of the robot end effector joint 5 of the present invention.

[0028] Figure 21 is the angular acceleration of the robot end effector joint 5 of the present invention.

[0029] Figure 22 is the angle of the robot end effector joint 6 of the present invention.

[0030] Figure 23 is the angular velocity of the robot end effector joint 6 of the present invention.

[0031] Figure 24 is the angular acceleration of the robot end effector joint 6 of the present invention.

[0032] Figure 25 is the angular acceleration of the robot end effector joint 6 of the present invention.

[0033] Figure 26 It is the iterative evolution curve of the robot flexibility index of the present invention.

[0034] Figure 27 It is the iterative evolution curve of the laser welding gun accessibility index of the present invention.

[0035] Figure 28 It is the iterative evolution curve of the robot's global joint motion index of the present invention.

[0036] Figure 29 It is the iterative evolution curve of the robot installation position of the present invention.

[0037] Figure 30 This is a flow chart of a robot trajectory planning method for electric vehicle stator welding provided by the present invention. DETAILED DESCRIPTION

[0038] The present invention is further described in detail below with reference to the accompanying drawings.

[0039] In a first aspect, the present invention provides a robot trajectory planning method for electric vehicle stator welding.

[0040] See also Figure 30To illustrate, the present invention provides a robot trajectory planning method for electric vehicle stator welding, and its specific implementation process is as follows:

[0041] Step S1: Robot initialization, including environment, model and related parameter settings;

[0042] Step S2: Use integer programming algorithm to find the welding path to ensure that the robot performs welding tasks in the optimal order;

[0043] Among them, when the integer programming algorithm is used for undirected graph path planning, for two nodes with an edge, only one variable is considered, and the formation of loops is prevented by adding inequality constraints. The number of sub-circles and the vertex set are determined according to the integer programming solution results.

[0044] The use of integer programming algorithm can accurately provide the optimal solution of welding path.

[0045] In the prior art, the constraints adopted for a given weighted undirected graph are:

[0046]

[0047] The above three constraints respectively mean that each node has only one outgoing arc, each node has only one incoming arc, and x ij The edge from node i to node j, x ij is an integer variable of 0 or 1, i and j both represent nodes, S represents a vertex set, and V represents a vertex set. Although integer programming is used, constraints that do not form a loop are not easy to represent.

[0048] Therefore, the present invention adopts an improved integer programming method: Specifically, for a given weighted undirected graph, the direction of the path can be ignored, that is, it is not necessary to distinguish the edge x from node i to node j. ij and edge x from node j to node i ji , only one of the variables needs to be considered, so only the case of i<j is considered. As long as there is an edge between i and j, then x ij =1(i<j). Therefore, the following relationship exists:

[0049]

[0050] Among them, i and j both represent nodes, x ij represents the edge from node i to node j, x ji represents the edge from node j to node i, w ij represents the weight of the edge, and n represents the number of nodes.

[0051] The constraints are as follows:

[0052]

[0053] x ij ∈{0,1},i<j

[0054] x ij ∈{0,1}

[0055] Among them, x ik represents the edge from node i to node k, x kj represents the edge from node k to node j.

[0056] The rewritten variable X={x ij} is reduced by half. However, the above constraints do not prevent the formation of loops. Therefore, suppose that integer programming is used to solve the vertex set containing m sub-circles. S1,S2,…,S m are all vertices, V is the vertex set, then perform the following operations in sequence until a vertex set containing only one subcircle is obtained, that is, m = 1.

[0057] Specifically, we first add m inequality constraints and solve them again using integer programming. The specific calculation formula is as follows:

[0058]

[0059] Among them, S k Represents the kth vertex in the vertex set.

[0060] Then, according to the result of resolving the integer programming, determine the number of subcircles m contained in the solution and the vertex set of m subcircles

[0061] Step S3: Use the fine geometric obstacle avoidance detection method to plan the welding path, respectively envelop the robot arm and obstacles using the capsule envelopment method and the sphere swept volume method, and use the geometric relationship between the two to complete the distance detection and further complete the obstacle avoidance process. The specific implementation process is as follows:

[0062] S3.1: Set the safe distance between the robot and obstacles to ensure that no collision occurs during the welding process;

[0063] S3.2: Use robot inverse kinematics to calculate joint angles to ensure the accessibility of each welding point of the robot;

[0064] S3.3: Determine possible collisions and handle them by detecting the geometric relationship between obstacles and robot links.

[0065] Step S4: trajectory planning;

[0066] The quintic spline interpolation function and the improved chicken swarm optimization algorithm are used for trajectory planning. The quintic spline interpolation function is mainly used for spline node interpolation on a given interval. Through integration and substitution of boundary conditions, an underdetermined system of equations containing unknown quantities and equations is obtained. Then, fixed boundary conditions are added to obtain a definite solution to the linear system of equations. The specific implementation process is as follows:

[0067] S4.1: Use the quintic spline interpolation method to smooth the planned welding trajectory to ensure the continuity and smoothness of the trajectory;

[0068] S4.2: Limit the robot's velocity, acceleration, and jerk by setting the initial velocity, initial acceleration, final velocity, and final acceleration of the robot's end effector.

[0069] S4.3: Introduce an optimization algorithm to iteratively optimize the trajectory, and calculate the joint angle condition number, global joint motion and robot flexibility indicators.

[0070] Among them, the optimization algorithm can specifically select the improved chicken swarm optimization algorithm.

[0071] In the improved chicken flock optimization algorithm, roosters, hens and chicks have different position update strategies. Roosters account for 20% of the population size, hens account for 40% of the population size, chicks account for 40% of the population size, and the chicken flock size is 50. Roosters randomly update their positions, and the update strategy is improved based on the normal distribution to enhance the search breadth. Hens usually follow roosters when foraging, and their position update strategy is constrained by the behavior of roosters and there is stealing behavior. By introducing an attraction mechanism to improve the search depth, the chicks can only follow hens, and their position update strategy is improved based on the hen's position.

[0072] Step S5: Establish a trajectory optimization model for the laser welding robot. Input finite parameters to plan the optimal smooth trajectory offline. The finite parameters mainly include the working environment axis range and the robot DH model parameters. The robot DH model parameters are as follows: the connecting rod offsets are 1.045, 0, 0, 1.025, 0, and 0, respectively; the connecting rod lengths are 0.5, 1.3, -0.055, 0, 0, and 0, respectively; and the connecting rod rotation angles are -90°, 0°, -90°, 90°, -90°, and 0°, respectively. First, the position information of the robot arm is extracted from the individual and the corresponding kinematic model is constructed; then, through a series of operations, the accessibility of the welding path and the movement sequence of the robot arm are determined, and the tasks of overlapping points are assigned using a nonlinear optimization method; a function is used to evaluate the collision risk of the robot arm during movement, and while ensuring that no collision occurs, multiple evaluation indicators are calculated, including the laser welding gun accessibility indicator, the robot flexibility indicator and the global joint motion indicator. The fitness function combines these indicators and measures the movement performance of the robot arm by setting different weight coefficients; finally, the function outputs the fitness value and a detailed result structure, which includes the movement path of each robot arm, the specific values ​​of each indicator, and the determination of the position, speed and acceleration of the robot arm at different time points.

[0073] Step S6: The robot performs welding operations according to the above optimal smooth trajectory.

[0074] In order to verify the performance of the laser welding robot trajectory planning method for electric vehicle stator welding of the present invention, the robot base coordinates and working environment axis range, robot DH model parameters and other parameters are set, the quintic spline interpolation is used to default the machine speed and acceleration, the number of individuals in the flock is 100, and the evolution generation number is 50. The test results are as follows Figures 1 to 29 shown.

[0075] Depend on Figure 1 It can be seen that the algorithm generation value iterative evolution curve of the present invention shows the change in fitness value of the CSO algorithm before and after the improvement during the evolution process. The improved CSO algorithm shows better stability and adaptability during the evolution process.

[0076] Depend on Figures 2 to 25After determining the motion parameters, a trajectory is generated in joint space using a quintic spline interpolation function. The robot completes the task in 16 seconds. The joint angular velocity curves have 0 rad / s at the initial and target nodes, and the overall trajectory speed is continuous and smooth. The joint angular acceleration curves have 0 rad / s at the initial and target nodes, and the range of the angular acceleration curves of the trajectory planned by the improved CSO algorithm meets the actual angular acceleration constraints. The initial impact values ​​of the joint angular acceleration curves are large for all six joints of the robot, and the terminal impact values ​​of joints 3 and 6 are also large. Joint 3 is primarily responsible for position transformation, while joint 6 is primarily responsible for adjusting the welding posture. Frequent adjustments to the welding posture result in greater impact on the robot. These results demonstrate that the trajectory planning method based on the improved CSO can generate continuous and smooth trajectories.

[0077] The iterative evolution curve of the robot flexibility index of the present invention is as follows: Figure 26 As shown in the figure, the robot has high flexibility and the welding posture adjustment is relatively flexible, thus completing the welding task in the area with large bending degree.

[0078] The iterative evolution curve of the laser welding gun accessibility index of the present invention is as follows: Figure 27 As shown in the figure, after 40 generations of iteration, the robot's welding clamp accessibility index converges to TA = 1, so there is no unreachable welding posture.

[0079] The iterative evolution curve of the robot's global joint motion index of the present invention is as follows: Figure 28 As shown in the figure, because the curved parts at the welding points that the robot is responsible for are more curved, it needs to adjust the welding posture more often. After the robot iterates to the 40th generation, although there is a slight increase, the total joint movement converges to the minimum value.

[0080] The iterative evolution curve of the robot installation position of the present invention is as follows: Figure 29 As shown in the figure, because the curved parts at the welding points that the robot is responsible for are more curved, it needs to adjust the welding posture more often. After the robot iterates to the 40th generation, although there is a slight increase, the total joint movement converges to the minimum value.

[0081] In a second aspect, the present invention provides a robot trajectory planning system for electric vehicle stator welding.

[0082] The present invention provides a robot trajectory planning system for electric vehicle stator welding, which is mainly used to implement a robot trajectory planning method for electric vehicle stator welding provided in the first aspect of the present invention, and specifically includes the following modules: a joint welding path planning module and a performance optimization trajectory planning module.

[0083] Among them, the joint welding path planning module uses integer programming algorithm and fine geometric obstacle avoidance detection method for path planning. By detecting the geometric relationship between obstacles and robot links, and using inverse kinematics to calculate joint angles to ensure weld accessibility, the robot can perform collision-free welding tasks in the optimal order.

[0084] Among them, the performance optimization trajectory planning module uses the quintic spline interpolation method to smooth the trajectory, limit the speed and acceleration, calculate indicators such as the joint angle condition number, global joint motion and robot flexibility, and then obtains a smooth laser welding robot trajectory through the cost function and collision detection algorithm.

[0085] In summary, the robot trajectory planning method and system for electric vehicle stator welding provided by the present invention can effectively suppress the impact of the robot working system, obtain an obstacle avoidance trajectory with optimal time and continuous acceleration, and can provide an effective reference for robot trajectory planning. The various models, parameters and robot operation data in the present invention have certain reference value for the robot trajectory planning method.

[0086] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A robot trajectory planning method for electric vehicle stator welding, characterized in that: The following steps are involved: Step S1: robot initialization; Step S2: finding the welding path using an integer programming algorithm to ensure that the robot performs the welding tasks in the optimal order; Step S3: Plan the welding path using a fine geometric obstacle avoidance detection method, envelop the robot arm and obstacles using capsule envelopment and sphere swept volume method respectively, and use the geometric relationship between the two to complete distance detection and complete obstacle avoidance processing; Step S4: trajectory planning; S4.1: Use the quintic spline interpolation method to smooth the planned welding trajectory to ensure the continuity and smoothness of the trajectory; S4.2: Limit the robot's speed, acceleration, and jerk by setting the robot's end effector's initial speed, initial acceleration, final speed, and final acceleration; S4.3: Introduce optimization algorithms to iteratively optimize the trajectory and calculate the joint angle condition number, global joint motion, and robot flexibility index; Step S5: Establish a robot trajectory optimization model, input finite parameters, and plan the optimal smooth trajectory offline; Step S6: The robot performs welding operations according to the optimal smooth trajectory.

2. A robot trajectory planning method for electric vehicle stator welding according to claim 1, characterized in that: In step S2, when the integer programming algorithm is used for undirected graph path planning, for two nodes with edges, only one variable is considered, and loops are prevented by adding inequality constraints, and the number of subcircles and vertex sets are determined according to the integer programming solution results.

3. A robot trajectory planning method for electric vehicle stator welding according to claim 1, characterized in that: In step S2, for a given weighted undirected graph, the direction of the path is not considered, that is, there is no need to distinguish the edge x from node i to node j. ij and edge x from node j to node i ji , only one of the variables is considered, so only the case of i<j is considered. As long as there is an edge between i and j, then x ij =1(i<j), so there is the following relationship: Among them, i and j both represent nodes, x ij represents the edge from node i to node j, x ji represents the edge from node j to node i, w ij represents the weight of the edge, and n represents the number of nodes; The constraints are as follows: x ij ∈{0,1},i<j x ij ∈{0,1} Among them, x ik represents the edge from node i to node k, x kj represents the edge from node k to node j; The rewritten variable X = {x ij } has been reduced by half, but the above constraints do not prevent the formation of loops; therefore, suppose that integer programming is used to solve the vertex set containing m subcircles. S1,S2,…,S m are all vertices, V is the vertex set, and through calculation, we get the vertex set containing only one subcircle, that is, m=1.

4. A robot trajectory planning method for electric vehicle stator welding according to claim 3, characterized in that: The specific implementation process of obtaining a vertex set containing only one subcircle, that is, m=1, is as follows: First, add m inequality constraints and solve them again using integer programming. The specific calculation formula is as follows: Among them, S k represents the kth vertex in the vertex set; Then, according to the result of the integer programming re-solution, determine the number of subcircles m contained in the solution and the vertex set of m subcircles 5. The robot trajectory planning method for electric vehicle stator welding according to claim 1 is characterized in that: The specific implementation process of step S3 is as follows: S3.1: Set the safe distance between the robot and obstacles to ensure that no collision occurs during welding; S3.2: Use robot inverse kinematics to calculate joint angles to ensure the accessibility of each welding point of the robot; S3.3: By detecting the geometric relationship between the obstacle and the robot link, determine possible collisions and handle them.

6. A robot trajectory planning method for electric vehicle stator welding according to claim 1, characterized in that: The quintic spline interpolation function is used for spline node interpolation on a given interval, and an underdetermined system of equations containing unknown quantities and equations is obtained by integration and substitution of boundary conditions, and then a definite solution to the linear system of equations is obtained by supplementing fixed boundary conditions.

7. A robot trajectory planning method for electric vehicle stator welding according to claim 1, characterized in that: The optimization algorithm is an improved chicken flock optimization algorithm. In the improved chicken flock optimization algorithm, roosters, hens and chicks have different position update strategies. Roosters account for 20% of the population size, hens account for 40% of the population size, chicks account for 40% of the population size, and the chicken flock size is 50. Roosters randomly update their positions, and the update strategy is improved based on the normal distribution to improve the search breadth. Hens usually follow roosters when foraging, and their position update strategy is constrained by the behavior of roosters and there is stealing behavior. The search depth is improved by introducing an attraction mechanism. Chicks can only follow hens, and their position update strategy is improved based on the position of the hens.

8. The robot trajectory planning method for electric vehicle stator welding according to claim 1, characterized in that: The finite parameters include: the working environment axis range and the robot DH model parameters, and the robot DH model parameters are: the connecting rod offsets are 1.045, 0, 0, 1.025, 0 and 0 respectively, the connecting rod lengths are 0.5, 1.3, -0.055, 0, 0 and 0 respectively, and the connecting rod rotation angles are -90°, 0°, -90°, 90°, -90° and 0° respectively.

9. The robot trajectory planning method for electric vehicle stator welding according to claim 1, characterized in that: The specific implementation process of step S5 is as follows: First, the position information of the robot arm is extracted from the individuals and the corresponding kinematic model is constructed. Then, the accessibility of the welding path and the movement sequence of the robot arm are determined through a series of operations, and the tasks of the overlapping points are allocated using a nonlinear optimization method. The function is used to evaluate the collision risk of the robot during movement, and multiple evaluation indicators are calculated while ensuring that no collision occurs, including the laser welding gun accessibility indicator, the robot flexibility indicator and the global joint motion indicator. The fitness function combines these indicators and measures the movement performance of the robot by setting different weight coefficients. Finally, the function outputs the fitness value and a detailed result structure, which includes the movement path of each robot, the specific values ​​of each indicator, and the determination of the position, speed and acceleration of the robot at different time points.

10. A robot trajectory planning system for electric vehicle stator welding, characterized in that: include: Joint welding path planning module and performance optimization trajectory planning module; The joint welding path planning module adopts integer programming algorithm and fine geometric obstacle avoidance detection method for path planning. By detecting the geometric relationship between obstacles and robot connecting rods and calculating the joint angles with the help of inverse kinematics, the robot can ensure the accessibility of welding points and perform collision-free welding tasks in the optimal order. The performance optimization trajectory planning module uses quintic spline interpolation to smooth the trajectory, limit the speed and acceleration, calculate the joint angle condition number, global joint motion and robot flexibility index, and then obtains a smooth laser welding robot trajectory through a cost function and a collision detection algorithm.

Images