A state feedback based robot arm obstacle avoidance method and system

Abstract

The application relates to the technical field of automation equipment, and discloses a mechanical arm obstacle avoidance method and system based on state feedback, which is used for solving the problem that when there are multiple obstacles with the same distance from a mechanical arm, the mechanical arm cannot accurately select the best priority obstacle to avoid, comprising the following steps: measuring the obstacles around the mechanical arm to obtain the volume and shape of the obstacles, analyzing to obtain an obstacle shape hindering coefficient, obtaining the moving data of the obstacles, evaluating to obtain an obstacle moving influence coefficient, comprehensively analyzing to obtain a priority avoidance index, selecting the obstacle to be avoided in priority according to the priority avoidance index, adjusting the obstacle avoidance distance to obtain an actual obstacle avoidance distance, generating a path to avoid collision with the obstacle when the distance between the mechanical arm and the obstacle reaches the actual obstacle avoidance distance, continuously detecting new obstacle information changes, and adjusting the path in real time to effectively obtain the best obstacle avoidance route and improve work efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of automation equipment, and more particularly to a method and system for avoiding obstacles of a robotic arm based on state feedback. Background Art

[0002] Robotic arm obstacle avoidance is a technology that uses sensors to detect obstacles in the environment and adjusts the robot's trajectory through path planning or control algorithms to avoid collisions. It generates a safe path based on the robot's current target task, ensuring the robot can complete its work efficiently and safely in complex or dynamic environments.

[0003] For example, the robotic arm obstacle avoidance method and robotic arm obstacle avoidance system disclosed in the invention patent announcement with the publication number CN115723121A include: a modeling step, a step of collecting and evaluating coordinates, a step of obtaining control variables, a step of establishing an occupancy function, and a step of finding an obstacle avoidance posture. The present invention pre-stores the data obtained from executing the modeling step, the step of establishing evaluation coordinates, the step of obtaining control variables, and the step of establishing an occupancy function in a database, thereby allowing the subsequent robotic arm to quickly assess whether a collision will occur when performing a task. If the assessment indicates that a collision will occur, the step of finding an obstacle avoidance posture is executed to avoid the obstacle. The present invention adopts a non-contact anti-collision design, which can improve the shortcomings faced by existing contact-type anti-collision designs.

[0004] For example, the dual neural network obstacle avoidance control method for an omnidirectional mobile manipulator, disclosed in the invention patent publication number CN117182912A, includes: designing the desired velocity-level motion trajectory of the omnidirectional mobile manipulator based on the designer's requirements; proposing a hybrid trajectory tracking and obstacle avoidance scheme for the omnidirectional mobile manipulator with joint and joint velocity constraints based on the kinematic model of the omnidirectional mobile manipulator; and designing a primal dual neural network based on linear variational inequalities. This method addresses the problem of ensuring smooth and accurate tracking of an omnidirectional mobile manipulator in the presence of static or moving obstacles, and verifies the effectiveness of the algorithm through simulation experiments.

[0005] However, in the process of implementing the technical solutions of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0006] In the existing technology, robotic arms generally avoid obstacles by detecting the distance between the robotic arm and the obstacle and performing obstacle avoidance processing based on the distance. However, if there are two or more obstacles at the same distance from the robotic arm, the robotic arm cannot accurately select the best obstacle to avoid first, resulting in the inability to generate the best obstacle avoidance route, thereby reducing work efficiency. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for robot arm obstacle avoidance based on state feedback to solve the problems existing in the above-mentioned background technology.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A method for avoiding obstacles of a manipulator based on state feedback comprises the following steps: Step 1: When the manipulator is in operation, a laser radar is used to measure obstacles around the manipulator to obtain the volume and shape of the obstacles, and an obstacle shape obstruction coefficient is obtained according to the volume and shape of the obstacles; Step 2: The laser radar is used to obtain the movement data of the obstacle, the movement data including the movement speed and the movement direction, and the obstacle movement influence coefficient is obtained according to the movement data evaluation; Step 3: The method for obtaining a priority avoidance index according to the obstacle shape obstruction coefficient and the obstacle movement influence coefficient is PA=a1×SO+a2×MI, where PA represents the priority avoidance index and SO represents the obstacle shape shape obstruction coefficient, MI represents the obstacle movement influence coefficient, a1 and a2 represent the weight coefficients of the obstacle shape obstruction coefficient and the obstacle movement influence coefficient; Step 4: Select the obstacle to be avoided first according to the priority avoidance index; Step 5: Adjust the obstacle avoidance distance according to the priority avoidance index of the obstacle to be avoided first, and obtain the actual obstacle avoidance distance; Step 6: When the robot arm reaches the actual obstacle avoidance distance from the obstacle, use the path planning algorithm according to the obstacle to be avoided first to generate a path to avoid collision with the obstacle; Step 7: During the movement of the robot arm to avoid obstacles according to the path, continuously detect new obstacle information changes and adjust the path in real time.

[0010] Preferably, the steps of using a laser radar to measure obstacles around the robotic arm and obtaining the volume and shape of the obstacles are as follows: the laser radar generates point cloud data of the obstacle by emitting laser pulses at the measurement points and measuring the time difference of the return signals; performing multi-point measurement of the obstacle using the laser radar to obtain obstacle point cloud data; connecting each point in the obstacle point cloud data to reconstruct the three-dimensional shape of the obstacle and obtain the volume of the obstacle; dividing the three-dimensional shape of the obstacle into n three-dimensional spaces on average, each of which is an equal-sized cube, recorded as obstacle subspaces, screening out obstacle subspaces with surfaces in contact with the outside world, recorded as effective obstacle subspaces; calculating the spatial irregularity of each effective obstacle subspace, clustering the spatial irregularity of each effective obstacle subspace using the K-means clustering method, and calculating the obstacle irregularity based on the clustering results.

[0011] Preferably, the step of calculating the spatial irregularity of each effective obstacle subspace is as follows: obtaining the actual volume of the obstacle in each effective obstacle subspace, obtaining the surface area of ​​the obstacle in each effective obstacle subspace that is actually in contact with the outside world, and obtaining the body surface ratio according to the actual volume and the actual surface area in contact with the outside world; obtaining the sphericity according to the actual volume and the actual surface area in contact with the outside world; and obtaining the spatial irregularity according to the body surface ratio and the sphericity. Where SI represents spatial irregularity, rt represents body surface area ratio, and sy represents sphericity.

[0012] Preferably, the steps of clustering the spatial irregularity of each effective obstacle subspace using the K-means clustering method are as follows: Step 1.1: Using spatial irregularity as a clustering feature, using the spatial irregularities of all effective obstacle subspaces as a data set, and using the spatial irregularities in the data set as data points; Step 1.2: Using the silhouette coefficient method to determine the optimal number of clusters K of the data set; Step 1.3: Randomly selecting K data points in the data set as initial cluster centers, for each data point, calculating its Euclidean distance to each initial cluster center, for each data point, traversing the K initial cluster centers, and assigning it to the cluster corresponding to the nearest initial cluster center; Step 1.4: After traversing all data points, obtaining initial cluster clusters, for each initial cluster cluster, performing mean calculation on the data points within it to obtain a new cluster center; Step 1.5: Repeating steps 1.3 and 1.4 until the cluster center no longer changes, obtaining the final cluster cluster and the final cluster center.

[0013] Preferably, the step of calculating the obstacle irregularity based on the clustering results is: calculating the ratio of the number of data points in each final cluster to the total number in the data set to obtain the weight of each final cluster; and performing weighted summation of the weight of each final cluster and the final cluster center to obtain the obstacle irregularity.

[0014] Preferably, the step of obtaining the obstacle shape obstruction coefficient according to the volume and shape of the obstacle is:

[0015] The method to obtain the obstacle shape obstruction coefficient based on the obstacle volume and obstacle irregularity is: Where SO represents the obstacle shape obstruction coefficient, IO represents the obstacle irregularity, and Ov represents the obstacle volume.

[0016] Preferably, the step of obtaining the obstacle movement influence coefficient based on the movement data evaluation is as follows: using a laser radar to obtain the movement speed of the obstacle; using the laser radar to obtain the angle between the relative movement direction of the obstacle and the movement direction of the robot arm; the method of obtaining the obstacle movement influence coefficient based on the movement speed of the obstacle and the angle between the relative movement direction of the obstacle and the movement direction of the robot arm is MI=V o ×cos(θ), where MI is the obstacle movement influence coefficient, V o It is expressed as the moving speed of the obstacle, and θ is the angle between the relative movement direction of the obstacle and the movement direction of the robotic arm.

[0017] Preferably, the step of selecting an obstacle to be avoided first according to the priority avoidance index is: comparing the priority avoidance indexes of the obstacles, and selecting the obstacle with the largest priority avoidance index as the obstacle to be avoided first.

[0018] Preferably, the step of adjusting the obstacle avoidance distance according to the priority avoidance index of the obstacle to be avoided is: setting the initial obstacle avoidance distance and the priority avoidance index threshold, and obtaining the actual obstacle avoidance distance according to the priority avoidance index and the initial obstacle avoidance distance. Among them AD 实际 Indicated as the actual obstacle avoidance distance, AD 初始 It represents the initial obstacle avoidance distance, PA represents the priority avoidance index, and PA0 represents the priority avoidance index threshold.

[0019] Preferably, a robotic arm obstacle avoidance system based on state feedback includes: an obstacle data acquisition module, which is used to collect obstacle information and transmit the obstacle information to an obstacle selection module; an obstacle selection module, which is used to receive the obstacle information transmitted by the obstacle data acquisition module, analyze the obstacle information, select obstacles to be avoided first based on the analysis results, and transmit the selected obstacles to be avoided first to an obstacle avoidance distance adjustment module; an obstacle avoidance distance adjustment module, which is used to receive the obstacles to be avoided first transmitted by the obstacle selection module, adjust the obstacle avoidance distance based on the obstacle information to obtain an actual obstacle avoidance distance, and transmit the obtained distance to an obstacle avoidance module; and an obstacle avoidance module, which is used to receive the actual obstacle avoidance distance transmitted by the obstacle avoidance distance adjustment module, and perform obstacle avoidance processing based on the actual obstacle avoidance distance.

[0020] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0021] 1. While the robotic arm is in motion, a lidar system is used to measure obstacles around the robotic arm, determining their volume and shape. This information is then used to determine the obstacle's shape obstruction coefficient. By determining the obstacle's volume and shape, the robotic arm can more accurately determine the actual degree of obstruction to its movement. This is particularly true for complex, irregularly shaped obstacles, effectively preventing incorrect obstacle avoidance or path planning failures caused by shape misjudgment. The real-time calculated shape obstruction coefficient reflects the impact of obstacles at different locations on the robotic arm's path, enabling the robotic arm to prioritize avoiding high-impedance obstacles in dynamic environments and improving the flexibility of its obstacle avoidance strategy.

[0022] 2. Use LiDAR to obtain obstacle movement data, including speed and direction, and evaluate the obstacle's movement impact coefficient based on this data. Based on the obstacle's speed and direction, the robot arm can dynamically predict the obstacle's future position and react in advance, effectively reducing the risk of collision. This is especially true in high-speed or complex motion environments. The movement impact coefficient helps the robot arm identify which obstacles have the greatest impact on its path, thereby optimizing its path selection.

[0023] 3. Prioritize obstacles based on the avoidance priority index. This index takes into account factors such as obstacle size, distance, shape, and speed, ensuring that the robot arm prioritizes avoiding obstacles that pose the greatest threat to its movement. This allows for more rational obstacle avoidance decisions, avoiding repeated avoidance attempts or wasted paths due to unclear judgments.

[0024] 4. The obstacle avoidance distance is adjusted based on the priority avoidance index of the selected obstacles to determine the actual obstacle avoidance distance. The actual obstacle avoidance distance is dynamically adjusted based on the priority avoidance index, allowing the robot arm to select the appropriate obstacle avoidance distance in different situations. For high-risk obstacles, increasing the obstacle avoidance distance provides a greater safety buffer; for low-risk obstacles, the obstacle avoidance distance can be reduced to ensure that space and time are not wasted. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A flow chart of a method for avoiding obstacles by a robotic arm based on state feedback is provided in an embodiment of the present application.

[0026] Figure 2 A structural diagram of a robotic arm obstacle avoidance system based on state feedback provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. In addition, the forms of the various structures described in the following embodiments are merely examples. The state feedback-based robotic arm obstacle avoidance method and system involved in the present invention are not limited to the various structures described in the following embodiments. All other implementations obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0028] The present invention provides a method for avoiding obstacles of a robotic arm based on state feedback, comprising the following steps:

[0029] When the robot arm detects two or more obstacles in the running state, and the distance between them is the same,

[0030] Step 1: When the robot arm is in operation, use the LiDAR to measure the obstacles around the robot arm to obtain the volume and shape of the obstacles. According to the volume and shape of the obstacles, the obstacle shape obstruction coefficient is obtained;

[0031] LiDAR is a sensor that measures the distance and shape of objects by emitting laser beams and receiving reflected signals. It uses a rapidly rotating laser emitter to scan the surrounding environment and calculates the position, size, and shape of obstacles using time differences and the speed of light. LiDAR offers high precision and resolution, generating detailed 3D point cloud data in complex environments. Therefore, it is commonly used for environmental perception and obstacle detection in robotic arms for obstacle avoidance.

[0032] In this embodiment, it should be specifically explained that the steps for using a laser radar to measure obstacles around the robotic arm and obtain the volume and shape of the obstacles are as follows:

[0033] LiDAR generates point cloud data of obstacles by emitting laser pulses to the measurement point and measuring the time difference of the return signal;

[0034] Use LiDAR to measure multiple points of obstacles and obtain obstacle point cloud data;

[0035] Connect each point in the obstacle point cloud data to reconstruct the three-dimensional shape of the obstacle and obtain the volume of the obstacle;

[0036] Divide the three-dimensional shape of the obstacle into n three-dimensional spaces, each of which is an equal-sized cube, recorded as an obstacle subspace. Filter out the obstacle subspace with a surface in contact with the outside world, and record it as an effective obstacle subspace.

[0037] The spatial irregularity of each effective obstacle subspace is calculated, and the spatial irregularity of each effective obstacle subspace is clustered using the K-means clustering method. The obstacle irregularity is calculated based on the clustering results.

[0038] By filtering out the obstacle subspaces that have surfaces in contact with the outside world, we can focus only on areas with strong obstacle avoidance relevance. When avoiding obstacles, the robot arm mainly needs to focus on external surfaces that may collide with it. Even if the shape of the internal subregion is complex, it will not affect the arm's motion path. Therefore, calculating the spatial irregularity of only the exposed areas can more accurately reflect the geometric features that have a direct impact on the obstacle avoidance task; if the spatial irregularity of all subregions, especially those in the internal areas, is calculated, it may increase unnecessary computational burden. Calculating only the exposed areas can significantly reduce the amount of calculation and improve the real-time and response speed of the system; because the internal subregions are not exposed to the robot arm's sensors (such as lidar or cameras), calculating their shape irregularity has no practical significance in obstacle avoidance scenarios. Therefore, by focusing on calculating the shape irregularity of the exposed areas, the analysis can be more focused on information that is useful for the task.

[0039] In this embodiment, it should be specifically explained that the steps for calculating the spatial irregularity of each effective obstacle subspace are:

[0040] Obtain the actual volume of the obstacle in each effective obstacle subspace, obtain the actual surface area of ​​the obstacle in each effective obstacle subspace that is in contact with the outside world, and obtain the body surface ratio based on the actual volume and the actual surface area in contact with the outside world. Where rt is the volume-to-surface ratio, ve is the actual volume, and sa is the surface area actually in contact with the outside world. A higher volume-to-surface ratio usually indicates a more regular shape, while a lower ratio indicates an irregular shape.

[0041] The method to obtain sphericity based on the actual volume and the actual surface area in contact with the outside world is: Where sy represents the sphericity. A sphericity close to 1 indicates that the shape is close to a sphere, and a smaller value indicates a more irregular shape.

[0042] The method to obtain spatial irregularity based on body surface ratio and sphericity is: Where SI represents spatial irregularity, rt represents body surface area ratio, and sy represents sphericity.

[0043] In this embodiment, it should be specifically explained that the steps of clustering the spatial irregularity of each effective obstacle subspace using the K-means clustering method are as follows:

[0044] Step 1.1: Use spatial irregularity as a clustering feature, and use the spatial irregularity of all valid obstacle subspaces as a data set, with the spatial irregularity in the data set as data points;

[0045] Step 1.2: Use the silhouette coefficient method to determine the optimal number of clusters K for the data set;

[0046] The silhouette coefficient method is a method for evaluating clustering quality and determining the optimal number of clusters K. It evaluates the compactness of the point within its cluster and the degree of separation from the nearest neighboring cluster by calculating the silhouette coefficient of each sample point. The value range of the silhouette coefficient is from -1 to 1. The higher the value, the better the clustering effect. When the value is close to 1, it means that the point has a high similarity with other points in the cluster and is far away from other clusters; when it is close to 0, it means that the point is at the boundary of two clusters and may not be suitable for the current cluster; and when it is close to -1, it means that the point may be incorrectly assigned to the current cluster. The key to the silhouette coefficient method is to calculate the average silhouette coefficient of all points under different K, and select the K with the highest average value as the optimal number of clusters. This method not only takes into account the compactness within the cluster, but also pays attention to the separation between clusters. Therefore, it can comprehensively measure the clustering effect, and is especially suitable for evaluating the internal consistency of clusters and the differences between clusters.

[0047] Step 1.3: Randomly select K data points in the data set as the initial cluster centers. For each data point, calculate the Euclidean distance from each initial cluster center. The calculation formula is: Where d(P,α) represents the Euclidean distance from the data point to the cluster center, where P represents the data point and α represents the initial cluster center. For each data point, traverse the K initial cluster centers and assign it to the cluster corresponding to the nearest initial cluster center.

[0048] Euclidean distance is the most intuitive and commonly used distance metric between two points. It originates from Euclidean geometry and represents the "straight-line distance" between two points in space. In the K-means clustering algorithm, Euclidean distance is used to calculate the distance between each data point and multiple initial cluster centers, thereby determining which cluster center each point is closest to, thereby completing cluster assignment.

[0049] Step 1.4: After traversing all data points, we get the initial clusters. For each initial cluster, we calculate the mean of the data points in it and get the new cluster center.

[0050] Step 1.5: Repeat steps 1.3 and 1.4 until the cluster center no longer changes, and obtain the final cluster and the final cluster center.

[0051] In this embodiment, it should be specifically explained that the steps for calculating the obstacle irregularity according to the clustering results are:

[0052] The weight of each final cluster is calculated by calculating the ratio of the number of data points in each final cluster to the total number of data points in the data set;

[0053] The weight of each final cluster is weighted and summed with the final cluster center to obtain the obstacle irregularity, which is calculated as follows: Where IO represents the irregularity of the obstacle, Q j Expressed as the weight of the jth final cluster, It is represented as the jth final cluster center, and K is the optimal number of clusters of the data set.

[0054] In this embodiment, it should be specifically explained that the steps for obtaining the obstacle shape obstruction coefficient based on the volume and shape of the obstacle are as follows:

[0055] The method to obtain the obstacle shape obstruction coefficient based on the obstacle volume and obstacle irregularity is: Where SO represents the obstacle shape obstruction coefficient, IO represents the obstacle irregularity, and Qv represents the obstacle volume.

[0056] Step 2: Use LiDAR to obtain obstacle movement data, including movement speed and direction, and evaluate the obstacle movement influence coefficient based on the movement data.

[0057] In this embodiment, it should be specifically explained that the steps for evaluating and obtaining the obstacle movement influence coefficient based on the movement data are as follows:

[0058] Use LiDAR to obtain the moving speed of obstacles;

[0059] Use LiDAR to obtain the angle between the relative movement direction of the obstacle and the movement direction of the robotic arm;

[0060] The method to obtain the obstacle movement influence coefficient is MI=V according to the obstacle movement speed and the angle between the obstacle's relative movement direction and the robot arm's movement direction. o ×cos(θ), where MI is the obstacle movement influence coefficient, V o It is expressed as the moving speed of the obstacle, and θ is the angle between the relative movement direction of the obstacle and the movement direction of the robotic arm.

[0061] By calculating the angle θ between the obstacle's relative motion direction and the robot's motion direction, we can determine whether the obstacle is approaching the robot's path. The smaller the angle θ (i.e., the closer the obstacle's motion direction is to the robot's motion direction), the greater the threat the obstacle poses to the robot. When the obstacle's motion direction aligns with the robot's motion direction, the impact coefficient reaches its maximum value. This means that the robot should prioritize avoiding these more threatening obstacles.

[0062] Step 3: The priority avoidance index is calculated based on the obstacle shape obstruction coefficient and the obstacle motion influence coefficient: PA = a1 × SO + a2 × MI, where PA represents the priority avoidance index and SO represents the obstacle shape obstruction coefficient. The more complex the obstacle shape and the greater the interference with the robot arm's movement, the more the system tends to prioritize avoiding it. In other words, complex or bulky obstacles, because they occupy more space, require longer detours, and are more difficult to avoid, are given a higher avoidance priority by the system. This ensures that the robot arm avoids these obstacles first during movement, reducing potential collision risk and path planning difficulties. This proportional relationship enables the obstacle avoidance system to more intelligently navigate complex environments and select the safest and most effective avoidance path. MI represents the obstacle motion influence coefficient. As the obstacle's speed, direction, and uncertainty increase, the system increases its priority avoidance level. In other words, obstacles with fast movement, complex trajectories, or unpredictable directions pose a greater threat to the robot arm and are more difficult to avoid. Therefore, the system tends to prioritize avoiding these dynamic obstacles. This proportional relationship ensures the robotic arm can promptly respond to ever-changing dynamic obstacles in the environment, preventing them from entering the robotic arm's working path and reducing the risk of collisions. By dynamically adjusting the avoidance priority, the system can more flexibly and safely handle complex mobile environments and avoid potential conflicts with fast-moving objects. a1 and a2 represent the weighting coefficients of the obstacle shape obstruction coefficient and the obstacle movement influence coefficient, with a1 + a2 = 1. The specific values ​​of a1 and a2 are determined by professionals based on actual conditions. For example, a1 and a2 can be 0.4 and 0.6.

[0063] Step 4: Select the obstacle to be avoided first according to the priority avoidance index;

[0064] In this embodiment, it should be specifically explained that the steps for selecting an obstacle to be avoided first according to the priority avoidance index are:

[0065] The priority avoidance indexes of the obstacles are compared, and the obstacle with the largest priority avoidance index is selected as the obstacle to be avoided first.

[0066] Step 5: Adjust the obstacle avoidance distance based on the priority avoidance index of the obstacles selected for priority avoidance;

[0067] In this embodiment, it should be specifically explained that the steps for adjusting the obstacle avoidance distance according to the priority avoidance index of the obstacle selected for priority avoidance are as follows:

[0068] Set the initial obstacle avoidance distance, and the method to obtain the actual obstacle avoidance distance based on the priority avoidance index and the initial obstacle avoidance distance is: Among them AD 实际Indicated as the actual obstacle avoidance distance, AD 初始 It represents the initial obstacle avoidance distance, PA represents the priority avoidance index, and PA0 represents the priority avoidance index threshold.

[0069] The priority avoidance index takes into account various factors of obstacles (such as size, shape, speed, and dynamic characteristics). By dynamically adjusting the actual obstacle avoidance distance, the system can more flexibly adapt to different environments and obstacle types. When a detected obstacle has a high priority avoidance index (such as a large or fast-moving obstacle), the system can automatically increase the actual obstacle avoidance distance to ensure sufficient space for avoidance. For smaller or stationary obstacles, the actual obstacle avoidance distance can be reduced when the priority avoidance index is low, improving the robot's movement efficiency.

[0070] In actual applications, different obstacles have different degrees of impact on the path. For low-priority obstacles, the system can choose a shorter detour; for high-priority obstacles, the system will maintain a larger obstacle avoidance distance to ensure safety while avoiding complex multiple detours. By dynamically adjusting the actual obstacle avoidance distance, the system can respond more effectively to obstacles, especially when dealing with fast or sudden obstacles, providing a longer lead time to avoid accidents caused by delayed judgment. By calculating the actual obstacle avoidance distance, the system can flexibly adjust according to the actual threat level of the obstacle, rather than uniformly adopting emergency stops or large detours. This can maintain the continuous movement of the robot arm, improve operational efficiency, and reduce production delays caused by frequent pauses.

[0071] Step 6: When the robot arm reaches the actual obstacle avoidance distance from the obstacle, it uses the path planning algorithm to generate a path that avoids collision with the obstacle based on the obstacles that are screened out for priority avoidance. The path planning algorithm can use algorithm A.

[0072] In this embodiment, it should be specifically explained that the steps of using the path planning algorithm to generate a path that avoids collision with obstacles are as follows:

[0073] Model the work area as a 3D grid map and mark obstacles in the map;

[0074] Use LiDAR to detect obstacles in real time, obtain obstacle information, mark obstacles on the environment map, and update the map to make the area where the obstacle is located an impassable area;

[0075] Use A algorithm for path search, which is a heuristic search algorithm used to find the shortest path from a starting point to an end point in a graph or grid;

[0076] Use the B-spline curve algorithm to smooth the path and make it smoother. The B-spline curve algorithm is a mathematical method for generating smooth curves and is widely used in computer graphics and path planning. It defines a piecewise polynomial curve through multiple control points to ensure the smoothness and continuity of the curve. The B-spline curve has the characteristic of local control, that is, each control point only affects the shape of the curve within a specific range. Therefore, when a control point is adjusted, the overall curve will not be affected. It can generate a smooth and natural path by adjusting the control points and the order of the curve (usually third order), which is very suitable for smoothing complex paths or trajectories.

[0077] Check whether the points or line segments on the path overlap with obstacles. Use the collision detection algorithm to determine whether each node or line segment in the path intersects with the obstacle. If so, the path segment is considered impassable.

[0078] When a new obstacle is detected or an existing path is blocked by an obstacle, an incremental path planning algorithm is used to recalculate the affected path portion instead of planning from scratch.

[0079] Step 7: During the obstacle avoidance process, the robot arm continuously detects new obstacle information or changes in the environment, and adjusts the path in real time to ensure the effectiveness of obstacle avoidance.

[0080] In this embodiment, it is necessary to specifically explain a robotic arm obstacle avoidance system based on state feedback, the system comprising:

[0081] The obstacle data acquisition module is used to collect obstacle information and transmit the obstacle information to the obstacle selection module;

[0082] The obstacle selection module is used to receive the obstacle information transmitted by the obstacle data acquisition module, analyze the obstacle information, select obstacles to be avoided first based on the analysis results, and transmit the selected obstacles to be avoided first to the obstacle avoidance distance adjustment module;

[0083] The obstacle avoidance distance adjustment module is used to receive the priority obstacles transmitted by the obstacle selection module, adjust the obstacle avoidance distance according to the information of the priority obstacles, obtain the actual obstacle avoidance distance, and transmit it to the obstacle avoidance module;

[0084] The obstacle avoidance module is used to receive the actual obstacle avoidance distance transmitted by the obstacle avoidance distance adjustment module and perform obstacle avoidance processing according to the actual obstacle avoidance distance.

[0085] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0086] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A robot arm obstacle avoidance method based on state feedback, characterized in that: The following steps are involved: Step 1: When the robot arm is in operation, use the LiDAR to measure the obstacles around the robot arm to obtain the volume and shape of the obstacles. The spatial irregularity is obtained based on the volume and shape of the obstacles, and the obstacle irregularity is calculated based on the clustering results, thereby further obtaining the obstacle shape obstruction coefficient. Step 2: Use LiDAR to obtain obstacle movement data, including movement speed and direction, and evaluate the obstacle movement influence coefficient based on the movement data. Step 3: Obtain the priority avoidance index PA based on the obstacle shape obstruction coefficient and the obstacle movement influence coefficient. , where PA represents the priority avoidance index, SO represents the obstacle shape obstruction coefficient, and MI represents the obstacle movement influence coefficient. 、 It is expressed as the weight coefficient of obstacle shape obstruction coefficient and obstacle movement influence coefficient; Step 4: Select the obstacle to be avoided first according to the priority avoidance index; Step 5: Adjust the obstacle avoidance distance based on the priority avoidance index of the obstacles selected to be avoided first, and obtain the actual obstacle avoidance distance; Step 6: When the robot arm reaches the actual obstacle avoidance distance, it uses the path planning algorithm to generate a path that avoids collisions with the obstacles based on the obstacles that are prioritized for avoidance. Step 7: While the robot arm is avoiding obstacles along the path, it continuously detects new changes in obstacle information and adjusts the path in real time. The use of the LiDAR to measure obstacles around the robot arm and obtain the volume and shape of the obstacles includes: LiDAR generates point cloud data of obstacles by emitting laser pulses to the measurement point and measuring the time difference of the return signal; Use LiDAR to measure multiple points of obstacles and obtain obstacle point cloud data; Connect each point in the obstacle point cloud data to reconstruct the three-dimensional shape of the obstacle and obtain the volume of the obstacle; Divide the three-dimensional shape of the obstacle into n three-dimensional spaces, each of which is an equal-sized cube, recorded as an obstacle subspace. Filter out the obstacle subspace with a surface in contact with the outside world, and record it as an effective obstacle subspace. Calculate the spatial irregularity of each effective obstacle subspace, cluster the spatial irregularity of each effective obstacle subspace using the K-means clustering method, and calculate the obstacle irregularity based on the clustering results; Calculating the spatial irregularity of each effective obstacle subspace includes: Obtain the actual volume of the obstacle in each valid obstacle subspace, obtain the actual surface area of ​​the obstacle in contact with the outside world in each valid obstacle subspace, and obtain the body surface ratio based on the actual volume and the actual surface area in contact with the outside world; The sphericity is obtained based on the actual volume and the actual surface area in contact with the outside world; According to the body surface ratio and sphericity, the spatial irregularity SI is obtained. ,in Expressed as spatial irregularity, Expressed as body surface ratio, Expressed as sphericity.

2. The method for avoiding obstacles of a robotic arm based on state feedback according to claim 1, characterized in that: The method of clustering the spatial irregularity of each effective obstacle subspace using the K-means clustering method includes: Step 1.1: Use spatial irregularity as a clustering feature, and use the spatial irregularity of all valid obstacle subspaces as a data set, with the spatial irregularity in the data set as data points; Step 1.2: Use the silhouette coefficient method to determine the optimal number of clusters K for the data set; Step 1.3: Randomly select K data points in the data set as initial cluster centers. For each data point, calculate its Euclidean distance to each initial cluster center. For each data point, traverse the K initial cluster centers and assign it to the cluster corresponding to the initial cluster center closest to it. Step 1.4: After traversing all data points, we get the initial clusters. For each initial cluster, we calculate the mean of the data points in it and get the new cluster center. Step 1.5: Repeat steps 1.3 and 1.4 until the cluster center no longer changes, and obtain the final cluster and the final cluster center.

3. The method for avoiding obstacles by a robotic arm based on state feedback according to claim 1, characterized in that: The obstacle irregularity calculated according to the clustering results includes: The weight of each final cluster is calculated by calculating the ratio of the number of data points in each final cluster to the total number of data points in the data set; The obstacle irregularity is obtained by taking the weighted sum of each final cluster and the final cluster center.

4. The method for avoiding obstacles by a robotic arm based on state feedback according to claim 1, characterized in that: The obstacle shape obstruction coefficient obtained by analyzing the volume and shape of the obstacle includes: The method to obtain the obstacle shape obstruction coefficient based on the volume and irregularity of the obstacle is: ,in Expressed as the obstacle shape hindrance coefficient, Expressed as obstacle irregularity, It is expressed as the volume of the obstacle, and e is the exponent.

5. The method for avoiding obstacles by a robotic arm based on state feedback according to claim 1, characterized in that: The obstacle movement influence coefficient obtained by evaluating the movement data includes: Use LiDAR to obtain the moving speed of obstacles; Use LiDAR to obtain the angle between the relative movement direction of the obstacle and the movement direction of the robotic arm; The method to obtain the obstacle movement influence coefficient is based on the obstacle movement speed and the angle between the obstacle's relative movement direction and the robot arm's movement direction: ,in Expressed as the obstacle movement influence coefficient, It is the moving speed of the obstacle. is the angle between the relative movement direction of the obstacle and the movement direction of the robotic arm.

6. The method for avoiding obstacles by a robotic arm based on state feedback according to claim 1, characterized in that: The obstacles selected for priority avoidance according to the priority avoidance index include: The priority avoidance indexes of the obstacles are compared, and the obstacle with the largest priority avoidance index is selected as the obstacle to be avoided first.

7. The method for avoiding obstacles by a robotic arm based on state feedback according to claim 1, characterized in that: The adjustment of the obstacle avoidance distance according to the priority avoidance index of the obstacle selected for priority avoidance includes: Set the initial obstacle avoidance distance and the priority avoidance index threshold, and the method to obtain the actual obstacle avoidance distance based on the priority avoidance index and the initial obstacle avoidance distance is: ,in Indicates the actual obstacle avoidance distance, It is represented as the initial obstacle avoidance distance, Expressed as the priority avoidance index, Expressed as the priority avoidance index threshold.

8. A robotic arm obstacle avoidance system based on state feedback, implementing a robotic arm obstacle avoidance method based on state feedback according to any one of claims 1 to 7, characterized in that: include: The obstacle data acquisition module is used to collect obstacle information and transmit the obstacle information to the obstacle selection module; The obstacle selection module is used to receive the obstacle information transmitted by the obstacle data acquisition module, analyze the obstacle information, select obstacles to be avoided first based on the analysis results, and transmit the selected obstacles to be avoided first to the obstacle avoidance distance adjustment module; The obstacle avoidance distance adjustment module is used to receive the priority obstacles transmitted by the obstacle selection module, adjust the obstacle avoidance distance according to the information of the priority obstacles, obtain the actual obstacle avoidance distance, and transmit it to the obstacle avoidance module; The obstacle avoidance module is used to receive the actual obstacle avoidance distance transmitted by the obstacle avoidance distance adjustment module and perform obstacle avoidance processing according to the actual obstacle avoidance distance.

Images