A trajectory planning method for a robotic arm, a mobile robot, a medium, and a product.

Abstract

This disclosure relates to a trajectory planning method for a robotic arm, a mobile robot, a medium, and a product. The method includes: determining the initial pose of the end effector of the robotic arm in a working environment model and its target pose in a predetermined operation task; determining multiple workspaces based on the initial pose and the target pose; performing trajectory planning based on the working environment model, and during the trajectory planning process, acquiring multiple collision models with differences in geometric modeling accuracy and collision detection efficiency, and performing collision detection on trajectory points in the corresponding workspaces using the collision models; after the trajectory planning is completed, obtaining the target motion trajectory of the robotic arm under the predetermined operation task. This embodiment solves the problem of high computational complexity in global collision detection using traditional collision models, improves the trajectory planning efficiency of the robotic arm, and thus ensures the task response efficiency of the mobile robot in scenarios with dense equipment distribution.

Description

Technical Field

[0001] This disclosure relates to the field of instrument control technology, and in particular to a trajectory planning method for a robotic arm, a mobile robot, a medium, and a product. Background Technology

[0002] With the rapid development of industries such as manufacturing, finance, and artificial intelligence, the security operation and maintenance of data centers has become a key link in ensuring the stable and reliable operation of various businesses. The use of mobile robots to carry out automated inspections of data centers has gradually become the mainstream trend in the industry.

[0003] When performing inspections in data centers, mobile robots not only need to monitor equipment status, identify potential hazards, and collect data, but also often need to use their onboard robotic arms to perform various precision operations, such as opening and closing cabinet doors, operating circuit breakers, touching equipment displays, and pressing dashboard buttons. For this purpose, robotic arms are typically equipped with complex end effectors such as multi-fingered dexterous hands and gripping mechanisms.

[0004] During the aforementioned operations, the robotic arm needs to possess autonomous obstacle avoidance capabilities to avoid collisions with surrounding equipment. Currently, mainstream trajectory planning methods generally rely on primitive geometric primitives defined by URDF (Unified Robot Description Format) to construct collider models. However, the complex structure of the end effector significantly increases the computational complexity of collision detection, causing a noticeable delay in planning response, which in turn restricts the task response efficiency of mobile robots in scenarios with densely distributed equipment. Summary of the Invention

[0005] This disclosure provides a trajectory planning method for a robotic arm, a mobile robot, a medium, and a product to address the problem of high computational complexity in collision models using primitive geometric primitives, thereby improving the task response efficiency of mobile robots in scenarios with densely distributed equipment.

[0006] One aspect of this disclosure provides a trajectory planning method for a robotic arm, comprising: Based on the working environment model corresponding to the robotic arm, determine the initial pose of the end effector on the robotic arm, and the target pose of the end effector in the predetermined operation task; Based on the initial pose and the target pose, a first working space and a second working space are determined; wherein, the first working space is a spatial region covering the initial pose, and the second working space is a spatial region covering the target pose; According to the work environment model, trajectory planning is performed, and during the trajectory planning process, a first collision detection is performed on the trajectory points located in the first work space using a first collision model, and a second collision detection is performed on the second trajectory points located in the second work space using a second collision model. After the trajectory planning is completed, the target motion trajectory of the robotic arm under the predetermined operation task is obtained; The first collision model has lower geometric modeling accuracy than the second collision model, and the first collision model has higher collision detection efficiency than the second collision model.

[0007] Another aspect of this disclosure provides a trajectory planning device for a robotic arm, comprising: The target pose determination module is used to determine the initial pose of the end effector on the robotic arm and the target pose of the end effector in the predetermined operation task based on the working environment model corresponding to the robotic arm. The workspace determination module is used to determine a first workspace and a second workspace based on the initial pose and the target pose; wherein the first workspace is a spatial region covering the initial pose, and the second workspace is a spatial region covering the target pose; The trajectory planning execution module is used to perform trajectory planning according to the work environment model, and in the process of executing trajectory planning, to perform first collision detection on trajectory points located in the first work space through a first collision model, and to perform collision detection on second trajectory points located in the second work space through a second collision model. The target motion trajectory determination module is used to obtain the target motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed. The first collision model has lower geometric modeling accuracy than the second collision model, and the first collision model has higher collision detection efficiency than the second collision model.

[0008] Another aspect of this disclosure provides a mobile robot, which includes a robot body, at least one robotic arm mounted on the robot body, and a control device communicatively connected to the robotic arm. The robotic arm is equipped with an end effector at its end. The control device includes: at least one processor and a memory communicatively connected to the at least one processor, the memory storing a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to execute the trajectory planning method of the robotic arm according to any embodiment of this disclosure.

[0009] Another aspect of this disclosure provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the trajectory planning method for a robotic arm according to any embodiment of this disclosure.

[0010] Another aspect of this disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the trajectory planning method for a robotic arm according to any embodiment of this disclosure.

[0011] The technical solution of this disclosure determines multiple workspaces based on the initial pose of the end effector of the robotic arm in the work environment model and its target pose in the predetermined operation task. During the trajectory planning process based on the work environment model, multiple collision models with different geometric modeling accuracy and collision detection efficiency are obtained. Through the collision models, collision detection is performed on the trajectory points in the corresponding workspaces. After the trajectory planning is completed, the target motion trajectory of the robotic arm under the predetermined operation task is obtained. This solves the problem of high computational complexity of traditional collision models for global collision detection, reduces trajectory search redundancy in the trajectory planning process, improves the trajectory planning efficiency of the robotic arm, and thus ensures the task response efficiency of the mobile robot in scenarios with dense equipment distribution.

[0012] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 A flowchart illustrating a trajectory planning method for a robotic arm provided in one embodiment of this disclosure; Figure 2 This is a schematic diagram of a working environment model provided in one embodiment of the present disclosure; Figure 3 A flowchart illustrating another trajectory planning method for a robotic arm provided in one embodiment of this disclosure; Figure 4 A flowchart illustrating another trajectory planning method for a robotic arm provided in one embodiment of this disclosure; Figure 5This is a schematic diagram of the structure of a trajectory planning device for a robotic arm provided in one embodiment of the present disclosure; Figure 6 This is a schematic diagram of the structure of a mobile robot provided in one embodiment of the present disclosure; Figure 7 This is a schematic diagram of the structure of a control device provided in one embodiment of the present disclosure. Detailed Implementation

[0015] To enable those skilled in the art to better understand the present disclosure, the technical solutions of the present disclosure will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present disclosure, and not all embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present disclosure.

[0016] It should be noted that the terms "predetermined," "initial," "target," "candidate," "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] Figure 1 This is a flowchart illustrating a trajectory planning method for a robotic arm according to an embodiment of this disclosure. This embodiment is applicable to trajectory planning for a robotic arm performing a predetermined operation task. The method can be executed by a trajectory planning device for the robotic arm, which can be implemented in hardware and / or software and can be configured in a terminal device. Figure 1 As shown, the method includes: S110. Based on the working environment model corresponding to the robotic arm, determine the initial pose of the end effector on the robotic arm, and the target pose of the end effector in the predetermined operation task.

[0018] Specifically, a robotic arm is a multi-degree-of-freedom motion mechanism mounted on a mobile robot. Its end effector is equipped with a specified behavior operation for completing a predetermined operation task. For example, the end effector can be a multi-finger dexterous hand, a gripping mechanism, a compound operation mechanism, etc., but is not limited to the given example.

[0019] The operating environment model is a three-dimensional spatial model that represents the environmental information of the mobile robot, such as the distribution of surrounding obstacles, spatial boundaries, and passable areas. For example, the reference coordinate system corresponding to the operating environment model can be the world coordinate system, the base coordinate system of the mobile robot, or the global coordinate system of the operating area, but it is not limited to the given example.

[0020] Specifically, when the mobile robot's operating environment is a static, known environment, the operating environment model can be a pre-built environment model read from the cloud, or a pre-modeled environment model completed offline and stored locally. When the mobile robot's operating environment is a dynamic, unknown environment, the operating environment model is constructed based on environmental perception data collected by sensing devices, which are rigidly mounted on the mobile robot along with the robotic arm. For example, the sensing devices include, but are not limited to, at least one of the following: 3D LiDAR, millimeter-wave radar, depth camera, binocular camera, and infrared sensor; each type of sensor can be one or more.

[0021] In an optional embodiment, the method further includes: acquiring first environmental data collected by a first sensor and second environmental data collected by a second sensor; performing data preprocessing on the first environmental data to obtain third environmental data, and performing object recognition on the second environmental data to determine the operation object data according to the operation object corresponding to the predetermined operation task; and constructing a working environment model corresponding to the robotic arm based on the third environmental data and the operation object data.

[0022] In this embodiment, the first sensor and the second sensor are rigidly mounted on the mobile robot along with the robotic arm, forming a fixed relative pose relationship to ensure spatial consistency between the two sensors and the robotic arm. The two sensors employ a complementary design; specifically, the environmental perception range of the first sensor is greater than that of the second sensor, and the environmental perception accuracy of the first sensor is lower than that of the second sensor.

[0023] Among them, the environmental perception range represents the spatial range in which the sensor can effectively collect environmental information, characterizing the sensor's coverage of the work area, and the environmental perception accuracy refers to the accuracy of the environmental perception data collected by the sensor, which directly affects the accuracy of the construction of the work environment model.

[0024] In one specific embodiment, before preprocessing the first environmental data to obtain the third environmental data, the method further includes: performing pose matching comparison between the first environmental data and the second environmental data and the body structure data of the mobile robot, respectively, and removing the body data of the mobile robot from the first environmental data and the second environmental data to avoid self-collision in subsequent trajectory planning, which would affect the safety and accuracy of trajectory planning.

[0025] In one specific embodiment, the first sensor is a 3D LiDAR, the second sensor is a depth camera, the first and third environmental data are point cloud data, the second environmental data is image data and depth data, and the fourth environmental data is point cloud data. 3D LiDAR has the core advantages of a wide scanning range and high scanning frame rate, enabling rapid full-area coverage of the work environment and making it suitable for obstacle avoidance scenarios in large-scale environments. However, its environmental perception accuracy is limited, with measurement errors typically ranging from 2-3 cm. Furthermore, point cloud data is prone to noise interference such as spikes and bumps. Traditional point cloud filtering relies on parameter tuning for specific scenarios, resulting in poor adaptability to dynamically changing work environments and making it difficult to meet the demands of high-precision modeling. Depth cameras can control measurement errors within 1 mm, but their low sampling frame rate and narrow scanning field of view prevent them from achieving large-scale modeling of the work environment and thus fail to meet the needs of full-area modeling.

[0026] Specifically, the 3D LiDAR and depth camera simultaneously acquire environmental data, and targeted processing operations are performed on the two sets of environmental data: The first point cloud data acquired by the 3D LiDAR undergoes data preprocessing. For example, a statistical filtering algorithm is used to remove isolated spikes in the first point cloud data to reduce environmental noise interference. A downsampling algorithm is then used to downsample the filtered first point cloud data, reducing the amount of environmental data while preserving the overall contour features of the environment. Finally, combined with parameters such as the size, shape, or feature points of the object being manipulated, object recognition is performed on the second environmental data to determine the object being manipulated.

[0027] In an optional embodiment, the step of determining the operation object data by performing object recognition on the second environmental data according to the operation object corresponding to the predetermined operation task includes: determining the object region image by performing target detection on the image data according to the operation object corresponding to the predetermined operation task; performing outer envelope fitting processing on the object region image to obtain an object fitting image; and aligning the object fitting image with the depth data to obtain the operation object data.

[0028] Specifically, an object region image of the target object is segmented from the image data using a target detection algorithm. The object region image is then subjected to outer envelope fitting processing to generate an object fitting image composed of an outer envelope cuboid. For example, the outer envelope cuboid can be expanded outward to match the measurement error of the three-dimensional LiDAR, such as 2 cm, and the object fitting image composed of the expanded outer envelope cuboid is obtained.

[0029] In one specific embodiment, the base coordinate system of the mobile robot is used as the reference coordinate system corresponding to the working environment model. Based on the coordinate transformation matrix obtained by the pre-calibration of the sensor, the third environmental data and the fourth environmental data are uniformly mapped to the base coordinate system of the mobile robot. The point cloud fusion algorithm is used to fuse the aligned third environmental data and the operation object data to obtain the working environment model.

[0030] Figure 2 This is a schematic diagram of a working environment model provided in one embodiment of the present disclosure. Specifically, the white rectangle on the left side near the end of the robotic arm represents the point cloud data corresponding to the depth camera, the discrete geometric bodies distributed around the robotic arm represent the point cloud data corresponding to the three-dimensional LiDAR, and the green rectangle connected to the bottom of the robotic arm represents the robot body of the mobile robot.

[0031] In actual operation scenarios of data computing centers, the layout of server racks is complex, the number of devices is large and the types are diverse, and the overall operating environment has strong differentiation and dynamic change characteristics. The structure of server racks, placement of miscellaneous items, and stacking of temporary equipment in different areas are not uniform. Traditional trajectory planning generally relies on manual on-site measurement, pre-drawing of 3D environment models and setting of obstacle parameters to complete obstacle avoidance configuration. This mode is only suitable for static and fixed scenarios. Faced with the complex working conditions of large-scale, highly dynamic and multi-variable computing centers, there are prominent problems such as heavy workload of manual modeling, lagging environment updates, and insufficient scenario adaptability. It cannot meet the actual needs of mobile robots for normalized and all-area autonomous operation.

[0032] This embodiment achieves complementary performance advantages of different sensors through multi-sensor joint modeling, enabling real-time dynamic 3D reconstruction of the on-site environment, accurate obstacle identification, and instantaneous updating of environmental parameters. It eliminates the reliance on manual pre-modeling and obstacle presets, adapts to varied layouts and dynamic obstacle scenarios, and provides full-domain, high-precision, and high-real-time environmental support for the trajectory planning of the robotic arm.

[0033] Specifically, the initial pose represents the position and attitude information of the end effector in the reference coordinate system corresponding to the working environment model before executing the predetermined operation task. For example, the initial pose can be obtained by inverse solving the joint angle sequence corresponding to each joint in the robotic arm, or it can be a default setting. The joint angle sequence can be collected in real time by angle sensors installed on each link and joint component, or it can be read in response to a reset command, but it is not limited to the example given above.

[0034] In one optional embodiment, determining the initial pose of the end effector on the robotic arm based on the working environment model corresponding to the robotic arm includes: obtaining the joint angle sequence corresponding to the robotic arm; verifying the joint angle sequence according to the joint limit range corresponding to the robotic arm to obtain a compliance verification result; if the compliance verification result indicates that it is valid, determining the original pose of the end effector in the robotic arm coordinate system according to the joint angle sequence; and converting the original pose into the initial pose according to the transformation matrix between the robotic arm coordinate system and the reference coordinate system of the working environment model.

[0035] Specifically, the joint limit range characterizes the angular limit range in which the link and joint components can achieve continuous and stable movement at the reference zero position. If the compliance verification result indicates invalidity, the trajectory planning process of the robotic arm ends.

[0036] The advantage of setting up a compliance verification mechanism is that it can eliminate joint angles that exceed the limits of joint movement, avoid motion interference and structural damage to the robotic arm, and at the same time avoid invalid trajectory calculations and reduce unnecessary consumption of computing resources.

[0037] Specifically, the target pose represents the position and attitude information of the end effector in the reference coordinate system corresponding to the working environment model in order to complete the predetermined operation task. The predetermined operation task represents the predetermined behavior operation performed by the robotic arm on the operation object. For example, the predetermined operation task may be opening and closing the cabinet door, operating the air switch, touching the equipment display screen, pressing the instrument panel button, etc., but is not limited to the given example.

[0038] In one optional embodiment, determining the target pose of the end effector in a predetermined operation task based on the working environment model corresponding to the robotic arm includes: obtaining a predetermined behavior operation and an operation object corresponding to the predetermined operation task; obtaining the object pose of the operation object from the working environment model; and determining the target pose of the end effector based on the predetermined behavior operation and the object pose.

[0039] For example, the operation object corresponding to the task of switching on / off cabinet door is the cabinet door handle, and the predetermined action can be to first clamp and then pull; the operation object corresponding to the task of controlling air switch is the air switch, and the predetermined action is to press.

[0040] Figure 2 Two end effectors are shown using different resolutions. The pose of the end effector closer to the mobile robot body and shown with lower resolution is the initial pose, while the pose of the end effector farther from the mobile robot body and closer to the object being manipulated and shown with higher resolution is the destination pose.

[0041] S120. Determine the first working space and the second working space based on the initial pose and the target pose.

[0042] In this embodiment, the first working space is a spatial region covering the initial pose, and the second working space is a spatial region covering the target pose. The first working space defines the reachable motion domain of the end effector from the initial pose to the transition pose, and the second working space defines the reachable motion domain of the end effector from the transition pose to the target pose. The two are closely connected by the boundary defined by the transition pose.

[0043] In an optional embodiment, determining the first working space and the second working space based on the initial pose and the target pose includes: determining a global working space based on the initial pose and the target pose, and dividing the global working area according to a preset division ratio to obtain the first working space and the second working space.

[0044] Specifically, the global workspace refers to the complete spatial region defined based on the initial pose and the target pose, covering all reachable poses of the robotic arm. Its boundary is determined by the kinematic constraints of the robotic arm and environmental obstacles. The global workspace includes the first workspace and the second workspace.

[0045] Specifically, the preset division ratio represents the spatial allocation weight in the direction from the initial pose to the target pose. For example, the preset division ratio can be 1:1, 7:3 or 2:3, etc. It can be set with reference to the safety margin requirements of the preset operation task. For example, the higher the safety margin, the higher the space ratio of the first operation space, and vice versa, the higher the space ratio of the second operation space, so as to ensure that the end effector has more accurate obstacle avoidance capability when approaching the operation object, and avoids misoperation due to collision when the operation object is a precision instrument.

[0046] In another optional embodiment, determining the first working space and the second working space based on the initial pose and the target pose includes: determining a global working space based on the initial pose and the target pose; obtaining a preset spatial envelope corresponding to the operation object of the predetermined operation task; determining the second working space based on the preset spatial envelope and the target pose; and determining the spatial region in the global working space other than the second working space as the first working space.

[0047] The preset spatial envelope refers to the three-dimensional protective area set around the object to ensure operational safety and accuracy. Its size and shape can be dynamically adapted according to the physical size of the object, the nature of the equipment, the safety level of the equipment, and the operational accuracy requirements of the predetermined operation task. For example, for objects that require pressing, such as air switches, the preset spatial envelope is a flat cuboid with a length-width-height ratio of 3:2:1 to balance the straightness of the pressing path and the lateral obstacle avoidance margin. For micron-level positioning objects, such as oscilloscope probes, the preset spatial envelope is a high-precision spherical area with a radius of no more than 0.5 mm.

[0048] In another optional embodiment, determining the first working space and the second working space based on the initial pose and the target pose includes: obtaining a preset spatial envelope corresponding to the operation object of the predetermined operation task; determining the second working space based on the preset spatial envelope and the target pose; and expanding outward from the spatial boundary of the second working space as a reference until the initial pose is covered to obtain the first working space.

[0049] S130. Based on the work environment model, perform trajectory planning, and during the trajectory planning process, perform first collision detection on the trajectory points located in the first work space using a first collision model, and perform collision detection on the second trajectory points located in the second work space using a second collision model.

[0050] For example, the trajectory planner used to perform the trajectory planning can be an RRTConnect (RapidlyExploring Random Tree Connect) planner, an LBKPIECE (LowerBoundary Knowledge Progressively Informed Exploration) planner, or a lower-boundary projection-based incremental exploration algorithm. (Rapidly Exploring Random Tree Star, Asymptotically Optimal Rapidly Expanding Random Tree) Planner or (Probabilistic Roadmap Star) planner, etc., but not limited to the examples given above.

[0051] In this embodiment, the geometric modeling accuracy of the first collision model is lower than that of the second collision model, and the collision detection efficiency of the first collision model is higher than that of the second collision model.

[0052] Specifically, the collision model refers to a geometrically encapsulated model built based on the rigid body's outline, used for spatial interference and obstacle collision determination. The geometric modeling accuracy characterizes the degree to which the overall outline of the robotic arm is reproduced. For example, the geometric modeling accuracy can be determined based on the attribute parameters of the geometric primitives defined in the collision model, such as shape, size, number, and discreteness.

[0053] In one alternative embodiment, the first collision model and the second collision model each employ a geometric primitive. Exemplarily, the geometric primitives used in the first collision model satisfy at least one of the following conditions compared to the second collision model: fewer primitives, higher degree of dispersion, smaller primitive size, and higher curvature of the primitive contour.

[0054] In another alternative embodiment, the geometric modeling accuracy of the end effector in the first collision model is lower than the geometric modeling accuracy of the end effector in the second collision model.

[0055] Since the end effector is the core component for performing the predetermined operation task and is also the most geometrically complex rigid body on the robotic arm, reducing the geometric modeling accuracy of the end effector in the first collision model can significantly improve the overall efficiency of collision detection. At the same time, improving the geometric modeling accuracy of the end effector in the second collision model ensures fine collision detection during the process of the end effector approaching the operation object. Compared with coarse collision detection, it can accurately identify close-range pose interference with small obstacles, thereby reducing the risk of damage to the precision structure of the operation object and reducing misoperation caused to the surrounding operation objects.

[0056] Specifically, the geometric modeling accuracy of the links and joints of the robotic arm in the first collision model is equal to or less than the geometric modeling accuracy of the end effector in the first collision model, and the geometric modeling accuracy of the links and joints of the robotic arm in the second collision model is equal to or less than the geometric modeling accuracy of the end effector in the second collision model.

[0057] In one specific embodiment, the geometric modeling accuracy of the links and joints is lower than that of the end effector. Since the end effector has a significantly wider range of motion than the links and joints, it is at higher risk of colliding with surrounding obstacles. The differentiated setting of the geometric modeling accuracy of different components on the robotic arm further balances the efficiency and accuracy of the overall collision detection.

[0058] Specifically, the geometric modeling accuracy of linkages and joints in the first and second collision models can be the same, or the geometric modeling accuracy in the first collision model can be lower than that in the second collision model. Therefore, the geometric modeling accuracy of the collision model is related not only to the properties of the geometric primitives, but also to the component volume of the different components in the robotic arm classified according to collision risk.

[0059] Based on the above embodiments, optionally, the first collision model is an envelope collision model, and the second collision model is a geometric collision model. The envelope collision model represents the outer contour of the rigid body through its minimum outer envelope, while the geometric collision model approximates the outer contour of the rigid body through multiple geometric bodies.

[0060] Specifically, the first collision detection includes whether there is a collision with its own structure, whether there is a collision with the mobile robot, and whether there is a collision with surrounding objects; the second collision detection includes whether there is a collision with its own structure, whether there is a collision with the mobile robot it is in, whether there is a collision with surrounding objects, and whether there is a collision with the object being operated on.

[0061] S140. After the trajectory planning is completed, the target motion trajectory of the robotic arm under the predetermined operation task is obtained.

[0062] Specifically, the target motion trajectory represents a continuous pose sequence after dual collision detection by the first collision model and the second collision model.

[0063] In an optional embodiment, obtaining the target motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed includes: determining the obtained motion trajectory as the target motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed.

[0064] In another optional embodiment, obtaining the target motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed includes: obtaining candidate motion trajectories of the robotic arm under the predetermined operation task after the trajectory planning is completed; performing collision verification on the candidate motion trajectories using a third collision model to obtain a collision verification result; and determining the candidate motion trajectory as the target motion trajectory of the robotic arm under the predetermined operation task if the collision verification result indicates that the collision verification is successful.

[0065] In this embodiment, the geometric modeling accuracy of the third collision model is higher than that of the second collision model, and the collision detection efficiency of the third collision model is lower than that of the second collision model.

[0066] Specifically, the collision verification results indicate whether the candidate motion trajectory, under the geometric representation of the higher modeling accuracy of the third collision model, can still ensure that the robotic arm maintains a safe distance from itself and all obstacles in the working environment, thus satisfying the collision-free constraint.

[0067] In one specific embodiment, the first collision model is an envelope collision model, the second collision model is a simplified geometric collision model, and the third collision model is a primitive geometric collision model. The simplified geometric collision model approximates the rigid body's outline using simplified geometry with low fitting, while the primitive geometric collision model accurately reconstructs the rigid body's outline using complex geometry with high fitting.

[0068] Based on the above embodiments, optionally, the step of obtaining the target motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed further includes: if the collision verification result indicates that the collision has failed, obtaining at least one candidate trajectory point in the collision verification result where a collision has occurred, iteratively optimizing the candidate motion trajectory according to each candidate trajectory point until the collision verification result indicates that the collision has passed, and determining the optimized candidate motion trajectory as the target motion trajectory.

[0069] This embodiment, by setting up a collaborative detection mechanism with multi-level collision models, ensures both the efficiency of the global trajectory search stage and the reliability of obstacle avoidance in the final motion trajectory, further balancing the response efficiency and safety accuracy requirements of the trajectory planning process.

[0070] The technical solution of this embodiment determines multiple workspaces based on the initial pose of the end effector of the robotic arm in the work environment model and its target pose in the predetermined operation task. During the trajectory planning process based on the work environment model, multiple collision models with different geometric modeling accuracy and collision detection efficiency are obtained. Through the collision models, collision detection is performed on the trajectory points in the corresponding workspaces. After the trajectory planning is completed, the target motion trajectory of the robotic arm under the predetermined operation task is obtained. This solves the problem of high computational complexity of traditional collision models for global collision detection, reduces trajectory search redundancy in the trajectory planning process, improves the trajectory planning efficiency of the robotic arm, and thus ensures the task response efficiency of the mobile robot in scenarios with dense equipment distribution.

[0071] It is understood that the above embodiments of this disclosure are illustrative examples using two workspaces as an example. In some embodiments, the workspaces corresponding to the robotic arm may be three or more. Along the movement sequence from the initial pose to the target pose, the geometric modeling accuracy of the collision model corresponding to each workspace decreases sequentially, and the collision detection accuracy of the collision model corresponding to each workspace increases sequentially.

[0072] The trajectory planning of robotic arms involves the coordinated optimization of multiple performance parameters, such as planning efficiency, planning quality, and planning stability. The dynamic balance of these multi-dimensional parameters is crucial to ensuring the overall operational stability of mobile robots. However, single trajectory planners, limited by their inherent performance constraints, often exhibit significant differences in adaptability to different operating environments. Consequently, they struggle to adapt to diverse operational tasks and complex, ever-changing working environments, severely hindering the widespread application of mobile robots.

[0073] Figure 3This is a flowchart of another trajectory planning method for a robotic arm provided in one embodiment of this disclosure. This embodiment further refines the steps of "performing trajectory planning according to the work environment model" and "obtaining candidate motion trajectories of the robotic arm under the predetermined operation task" in the above embodiment. In this embodiment, performing trajectory planning according to the work environment model includes: performing trajectory planning in parallel according to the work environment model using multiple trajectory planners; obtaining the target motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed includes: obtaining an initial motion trajectory set output by the multiple trajectory planners after all multiple trajectory planners have been executed, and traversing the initial motion trajectories in the initial motion trajectory set; determining at least one quality parameter value based on the traversed initial motion trajectories, and determining the trajectory quality score corresponding to each initial motion trajectory based on each quality parameter value; and determining the target motion trajectory of the robotic arm under the predetermined operation task based on the initial motion trajectory with the smallest trajectory quality score after traversal.

[0074] like Figure 3 As shown, the method includes: S210. Based on the working environment model corresponding to the robotic arm, determine the initial pose of the end effector on the robotic arm, and the target pose of the end effector in the predetermined operation task.

[0075] S220. Determine the first working space and the second working space based on the initial pose and the target pose.

[0076] S210-S220 in this embodiment are the same as those in the above embodiments. Figure 1 The S110-S120 shown are the same or similar, and will not be described again here.

[0077] S230. Using multiple trajectory planners, trajectory planning is performed in parallel according to the work environment model. During the trajectory planning process, a first collision detection is performed on the trajectory points located in the first work space using a first collision model, and a second collision detection is performed on the second trajectory points located in the second work space using a second collision model.

[0078] Specifically, the adaptability of different trajectory planners to the same working environment will exhibit a non-uniform, discrete distribution. Adaptability represents quantitative parameters for evaluating the performance of each trajectory planner under the same working environment. Different working environments may have different performance requirements for trajectory planners, therefore the performance parameters used to characterize adaptability may differ. For example, normalized working environments typically require trajectory planners to have continuous operational stability to ensure the reliable execution of repetitive tasks; working environments requiring precise operations require trajectory planners to have high-quality trajectory planning capabilities to meet the demands of high-precision tasks.

[0079] For example, the performance parameters used to characterize fitness include, but are not limited to, at least one of planning efficiency, trajectory smoothness, spatial coverage, planning stability, robustness, and planning quality, wherein planning quality can be determined based on one or more trajectory quality parameters.

[0080] In one specific embodiment, the trajectory planners include the RRTConnect planner, the LBKPIECE planner, and... Planner and Planner.

[0081] Among them, the RRTConnect planner is a trajectory planner with high planning efficiency in the diversified trajectory planning system. Its core principle is to simultaneously construct search trees from the initial pose and the target pose, explore two workspaces at the same time, and attempt to connect the two search trees. Compared with the one-way search mode, it greatly shortens the trajectory search time, enabling it to output obstacle avoidance motion trajectories in a very short time. The LBKPIECE planner is a trajectory planner with high spatial coverage in the diversified trajectory planning system. Its core principle is to perform gridded projection processing on the workspace, track the spatial distribution of explored trajectory points in real time, and guide sampling points to perform directional expansion and sparse compensation sampling to unexplored spatial areas, thereby outputting motion trajectories with comprehensive coverage. The planner is a trajectory planner with high planning quality in the diversified trajectory planning system. Its core principle is to quickly find an obstacle avoidance trajectory and then continuously optimize the obstacle avoidance trajectory so that the planned trajectory can gradually approach the optimal solution in mathematical theory under preset quality constraints. The planner is a trajectory planner with high planning stability in the diversified trajectory planning system. Its core principle is to convert the operation environment model into a topological structure, construct a global undirected road network through random incremental sampling, continuously optimize the node connection relationship and path cost weight, eliminate redundant and invalid nodes and unstable links, and weaken the planning deviation caused by environmental disturbances by relying on the iterative update mechanism of the progressively optimal road network.

[0082] For example, in dynamic work scenarios, the poses of obstacles in the work environment change, and the performance advantage of the RRTConnect planner in planning efficiency can adapt to the real-time planning requirements of this scenario; in narrow and complex work scenarios, obstacles are densely distributed within the mobile robot's work area, and the LBKPIECE planner, relying on its performance advantage of full-domain spatial coverage, can overcome the occlusion limitations of complex obstacles and adapt to the passage requirements of this scenario; in refined work scenarios, The planner's high-quality planning capabilities can adapt to the application constraints of this scenario, such as the operating cost and work quality of mobile robots. Leveraging its performance advantage in planning stability, the planner can quickly extract highly consistent motion trajectories from a pre-built environmental topology network, making it more adaptable to routine operation scenarios with long-term repetition and fixed working conditions.

[0083] S240. After all the trajectory planners have been executed, the initial motion trajectory set output by the multiple trajectory planners is obtained, and the initial motion trajectory in the initial motion trajectory set is traversed.

[0084] Specifically, the initial motion trajectory set contains one or more initial motion trajectories that satisfy the collision-free constraint, output by each trajectory planner.

[0085] S250. Based on the initial motion trajectory obtained through traversal, determine at least one quality parameter value, and based on each of the quality parameter values, determine the trajectory quality score corresponding to the initial motion trajectory.

[0086] In an optional embodiment, the trajectory quality parameter corresponding to the quality parameter value is trajectory travel, trajectory smoothness, or trajectory execution time, and the trajectory travel includes trajectory length and / or the joint curvature corresponding to the robotic arm.

[0087] Specifically, trajectory length refers to the total length of the Cartesian space path formed by the continuous movement of the end effector along the workspace, and joint radians refer to the cumulative value of the angular displacement generated by the overall movement of the links and joint components of the robotic arm, representing the total radian rotated by the motor, and can reflect the energy consumption and execution efficiency of the mobile robot. The shorter the trajectory length and joint radians, the higher the trajectory quality of the initial motion trajectory.

[0088] by Taking a axial robotic arm as an example, the joint curvature Satisfy the following formula: in, Indicates the first The joint in the first Joint angles at each sampling trajectory point This indicates the number of samples for the initial trajectory points.

[0089] Specifically, trajectory smoothness includes geometric smoothness and / or motion smoothness. Geometric smoothness refers to the continuity of the initial motion trajectory's contour and curvature variation characteristics in Cartesian space and joint space. Motion smoothness refers to the degree of fluctuation in the motion parameters of the robotic arm's links and joints caused by the initial motion trajectory after time parameterization. For example, motion parameters include, but are not limited to, at least one of velocity, acceleration, and jump. Higher geometric smoothness and motion smoothness indicate higher trajectory quality of the initial motion trajectory.

[0090] For example, motion smoothness Satisfy the following formula: in, Indicates the trajectory execution time. Indicates a joint index. Indicates the first Each joint at any time The jerk, or the first derivative of acceleration with respect to time, is used to characterize the intensity of the impact of joint movement. This indicates the number of samples for the initial trajectory points. Indicates the first The joint in the first Acceleration at each sampling trajectory point Indicates the first The sampling time interval corresponding to each sampling trajectory point.

[0091] Specifically, trajectory execution time refers to the total time taken for the robotic arm to complete the initial motion trajectory after time parameterization processing.

[0092] For example, trajectory execution time ,in, Indicates the end time. Indicates the start time.

[0093] In one specific embodiment, based on the initial motion trajectory obtained through traversal, the quality parameter values ​​corresponding to joint curvature, motion smoothness, and trajectory execution time are determined, and a trajectory quality score is generated. Satisfying the formula: ,in, , and This represents the weight values ​​corresponding to joint curvature, motion smoothness, and trajectory execution time, respectively.

[0094] For example, in scenarios requiring stable operation, the weight value of trajectory smoothness can be set to the highest to reduce motion impact during the execution of predetermined operation tasks by the robotic arm, thereby ensuring the operational stability of the robotic arm in long-term continuous operation and reducing the risk of vibration noise and joint wear. In scenarios requiring high-efficiency operation, the weight value of trajectory execution time can be set to the highest to achieve real-time planning and rapid response in dynamic operation environments. In scenarios requiring low-loss operation, the weight value of trajectory travel can be set to the highest to reduce the overall motion loss of the robotic arm and achieve low maintenance costs for long-term repetitive operations.

[0095] S260. After the traversal is completed, the initial motion trajectory with the smallest trajectory quality score is determined as the target motion trajectory of the robotic arm under the predetermined operation task.

[0096] The technical solution of this embodiment introduces multiple trajectory planners to construct a diversified parallel planning system. Based on trajectory travel, trajectory smoothness, or trajectory execution time, it achieves the selection of motion trajectories. Under the premise of meeting the requirements of the scenario in terms of response efficiency, it solves the problem that a single trajectory planner has inherent limitations in complex and ever-changing working conditions. Relying on the algorithm characteristics of different trajectory planners, it complements and adapts to diverse operational requirements, thereby improving the operational robustness of the mobile robot and taking into account the high efficiency of the mobile robot's responsiveness, operational stability, and scenario versatility.

[0097] Figure 4 This is a flowchart illustrating another trajectory planning method for a robotic arm according to an embodiment of this disclosure. This embodiment further optimizes the trajectory planning method for the robotic arm described in the above embodiment, such as... Figure 4 As shown, the method includes: S310. Based on the working environment model corresponding to the robotic arm, determine the initial pose of the end effector on the robotic arm, and the target pose of the end effector in the predetermined operation task.

[0098] S320. Determine the first working space and the second working space based on the initial pose and the target pose.

[0099] S330. Based on the work environment model, perform trajectory planning, and during the trajectory planning process, perform first collision detection on the trajectory points located in the first work space using a first collision model, and perform collision detection on the second trajectory points located in the second work space using a second collision model.

[0100] S340. After the trajectory planning is completed, the target motion trajectory of the robotic arm under the predetermined operation task is obtained.

[0101] In an optional embodiment, S310-S340 are the same as those in the above embodiments. Figure 1 S110-S140 shown are the same as or similar to those in the above embodiments. Figure 3 The S210-S260 shown are the same or similar, and will not be described again here.

[0102] In another optional embodiment, obtaining the target motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed includes: after multiple trajectory planners have been executed, obtaining an initial motion trajectory set output by the multiple trajectory planners, and traversing the initial motion trajectories in the initial motion trajectory set; determining at least one quality parameter value based on the traversed initial motion trajectories, and determining the trajectory quality score corresponding to each initial motion trajectory based on each quality parameter value; after the traversal is completed, determining the initial motion trajectory with the smallest trajectory quality score as the candidate motion trajectory of the robotic arm under the predetermined operation task; performing collision verification on the candidate motion trajectory using a third collision model to obtain a collision verification result; and if the collision verification result indicates that the collision verification is successful, determining the candidate motion trajectory as the target motion trajectory of the robotic arm under the predetermined operation task.

[0103] S350. The first time parameterization algorithm is used to perform time-series mapping processing on the target motion trajectory to obtain the algorithm processing result.

[0104] In this embodiment, the dynamic performance of the first time parameterization algorithm is better than that of the second time parameterization algorithm, and the robustness of the first time parameterization algorithm is lower than that of the second time parameterization algorithm.

[0105] S360. If the algorithm processing result indicates failure, a second time parameterization algorithm is used to perform time-series mapping processing on the target motion trajectory to obtain a time parameterized trajectory.

[0106] In one specific embodiment, the first time parameterization algorithm is the TOTG algorithm (Time Optimal Trajectory Generation), and the second time parameterization algorithm is the IPTP algorithm (Iterative Parabolic Time Parameterization).

[0107] The TOTG algorithm, based on numerical integration or convex optimization, minimizes trajectory execution time by solving the limiting velocity curve, under the premise of satisfying the maximum extremum constraints of joint velocity and acceleration, so that the robotic arm operates close to its performance limit at every moment. The TOTG algorithm generates trajectories with higher smoothness and optimal execution efficiency, but its computational cost is relatively large, and it is prone to convergence failure when the path points are unevenly distributed or there are singular points.

[0108] For example, when the motion parameters such as angular velocity and angular acceleration of the robotic arm joints exceed the limit constraint range, the TOTG algorithm cannot complete the smooth time allocation and interpolation solution, and the iterative calculation process is prone to divergence, which in turn leads to calculation errors. In addition, when the algorithm parameters such as time sampling step size, convergence threshold, and maximum number of iterations are unreasonable, problems such as calculation timeout, numerical overflow, and local unsolvable problems are likely to occur.

[0109] Among them, the IPTP algorithm is based on a constant acceleration parabolic hybrid model between adjacent path points. Through a heuristic process of "initialization time allocation - over-limit detection - iterative correction", it adjusts the time span of each path segment to meet physical constraints. Although the motion efficiency of the IPTP algorithm is conservative and acceleration discontinuity is prone to occur at the connection of trajectory points, which can easily cause slight vibrations, it is robust, computationally efficient, and has no convergence risk. It can ensure the stable realization of time parameterization of the trajectory and ensure the reliability and stability of the time allocation process.

[0110] S370. Based on the time parameterized trajectory, control the robotic arm to execute the predetermined operation task.

[0111] In an optional embodiment, the method further includes: if the algorithm processing result indicates success, obtaining the time parameterized trajectory in the algorithm processing result, and controlling the robotic arm to perform the predetermined operation task according to the time parameterized trajectory.

[0112] The motion trajectory directly output by the trajectory planner has problems such as an uncertain number of trajectory points, uneven time intervals, and slight jagged noise. However, when controlling a robotic arm to perform a predetermined operation, the motion trajectory in the control command must meet conditions such as a fixed number of points, a fixed period, and smoothness without impact. Too many trajectory points will slow down the robotic arm's movement rhythm and delay the response. At the same time, the communication link has bandwidth limitations, and a large number of trajectory points will cause congestion on the serial port or bus. In particular, some robotic arms have limited buffers in their command receiving devices, which cannot handle high-density trajectory points. Uneven time intervals cannot be adapted to the fixed drive frequency required to control the robotic arm, and noise in the motion trajectory will impact the motor, thus affecting the motion stability of the robotic arm.

[0113] In an optional embodiment, controlling the robotic arm to perform the predetermined operation task based on the time parameterized trajectory includes: compressing the time parameterized trajectory and controlling the robotic arm to perform the predetermined operation task based on the obtained compressed time parameterized trajectory.

[0114] For example, the compression algorithms used in the compression processing include, but are not limited to, equal-interval resampling algorithms, Douglas-Pock algorithms, curvature-based compression methods, or spline curve fitting algorithms. The principle of equal-interval resampling is to perform linear interpolation on the original trajectory based on a fixed number of points or a fixed step size, generating a sequence of trajectory points with equal intervals. The principle of the Douglas-Pock algorithm is to use a recursive strategy to iteratively identify and remove redundant points with small deviations from the original path, achieving data simplification while preserving key trajectory morphological features. The principle of curvature-based compression is to dynamically adjust the sampling density based on the local curvature characteristics of the trajectory, retaining more trajectory points in areas of high curvature and reducing the number of sampling points in areas of low curvature, achieving a balance between data volume and trajectory morphological fidelity. The principle of spline curve fitting algorithms is to fit the original trajectory points as a whole using a B-spline mathematical model, constructing a continuous trajectory function expression, and then resampling on the fitted curve as needed according to application requirements to generate an optimized trajectory sequence.

[0115] In one specific embodiment, a spline curve fitting algorithm is used to compress the target motion trajectory. The spline curve fitting algorithm has at least the following advantages: 1. Local support: Modifying local control points on the curve only affects the corresponding local trajectory segment and does not interfere with the shape of the entire trajectory, making the algorithm highly robust when handling motion trajectories with complex obstacle avoidance paths; 2. High-order motion continuity: The spline curve fitting algorithm can ensure the second-order continuity of the trajectory in position, velocity, and acceleration, eliminating abrupt shocks during motor torque switching and significantly improving the stability of motion control; 3. High data compression ratio: Only a few control points are needed to characterize complex curve shapes, greatly reducing the communication bandwidth occupied between the host computer and the underlying driver, and effectively avoiding the risk of data overflow; 4. Low-pass filtering characteristics: The fitting process can automatically eliminate sawtooth noise in the motion trajectory without the need for additional filtering processing, simplifying the overall control process.

[0116] The technical solution in this embodiment solves the problem of the inability to balance dynamic performance and robustness by adopting a degradation strategy to achieve time parameterization of the trajectory. It prioritizes the activation of the first time parameterization algorithm with better dynamic performance to leverage its performance advantages. When abnormal situations such as algorithm non-convergence occur, the second time parameterization algorithm with higher robustness is automatically activated to maintain the continuity of the control flow. By utilizing the synergistic linkage of the two algorithms, the reliability of task execution in complex working conditions is improved, and a dynamic balance between task execution efficiency and stability of the mobile robot is achieved.

[0117] The following are embodiments of the trajectory planning device for a robotic arm provided in this disclosure. This device and the trajectory planning method for a robotic arm described above belong to the same inventive concept. For details not described in detail in the embodiments of the trajectory planning device for a robotic arm, please refer to the content on the trajectory planning method for a robotic arm described above.

[0118] Figure 5 This is a schematic diagram of the structure of a trajectory planning device for a robotic arm provided in one embodiment of this disclosure. Figure 5 As shown, the device includes: a target pose determination module 410, a workspace determination module 420, a trajectory planning and execution module 430, and a target motion trajectory determination module 440.

[0119] The target pose determination module 410 is used to determine the initial pose of the end effector on the robotic arm and the target pose of the end effector in the predetermined operation task based on the working environment model corresponding to the robotic arm. The workspace determination module 420 is used to determine a first workspace and a second workspace based on the initial pose and the target pose; wherein, the first workspace is a spatial region covering the initial pose, and the second workspace is a spatial region covering the target pose; The trajectory planning execution module 430 is used to perform trajectory planning according to the work environment model, and in the process of executing trajectory planning, to perform first collision detection on trajectory points located in the first work space through a first collision model, and to perform collision detection on second trajectory points located in the second work space through a second collision model. The target motion trajectory determination module 440 is used to obtain the target motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed; The first collision model has lower geometric modeling accuracy than the second collision model, and the first collision model has higher collision detection efficiency than the second collision model.

[0120] The technical solution of this embodiment determines multiple workspaces based on the initial pose of the end effector of the robotic arm in the work environment model and its target pose in the predetermined operation task. During the trajectory planning process based on the work environment model, multiple collision models with different geometric modeling accuracy and collision detection efficiency are obtained. Through the collision models, collision detection is performed on the trajectory points in the corresponding workspaces. After the trajectory planning is completed, the target motion trajectory of the robotic arm under the predetermined operation task is obtained. This solves the problem of high computational complexity of traditional collision models for global collision detection, reduces trajectory search redundancy in the trajectory planning process, improves the trajectory planning efficiency of the robotic arm, and thus ensures the task response efficiency of the mobile robot in scenarios with dense equipment distribution.

[0121] In one alternative embodiment, the geometric modeling accuracy of the end effector in the first collision model is lower than the geometric modeling accuracy of the end effector in the second collision model.

[0122] In an optional embodiment, the target motion trajectory determination module 440 includes: The collision verification unit is used to obtain the candidate motion trajectory of the robotic arm under the predetermined operation task after the trajectory planning is completed. The candidate motion trajectory is collided using a third collision model to obtain a collision verification result. The geometric modeling accuracy of the third collision model is higher than that of the second collision model, and the collision detection efficiency of the third collision model is lower than that of the second collision model. If the collision verification result indicates that the collision has passed, the candidate motion trajectory is determined as the target motion trajectory of the robotic arm under the predetermined operation task.

[0123] In an optional embodiment, the trajectory planning execution module 430 is specifically used for: Multiple trajectory planners are used to perform trajectory planning in parallel according to the work environment model; The target trajectory determination module 440 includes: The initial motion trajectory set filtering unit is used to obtain the initial motion trajectory set output by the multiple trajectory planners after all multiple trajectory planners have been executed, and to traverse the initial motion trajectories in the initial motion trajectory set. Based on the initial motion trajectory obtained through traversal, at least one quality parameter value is determined, and based on each of the quality parameter values, the trajectory quality score corresponding to the initial motion trajectory is determined. After the traversal is completed, the initial motion trajectory with the lowest trajectory quality score is determined as the target motion trajectory of the robotic arm under the predetermined operation task. The trajectory quality parameters corresponding to the quality parameter values ​​are trajectory travel, trajectory smoothness, or trajectory execution time, and the trajectory travel includes trajectory length and / or the joint curvature corresponding to the robotic arm.

[0124] In an optional embodiment, the method further includes: The pre-defined operation task execution module is used to perform time-series mapping processing on the target motion trajectory using a first time parameterization algorithm to obtain the algorithm processing result. If the algorithm processing result indicates failure, a second time parameterization algorithm is used to perform time-series mapping processing on the target motion trajectory to obtain a time parameterized trajectory. Based on the time-parameterized trajectory, the robotic arm is controlled to perform the predetermined operation task; The first time parameterization algorithm has better dynamic performance than the second time parameterization algorithm, but the robustness of the first time parameterization algorithm is lower than that of the second time parameterization algorithm.

[0125] In an optional embodiment, the device further includes: The working environment model construction module is used to acquire the first environmental data collected by the first sensor and the second environmental data collected by the second sensor. The first environmental data is preprocessed to obtain the third environmental data, and the second environmental data is object identified to determine the operation object data according to the operation object corresponding to the predetermined operation task. Based on the third environmental data and the operation object data, a working environment model corresponding to the robotic arm is constructed; The first sensor has a larger environmental perception range than the second sensor, and the first sensor has a lower environmental perception accuracy than the second sensor. The first sensor and the second sensor are rigidly mounted on the mobile robot along with the robotic arm.

[0126] In an optional embodiment, the second environmental data includes image data and depth data, and the work environment model construction module is specifically used for: Based on the operation object corresponding to the predetermined operation task, target detection is performed on the image data to determine the object region image; The object region image is subjected to outer envelope fitting processing to obtain the object fitting image; Align the fitted image of the object with the depth data to obtain the object data for operation.

[0127] The trajectory planning device for the robotic arm provided in this disclosure can execute the trajectory planning method for the robotic arm provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects of the execution method.

[0128] Figure 6 This is a schematic diagram of the structure of a mobile robot provided in one embodiment of the present disclosure, as shown below. Figure 6 As shown, the mobile robot 500 includes a robot body 510, at least one robotic arm 520 mounted on the robot body 510, and a control device communicatively connected to the robotic arm 520. Figure 6 (Not shown in the image), an end effector 521 is installed at the end of the robotic arm 520. Specifically, the number of control devices can be one or matched with the number of robotic arms 520, with each control device communicating with a corresponding robotic arm 520.

[0129] Based on the above embodiments, optionally, the mobile robot 500 further includes a first sensor 540 and a second sensor 550, the first sensor 540 and the second sensor 550 being rigidly mounted on the robot body 510 with the robotic arm 520, the first sensor 540 being used to collect first environmental data, and the second sensor 550 being used to collect second environmental data.

[0130] Specifically, the environmental perception range of the first sensor 540 is greater than that of the second sensor 550, and the environmental perception accuracy of the first sensor 540 is lower than that of the second sensor 550.

[0131] Figure 7 This is a schematic diagram of a control device provided according to one embodiment of the present disclosure. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0132] like Figure 7 As shown, the control device 530 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor 11. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from the storage unit 18. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14.

[0133] Input / output (I / O) interface 15 is also connected to bus 14. Multiple components in control device 530 are connected to I / O interface 15, including: input unit 16, output unit 17, storage unit 18, and communication unit 19. Communication unit 19 allows control device 530 to exchange information or data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0134] The processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. The processor 11 executes the various methods and processes described above, such as the trajectory planning method for the robotic arm provided in the above embodiments.

[0135] In some embodiments, the trajectory planning method for the robotic arm provided in the above embodiments can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the control device 530 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the trajectory planning method for the robotic arm described above can be performed.

[0136] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of embodiments of this disclosure.

[0137] The computer program may be written in any combination of one or more programming languages. These computer programs may be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the functions / operations specified in the flowchart and / or block diagram are performed. The computer program may execute entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0138] In the context of this application, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0139] The systems and technologies described herein can be implemented in any combination of backend components, middleware components, or frontend components in a computing system that interconnects its components via digital data communication of any form or medium. The computing system may include clients and servers, which typically interact via a communication network, with the client-server relationship established by computer programs running on the respective computers and having a client-server relationship with each other.

[0140] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this disclosure can be achieved, and this is not limited herein.

[0141] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A trajectory planning method of a robot arm, characterized by, The method comprises the following steps: determining an initial pose of an end effector on a mechanical arm and a target pose of the end effector in a predetermined operation task according to a work environment model corresponding to the mechanical arm; determining a first work space and a second work space according to the initial pose and the target pose; the first work space is a space region covering the initial pose, and the second work space is a space region covering the target pose; performing trajectory planning according to the work environment model, and in the process of performing trajectory planning, performing first collision detection on a trajectory point located in the first work space by a first collision model and performing collision detection on a second trajectory point located in the second work space by a second collision model; obtaining a target motion trajectory of the mechanical arm in the predetermined operation task after the trajectory planning is performed; the geometric modeling accuracy of the first collision model is lower than that of the second collision model, and the collision detection efficiency of the first collision model is higher than that of the second collision model.

2. The method of claim 1, wherein, The geometric modeling accuracy of the end effector in the first collision model is lower than that of the end effector in the second collision model.

3. The method of claim 1, wherein, The method further comprises the following steps: obtaining a candidate motion trajectory of the mechanical arm in the predetermined operation task after the trajectory planning is performed; performing collision verification on the candidate motion trajectory by a third collision model to obtain a collision verification result; the geometric modeling accuracy of the third collision model is higher than that of the second collision model, and the collision detection efficiency of the third collision model is lower than that of the second collision model; in a case where the collision verification result indicates that the collision verification is passed, determining the candidate motion trajectory as the target motion trajectory of the mechanical arm in the predetermined operation task.

4. The method of claim 1, wherein, The method further comprises the following steps: performing trajectory planning in parallel by a plurality of trajectory planners according to the work environment model; The method further comprises the following steps: obtaining an initial motion trajectory set output by the plurality of trajectory planners after the plurality of trajectory planners are all performed, and traversing the initial motion trajectory in the initial motion trajectory set; determining at least one quality parameter value according to the traversed initial motion trajectory, and determining a trajectory quality score corresponding to the initial motion trajectory according to each quality parameter value; after the traversal is completed, determining an initial motion trajectory with the minimum trajectory quality score as the target motion trajectory of the mechanical arm in the predetermined operation task; the trajectory quality parameter corresponding to the quality parameter value is trajectory stroke, trajectory smoothness or trajectory execution time, and the trajectory stroke comprises trajectory length and / or joint arc of the mechanical arm.

5. The method of claim 1, wherein, The method further comprises the following steps: performing time sequence mapping processing on the target motion trajectory by using a first time parameterization algorithm to obtain an algorithm processing result. In a case where the algorithm processing result indicates a failure, a second time parameterization algorithm is used to perform time sequence mapping processing on the target motion trajectory to obtain a time parameterization trajectory; According to the time parameterization trajectory, the mechanical arm is controlled to perform the predetermined operation task; The first time parameterization algorithm has better dynamics performance than the second time parameterization algorithm, and the first time parameterization algorithm has lower robustness than the second time parameterization algorithm.

6. The method according to any one of claims 1 to 5, characterized in that, The method further comprises: obtaining first environment data collected by a first sensor and second environment data collected by a second sensor; performing data preprocessing on the first environment data to obtain third environment data, and performing object recognition on the second environment data to determine operation object data according to an operation object corresponding to the predetermined operation task; constructing a working environment model corresponding to the mechanical arm according to the third environment data and the operation object data; The environment perception range of the first sensor is greater than that of the second sensor, and the environment perception accuracy of the first sensor is lower than that of the second sensor. The first sensor and the second sensor are rigidly installed on a mobile robot with the mechanical arm.

7. The method of claim 6, wherein, The second environment data includes image data and depth data. The operation object data is determined by performing object recognition on the second environment data according to an operation object corresponding to the predetermined operation task, comprising: performing target detection on the image data to determine an object region image according to an operation object corresponding to the predetermined operation task; performing envelope fitting processing on the object region image to obtain an object fitting image; aligning the object fitting image with the depth data to obtain operation object data.

8. A mobile robot, characterized by The mobile robot comprises a robot body, at least one mechanical arm installed on the robot body, and a control device in communication connection with the mechanical arm; The end of the mechanical arm is provided with an end effector. The control device comprises at least one processor and a memory in communication connection with the at least one processor. The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor to enable the at least one processor to execute the trajectory planning method of the mechanical arm according to any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for enabling the processor to execute the trajectory planning method of the mechanical arm according to any one of claims 1-7.

10. A computer program product, characterised in that, The computer program is executed by the processor to implement the trajectory planning method of the mechanical arm according to any one of claims 1-7.

Images