680 results about "Collision free" patented technology

The invention provides a teaching-free welding robot path planning system based on visual autonomous learning, and relates to the technical field of industrial welding, and the teaching-free welding robot path planning system comprises a pose compensation module which is used for arranging three infrared reflective mark points on the surface of a clamp base in an L shape, collecting original point clouds of the mark points and a workpiece by using a stereo camera, and preprocessing the point clouds; generating a pose compensation matrix based on the offset of the actual coordinate and the theoretical coordinate of the mark point, and outputting a de-noising point cloud data set and a compensation matrix; and the pose correction module is used for fitting a workpiece plane equation by adopting a random sampling consistency algorithm on the basis of the de-noised point cloud data set, acting a pose compensation matrix on the workpiece plane equation, correcting a pose error caused by clamp drifting, calculating a welding seam track end point coordinate and outputting a corrected welding seam pose parameter and a scanning pose instruction. According to the method, clamp drift can be compensated, weld joint poses can be corrected, three-dimensional features are extracted, a collision-free path is planned, and the welding precision, efficiency and automation degree are improved.

The invention provides a cooperative mechanical arm self-adaptive obstacle avoidance method in a dynamic scene, which comprises the following steps of: 1, setting an initial pose and a target pose according to a work task of a cooperative robot, and performing global path planning in an initial environment; 2, an operator completes an assembly task in an assembly area, and the collaborative robot operates according to a preset route and completes the assembly task in a matched mode; 3, the depth camera obtains a human body skeleton and skeleton node information, and a human body bounding box is generated; 4, whether the distance between the arm bounding box and the robot is smaller than the safety distance or not is judged, and if yes, collision early warning is triggered; 5, a new collision-free track is re-planned based on the self-adaptive obstacle avoidance strategy of the artificial potential field method, and the mechanical arm moves according to the new track to avoid the obstacle; and 6, after the obstacle leaves the safety early warning range, the mechanical arm returns to the global path and finally arrives at the target point.

The invention relates to a multi-product-line-oriented robot welding integration method, device and equipment and a medium, and the method comprises the steps that three-dimensional geometrical characteristics and material attribute data of multiple types of workpieces are collected, and geometric parameters and material mechanical properties of weld joints are extracted; an order demand template is intelligently matched, and initial welding parameters and a clamp configuration scheme are generated; the health state of production equipment is evaluated in real time, and the collision-free movement track of the mechanical arm is planned; dynamically analyzing the joint angle sequence to generate a high-precision servo control signal; the welding process is monitored in real time, and current and molten pool temperature parameters are corrected; and iteratively optimizing the algorithm weight of the process knowledge base to improve the adaptive capacity. Through cooperation of multi-source data fusion and an intelligent algorithm, rapid switching of multiple kinds of workpieces, self-adaptive compensation of the equipment operation state and closed-loop optimization of welding process parameters are achieved, and the equipment utilization rate and the welding quality stability in a flexible production scene are remarkably improved.

A method for robot path planning using skills extracted from human-taught motion programs applied to a new obstacle environment. A three-dimensional convolutional neural network is used to extract features characterizing an obstacle environment, where the feature vector representation of the obstacles overcomes problems encountered when using point cloud obstacle data. The obstacle feature data and robot path start and goal points are provided to an encoder / decoder neural network system which is trained to extract skills from a database of human-generated motion programs. The encoder / decoder neural network system produces a distribution of waypoints for the current obstacle environment and start / goal points. The distribution of waypoints is used to perform a final collision-free path generation using either a rapidly-exploring random tree (RRT) technique or an optimization-based technique.

The invention discloses a mobile robot autonomous exploration method based on deep reinforcement learning, and the method comprises the steps: starting laser radar equipment, a calculation unit and a mobile robot body power supply in an unknown environment, and collecting the point cloud data of the current environment through a laser radar; storing the point cloud data based on the complementary hole structure, constructing a three-dimensional point cloud map, and constructing a two-dimensional grid occupation map after projection; constructing an information graph based on the two-dimensional grid-occupying map, and sparsifying the information graph by using a bidirectional A * algorithm to form a sparse information graph; a reinforcement learning strategy network is constructed, a target function of the reinforcement learning strategy network is designed, the input of the reinforcement learning strategy network is a sparse information graph, and the output of the reinforcement learning strategy network is a target path point at the next moment; a path planner generates a collision-free path according to the target path point, and a local planner generates a linear velocity value and an angular velocity value of the current robot according to the collision-free path; a motion controller of the robot generates a motion command according to the speed value, the mobile robot is controlled to reach the target path point, and the steps are repeated until the robot autonomous exploration task based on deep reinforcement learning is completed; the method has the advantages that a model-free deep reinforcement learning technology is adopted to learn an exploration strategy in a simulation system, the method can be seamlessly transplanted to a real physical environment, higher-efficiency autonomous exploration can be realized, the adaptability to the environment is high, the application scene is wider, and a foundation is laid for completely autonomous operation of the mobile robot.

The invention discloses an automatic fastening system and method for overhead line system cantilever bolts. The system comprises a maintenance car, a visual ranging parking unit, a laser radar sensing unit, a mechanical arm execution unit, a force feedback control unit, a tail end execution unit, a digital twin unit and an operation management platform. The method comprises the following steps: S1, guiding a maintenance car to stop below a cantilever of the overhead line system; s2, scanning an overhead line system cantilever, generating a three-dimensional model and positioning a bolt; s3, planning a collision-free path; s4, smooth alignment of the bolt sleeve is achieved; s5, bolt fastening is completed according to the preset torque; s6, synchronously displaying the operation progress, and storing whole-process operation data; and S7, performing anomaly detection and task continuous transmission on the whole process, and performing data archiving and early warning. Intelligent and automatic bolt fastening operation is achieved for the first time in the field of railway contact network maintenance, and an effective solution is provided for improving the railway operation and maintenance efficiency and safety.

An example method executed by a controller onboard a spacecraft generates a mission plan in real-time that guides the spacecraft along a space autonomous mission to rendezvous with two or more orbiting target objects. The method includes establishing potential permutations for all possible unique maneuver sequences in which to visit the two or more orbiting target objects, and for each potential permutation, determining a collision-free maneuver plan for defining orbit trajectories to intercept each of the two or more orbiting target objects for each of the possible unique maneuver sequences. An optimal permutation is determined that meets viewing constraints of the two or more orbiting target objects, a viewing priority, and fuel constraints of the spacecraft. A mission visit plan is generated using the optimal permutation, and the spacecraft executes the plan to rendezvous with and inspect the two or more orbiting target objects.

The invention discloses an unmanned ship distributed cooperative dynamic target surrounding control method under multiple constraints, and relates to the field of unmanned ship cooperative control. In order to solve the problems that multiple constraint coupling is not fully considered in cooperative surrounding of multiple unmanned ships, target state estimation is insufficient in an intermittent communication environment and the like, the following technical scheme is provided: firstly, a motion model of the unmanned ships and a target is established; a distributed internal model observer is arranged for each unmanned ship, and the internal motion state of the target is estimated; then designing a distributed type progressive average surrounding guidance algorithm, and performing distributed scheduling on the expected speed and heading of the unmanned ship, so that the unmanned ship can still keep safe surrounding when the speed or direction of the target is changed; the input saturation constraint of a propeller or a steering engine is considered, so that each unmanned ship can be autonomously controlled according to an expected trajectory and speed; and finally, stable and collision-free surrounding of the dynamic target is realized under multiple constraint conditions. The method can be applied to scenes such as ocean escort and maritime search and rescue which need to perform collaborative operation around a dynamic target.

The invention discloses a mechanical arm path planning method based on Transform and a diffusion model, and belongs to the technical field of robots, and the method comprises the following steps: collecting environment information to generate a reference path, and constructing a training data set; a conditional diffusion Transform prediction network is constructed, features are extracted, and multi-modal path prediction is realized; gaussian noise is applied to the reference path, and denoising training is carried out on the conditional diffusion Transform prediction network; in combination with a diffusion model and a cost guidance mechanism, optimizing a noise path sampled from Gaussian distribution until a smooth collision-free path is generated; the candidate paths are evaluated, and a mechanical arm joint control instruction is generated; and the mechanical arm executes the planning track and carries out real-time sensing and online re-planning. According to the method, environment perception, Transform coding and diffusion generation are organically combined, a plurality of feasible tracks which are smooth and capable of avoiding obstacles are rapidly generated in a complex obstacle scene, and the method has good generalization ability and can adapt to different scenes and dimension changes.

The invention relates to the field of computing power center operation and maintenance, in particular to a computing power center multi-robot collaborative operation and maintenance system and method based on a multi-modal large model, and the method comprises the steps: collecting equipment operation data in real time; carrying out data preprocessing on the acquired multi-modal data; the preprocessed single-mode features are input into a multi-mode large model, a unified multi-mode feature vector is generated through an adaptive feature fusion algorithm, the equipment fault type, position and occurrence probability are predicted based on the fused features, and a multi-robot cooperative operation and maintenance strategy is generated; the cooperative control center distributes tasks to the optimal robot combination through a Hungary algorithm according to an operation and maintenance strategy; a reinforcement learning algorithm is adopted to plan a collision-free path for each robot, and real-time dynamic adjustment is carried out to cope with environmental changes; and the operation and maintenance robot executes inspection, fault positioning and basic maintenance tasks, and uploads execution data to the cooperative control center in real time. The complex operation and maintenance requirements of the computing power center are effectively met, and the operation and maintenance efficiency is improved.

The invention discloses a cooperative positioning and trajectory planning method based on an air-ground robot system, and the method comprises the steps: enabling unmanned vehicles to construct a global environment map through a laser radar, achieving the positioning of the unmanned vehicles, obtaining the distances between the unmanned vehicles through the combination of an IMU (Inertial Measurement Unit) and a UWB (Ultra Wideband) distance measurement chip of the unmanned vehicles, and carrying out the real-time pose estimation. Based on the fused pose and global map information, a method of combining mixed A * search and nonlinear trajectory optimization is adopted, a collaborative trajectory planning problem is converted into a multi-constraint optimization model, the sight distance quality between air-ground robots, an unmanned vehicle team-shaped geometric structure and an unmanned aerial vehicle rapid navigation problem are comprehensively considered, and the space-ground robot cooperative trajectory planning method is established. And generating a collision-free optimized cooperative trajectory meeting multi-robot kinematics constraints in real time. According to the method, the positioning precision and the coordination degree of the air-ground coordination system in a complex dynamic environment are remarkably improved, and a safe, efficient and real-time heterogeneous multi-robot trajectory planning method is provided for various application scenes such as routing inspection, search and rescue, environment monitoring and the like.

The invention provides a robot skill learning and control method and device oriented to rich contact operation tasks. The method provided by the invention comprises the following steps: collecting human demonstration data; the human demonstration data at least comprises end effector pose data and stiffness matrix data; on the basis of the end effector pose data and the stiffness matrix data, a generalization motion track and a generalization stiffness track are obtained through time alignment, probability distribution modeling and new state constraint adaptation; constructing an obstacle model through environmental perception, taking the generalized motion trajectory and the generalized rigidity trajectory as reference sampling candidate trajectories, screening a collision-free trajectory in combination with motion consistency, rigidity consistency and smoothness cost, and performing iterative optimization to obtain an optimized motion trajectory and an optimized rigidity trajectory; and converting the optimized motion track and the optimized rigidity track into a robot joint torque instruction based on variable impedance control, and controlling the robot to execute a rich contact operation task based on the robot joint torque instruction.

The invention relates to a distributed trajectory planning method and system for an unmanned aerial vehicle cluster in an obstacle environment. The method comprises the steps of initializing a system, constructing a local Euclidean symbol distance site map, and realizing real-time interaction of state information of the unmanned aerial vehicle. An initial collision-free path is generated using jump point search. And solving an optimal control point through an L-BFGS algorithm through multi-constraint trajectory optimization in combination with a load swing dynamics model, four-rotor dynamics limitation, cluster collision avoidance and environment obstacle avoidance requirements. And a dynamic time redistribution strategy is adopted, and the track time interval is adjusted according to speed and acceleration overrun conditions. The method further comprises an adaptive re-planning mechanism, and local target points are updated in real time and adjacent aircraft collaborative optimization is triggered based on local map boundary detection and quadrotor track safety detection. According to the method, efficient, safe and stable trajectory planning of the hanging load quad-rotor unmanned aerial vehicle cluster in a complex environment is realized, the requirement of autonomously and efficiently completing tasks is met, and the task execution efficiency and safety are improved.

The invention relates to a robot dynamic grabbing control method based on visual sense and tactile sense depth fusion. The method comprises the following steps that firstly, based on visual information, surface geometric features of a target object are extracted, grabbing adaptability scores are calculated, an optimal grabbing point is selected, and a collision-free approaching track is planned; secondly, in a contact establishment stage, detecting a contact event through a touch sensor, fusing vision and touch data to unify a coordinate system, calculating a vision-touch consistency score, and applying an initial grabbing force; and thirdly, in the stable holding stage, slip detection is conducted through wavelet packet energy entropy and pressure gradient, and the grabbing force and impedance parameters are dynamically adjusted in combination with self-adaptive impedance control. According to the method, intelligent control over the whole process from approaching to contact to stable holding is achieved, and the adaptability, stability and safety of robot grabbing in the dynamic environment are improved.

The invention discloses a non-standard weldment-oriented robot efficient collision-free path planning method and system, and belongs to the technical field of robot welding, and the method comprises the following steps: S1, obtaining three-dimensional model data of a non-standard weldment and an operation environment, and extracting welding seam point cloud data; s2, carrying out feature extraction and downsampling on the welding seam point cloud data to generate a group of key path points; s3, global optimization is conducted on the welding path through an improved genetic algorithm, candidate paths are generated, and the welding sequence of the multiple sections of welding seams and the welding gun postures at the key path points are optimized at the same time through the improved genetic algorithm; s4, evaluating the candidate paths generated by the improved genetic algorithm through a multi-target fitness function including path length, attitude stability and collision risk; and S5, outputting an optimal collision-free welding path according to an evaluation result, and converting the optimal collision-free welding path into a smooth track which can be executed by the robot. According to the invention, the high-efficiency, high-quality and collision-free welding gun motion trail can be automatically generated.

A method for operating an adjustment system for an interior of a motor vehicle, wherein the adjustment system has motor-adjustable interior elements, which are adjustable between different configurations by respective drive arrangements with actuators via adjustment kinematics, wherein a control arrangement is provided, by which the drive arrangements are activated in an adjustment routine in order to adjust the motor-adjustable interior elements from an initial configuration into an end configuration via the adjustment kinematics. The control arrangement has an obstacle representation of objects in the interior for collision testing during the adjustment. A path planning routine is carried out by the control arrangement, in which a collision-free adjustment path from the initial configuration into the end configuration is determined on the basis of a kinematics model of the adjustment kinematics and of the obstacle representation, and that the activation in the adjustment routine is carried out by the control arrangement.

The invention discloses a pull-type robot blind guiding system and method, and relates to the technical field of mobile service robots, the pull-type robot blind guiding system comprises a robot body, the robot body comprises an environment sensing and recognition module and a path planning module, the environment sensing and recognition module is used for obtaining three-dimensional point cloud information of the surrounding environment, and the path planning module is used for planning the three-dimensional point cloud information of the surrounding environment; generating an environment composite map according to the three-dimensional point cloud information; the path planning module analyzes the environment composite map by introducing artificial potential field factors, and generates a safe and collision-free initial path for the blind; the robot body is physically connected with the blind person through the rigid guide rod. Multi-sensor coupling is introduced, and artificial potential field factors are added, so that the obstacle avoidance capability of the blind guiding robot in a complex environment is improved, safe walking of the blind in indoor and outdoor scenes is ensured, and the life autonomy of the blind is remarkably improved.

The invention discloses a truck-mounted crane group operation safety monitoring system and method based on multi-source perception and collaborative decision, and particularly relates to the field of safety monitoring, the truck-mounted crane group operation safety monitoring system comprises a distributed perception module, an edge calculation module, a graded early warning and execution module, a digital twin platform module and a man-machine interaction module; collected multi-source data are fused through Kalman filtering and then transmitted to an edge computing node; generating a dynamic safety envelope region based on an adaptive kinematics model, predicting a group conflict risk through a graph theory algorithm, and calculating collision time; the digital twin platform fuses a BIM model and real-time data, hoisting path rehearsal, load swing prediction and foundation bearing capacity visualization are carried out, an improved RRT * algorithm is adopted to plan a collision-free path, and a Lagrange dynamics model is adopted to predict a swing track; the grading early warning module triggers differential response according to the risk grade; and risk warning and path guidance are provided through the AR-HUD and the three-dimensional situation billboard.

The invention discloses a virtual assembly and anti-collision simulation method for a double-swing-head five-axis numerical control machine tool, and the method comprises the following steps: S1, building a geometric model of the double-swing-head five-axis numerical control machine tool based on a machine tool kinematic model, and the geometric model comprises parameterized three-dimensional models of a main shaft part, a swing head mechanism, a workbench and a machine tool body; s2, according to the constraint relation of the kinematic chains of the machine tool, an assembly hierarchical relation is established in the virtual environment, and virtual assembly of all parts is achieved through dynamic constraint driving; s3, on the basis of a machine tool kinematics forward and inverse solution algorithm, simulating the motion trail of the double swing heads in the five-axis linkage process; s4, through a real-time collision detection algorithm, detecting interference conditions among the cutter, the swing head, the workpiece, the clamp and the machine tool body, and generating an anti-collision early warning signal; and S5, optimizing a motion path of the machine tool in combination with a simulation result, and outputting a collision-free machining code. The assembly precision can be improved, and the physical debugging cost can be reduced.

The invention relates to a double-arm intelligent collaborative tea picking method and device based on multi-strategy dynamic scheduling, and belongs to the field of robot automation. The method comprises the following steps: obtaining three-dimensional coordinate information of a target picking point; identifying and positioning potential branch and leaf shields, and generating space data of the shields; the method comprises the following steps: simultaneously projecting a shielding object and a clamping jaw geometric contour under an expected grabbing pose on a plane where a target picking point is located, analyzing the overlapping amount, the proximity and / or the density of the projections of the shielding object and the clamping jaw geometric contour, and generating a standardized shielding evaluation index; comparing the shielding evaluation index with a preset decision threshold, and if the shielding evaluation index is greater than the preset decision threshold, executing a master-slave collaborative picking method based on multi-stage positioning; otherwise, executing the parallel picking method based on dynamic collision-free region division. The problems that a traditional single-arm tea picking robot is low in operation efficiency, high in collision risk and insufficient in intelligent degree in a dense and unstructured tea tree environment are solved.

The invention provides a coffee machine mechanical arm acceleration compensation control method and device based on toppling spill prevention, and relates to the technical field of mechanical arm control. A coffee machine calls a real-time acceleration sequence from an acceleration control library after receiving a real-time order, and a pipetting mechanical arm moves along a pipetting collision-free track in the process of clamping an order coffee cup to move along the pipetting collision-free track; the liquid gravity center offset is measured in real time through a six-axis force sensor integrated on a clamping jaw, and lagging shake offset of a real-time acceleration sequence is carried out according to the real-time gravity center offset. The technical problem that in the prior art, when a mechanical arm of a coffee machine transports coffee drinks containing heterogeneous liquid, the displacement control stability is insufficient, and pouring and scattering are likely to happen is solved. The technical effects that the coffee liquid is effectively prevented from spilling and leaking in the transportation process, the transportation stability of complex coffee drinks containing ice blocks, foam and the like is remarkably improved, safety and high efficiency of collaborative operation of the multiple mechanical arms are ensured, and the delay or overflow risk caused by liquid shaking is greatly reduced are achieved.

The invention discloses a global reference line generation method for an automatic driving vehicle and a related device, and relates to the technical field of automatic driving navigation, and the method comprises the steps: generating a collision-free original path of the vehicle based on an improved A * algorithm according to a starting point position and a terminal point position of the vehicle in a grid map; removing redundant points in the collision-free original path of the vehicle to obtain an original reference path; numerical smooth optimization is carried out on discrete points of the original reference path, and a cubic spline interpolation method is adopted to generate a basic reference line meeting the geometric continuity of the vehicle; establishing a Frenet coordinate system based on the basic reference line, and carrying out transverse and longitudinal motion primitive sampling to obtain a vehicle kinematics constraint and a boundary constraint; according to the constraint, establishing a quadratic programming multi-objective optimization function; and solving the quadratic programming multi-objective optimization function to obtain a global reference line. According to the invention, the robustness and security of reference line generation in a complex scene can be improved.

The invention provides a collision-free global path planning method for a mobile robot, belongs to the field of mobile robot path planning, and is used for solving the problems of intention misjudgment, low role allocation efficiency and high collision risk in a dynamic scene caused by dependence on statistical association in related technologies. According to the method, agent intention reasoning is achieved by constructing a mental atlas containing a causal relationship, efficient role allocation is performed based on a quantum role superposition state, path search is optimized in combination with a conflict risk field, role collapse is triggered when a conflict risk exceeds a threshold value, and cooperation of adjacent agents is guaranteed through a quantum entanglement cooperation state. According to the scheme, intention prediction accuracy and role allocation efficiency are improved, and collision-free and efficient global path planning in a dynamic environment is realized.

The invention relates to the technical field of AGV intelligent transportation, and discloses a lyocell fiber raw material intelligent transportation method and system based on AGV trajectory planning, and the method comprises the steps: obtaining a collision-free path of an AGV from an initial position to a target position through employing a self-adaptive target guidance RRT random search algorithm; deleting redundant nodes of the collision-free path, and fitting the path by adopting an NURBS curve to obtain a fitted path; segmenting the fitting path according to the curvature value of the fitting path and a preset segmentation number; calculating the speed constraint of each segment point in combination with the resultant gravity center and curvature value of the AGV and the carried raw material; and finally, performing acceleration and deceleration planning on each segmented path when the AGV moves along the fitting path by using an S-shaped speed curve to obtain motion control parameters of the AGV, and controlling the AGV to transport raw materials along the fitting path. According to the method, the optimal track planning of the AGV for transporting the Lyocell pulp raw material can be realized, the path accuracy and the transportation stability of the trolley for transporting the large-weight pulp raw material are ensured, and the continuity of Lyocell production is ensured.

The invention relates to an accurate collision-free automatic docking control method which is characterized by comprising the following steps: S1, remote guidance: triggering an automatic docking signal, establishing a three-dimensional semantic map, completing autonomous positioning, planning a docking path, performing real-time obstacle avoidance navigation, and executing safe and stable movement to reach a pre-docking position; s2, posture adjustment: when reaching the pre-docking position, determining the target posture of the target docking object and the posture of the object to be docked, and adjusting the posture of the object to be docked to the pre-docking posture; and S3, short-range docking: based on motion compensation, detecting a motion error, generating a motion correction instruction according to error signal feedback, and realizing collision-free accurate docking. The movement function and the body position changing function of the intelligent wheelchair are achieved through the intelligent wheelchair, smooth and stable switching can be achieved in the changing process of the intelligent wheelchair, and meanwhile the intelligent wheelchair is an indoor power assisting system, has the obstacle avoiding walking function and can be in collision-free automatic butt joint with a bed body and a closestool.

The invention relates to the technical field of mechanical arm trajectory planning, and particularly discloses a picking mechanical arm trajectory planning method and device, a terminal and a medium. The method comprises the steps that a three-dimensional scene model of the mechanical arm and the working environment of the mechanical arm is established; a collision detection model of the mechanical arm in the operation process is constructed, and a collision-free planned path is determined; adopting a quintic polynomial interpolation function to perform preliminary fitting on a joint track of the mechanical arm, and determining a quintic polynomial preliminary track function of the mechanical arm on the planned path; performing smooth optimization on the preliminary trajectory function of the planned path by using a cubic B-spline curve, and constructing a calculation trajectory function of the picking mechanical arm; and by taking the shortest movement time of the picking mechanical arm as an optimization target, optimizing the calculated trajectory function by using an improved particle swarm algorithm to obtain a picking mechanical arm optimized trajectory function with optimal time. According to the invention, high-precision trajectory planning of the picking mechanical arm in a complex agricultural operation environment is realized.

The invention discloses a quadrotor unmanned aerial vehicle autonomous flight path planning method and system, and the method comprises the steps: introducing a global trajectory updating mechanism to update a global trajectory in real time, and carrying out the re-planning of a generated guide trajectory, and obtaining a global trajectory; based on the generated global trajectory, generating a collision trajectory detection and obstacle avoidance path, and when the global trajectory passing through the obstacle is detected, searching a collision-obstacle-free path of which the starting point is a trajectory point entering the obstacle and the end point is a trajectory point passing through the corresponding obstacle by using A *; performing trajectory optimization on the generated collision-obstacle-free path to generate a safe and smooth trajectory, solving an optimization function and performing trajectory optimization based on a MINCO trajectory to generate an optimized trajectory; and taking the optimized trajectory meeting the dynamics feasibility as an expected trajectory for the unmanned aerial vehicle to execute the autonomous flight task. The problems of local path oscillation, high energy consumption and poor dynamic environment adaptability of the unmanned aerial vehicle are solved.

The invention relates to an unmanned aerial vehicle cluster safety path planning reinforcement learning method and device. The method comprises the following steps: inputting unmanned aerial vehicle group states, task targets and environment observation information into a deep reinforcement learning strategy of a continuous action space, and generating a global guidance quantity; an explicit security constraint set between the unmanned aerial vehicles and between the unmanned aerial vehicles and obstacles is constructed, through graph structure modeling and a GNN adaptive mechanism, the unmanned aerial vehicles / obstacles are represented by dynamic graphs, and neighborhood and relative position speed information is utilized to predict CBF parameters; and performing safety projection correction on the guide quantity based on the constraint and the CBF parameter to obtain an actual control instruction corresponding to the safety path. After the instruction is executed, feedback information such as state deviation and constraint tensity is collected and used for adjusting a reinforcement learning strategy, loop iteration is carried out to continuously output an optimization path, and whole-course collision-free planning of the cluster is achieved. By adopting the method, stable and reliable safety protection in a complex dynamic environment can be realized.

The invention relates to the technical field of mechanical arms, in particular to a chemical emergency disposal-oriented manipulator cross-modal sensing control system which comprises an environment scanning module, a visual positioning module, an obstacle avoidance planning module, a pose compensation module, a decision mapping module and a grabbing updating module. Obtaining environment state information; an operation view angle image of the manipulator is collected, visual features of the operation view angle image are solved, and pose information is obtained; obstacle avoidance correction is conducted on the motion trail of the mechanical arm, and a collision-free approaching path is obtained; the operation view angle image and the multi-mode micro-touch signal are fused, error compensation is carried out on the pose information, and an accurate grabbing pose is obtained; mapping object surface physical attributes sensed by the multi-mode micro-tactile signals to obtain a self-adaptive grabbing instruction; the manipulator is controlled to conduct grabbing through the self-adaptive grabbing instruction, and environment state information is updated; according to the invention, the accuracy of cross-modal sensing control of the manipulator can be improved.

The invention discloses a multi-station robot photographing task allocation method for hub surface defect detection, and the method comprises the steps: forming a hub triangular patch model, and carrying out the Gaussian mapping of the hub triangular patch model; finding out the task area boundary of each station robot based on a simulated annealing algorithm, wherein the included angle between the normal vector of the triangular patch in the same task area and the Z axis is in a set range; generating a camera viewpoint of each station robot for completing task area shooting, and adjusting the viewpoint with collision into a collision-free viewpoint; planning a local optimal path passing through all collision-free camera viewpoints based on a greedy algorithm; and carrying out collision detection on the local optimal path, and re-planning the road section with collision. According to the multi-station wheel hub defect detection method, the task area corresponding to each station is determined according to the wheel hub curved surface features by adopting the simulated annealing algorithm, finally, automatic balanced distribution of the multi-station robot operation tasks is achieved, collision detection is conducted on the generated shooting path, and the reliability of multi-station wheel hub defect detection is improved.