78 results about "Robot motion planning" patented technology

The invention relates to a planning method for a simulated path of a robot under a complex dynamic scene and a robot motion planning simulation platform for achieving the method. According to the method, path security evaluation criteria are established according to a collision possibility and an arrival possibility, an algorithm framework for planning two layers of interactive paths is adopted, motion planning is divided into an environment exploration layer and a partial path planning layer, the two layers perform information interaction by means of a self-adaptive path buffering area, and a security update search tree is taken as a top planner and applied to the environment exploration layer in the two-layer interactive framework. The simulation platform comprises a problem module, a planning module and an execution module. A safe and real-time path planning scheme is provided according to a path planning strategy selected by simulating a human path, the defect that only the feasibility of the path is considered but the continuity of the path is not concerned in the prior art is overcome, and the motion planning of the robot in a complex dynamic environment is realized.

The invention discloses a quadruped robot motion planning method oriented to an unknown rough terrain. A quadruped robot uses a crawling gait in a rough terrain. In a single-legged swing phase, a swinging leg swings along a planned foot trajectory and the center of gravity of the robot moves forwards along the planned path by means of the motion of supporting legs. In a four-foot stance phase, the center of gravity of the robot laterally moves while moving forwards along the planned path. The quadruped robot motion planning method achieves robot online trajectory planning based on COG and ZMP dual-stability criteria, takes full account of the possibility of advanced landing and lagged landing of the swinging leg for the unknown terrain, and estimates the parameter of the unknown terrain by the landing planning and the sensing strategy of the swinging leg. Thus, the robot adaptively adapts to the height and the slope of the unknown terrain, and achieves continuous adaptive walking for the unknown rough terrain according to a motion planning result.

The invention provides a quadruped robot motion planning method for facing complex terrain. According to moving object of a quadruped robot, and local environment information and terrain information which are detected by a sensor, movement direction at the current moment of the quadruped robot is confirmed and generate local moving object of the quadruped robot in Cartesian space. According to the local moving object of the quadruped robot in the Cartesian space and the terrain information detected by the sensor, a motion path sequence is generated in quadruped robot configuration space through a rapid random tree algorithm, and the quadruped robot moves according to the motion path sequence until reaches the local moving object. Above steps are repeated until the quadruped robot reaches a set target. Due to the fact that a body motion planning in the Cartesian space and each joint configuration planning in the configuration space are combined, and a motion path sequence of each joint of the quadruped robot is generated according to present environment of the quadruped robot, safe and stable voluntary movement of the quadruped robot in the complex terrain is ensured.

The invention discloses an RRT (Rapidly exploring Random Tree) algorithm based path planning method. On the basis of the classic algorithm, pruning operation is performed and tiny branches on a tree generated by RRT algorithm are cut off, so that the quantity of nodes travelled through by a finally generated path is reduced and the path bending frequency is reduced. Therefore, the path is more suitable for robot walking. The RRT algorithm is a path planning algorithm disclosed by LaValle in a paper for the first time in 1998 and is a classic algorithm frequently used in MDOF robot motion planning and plan path planning. However, length of paths generated by the RRT algorithm has comparatively large redundancy and the linearity of the paths is poor. Through the improvement implemented through pruning in the invention, path redundancy is reduced and the linearity is improved substantially.

The invention discloses a control method of an industrial robot demonstration planner with a motion planning function. The method comprises: establishing a new demonstration file or opening a demonstration file, setting the system parameters and demonstration parameters of a robot, controlling the robot in a robot task space to carry out hand-operated demonstration, and utilizing the programming instruction of a VAL-III language to carry out demonstration programming; performing robot kinetics inversion, converting a three-dimensional locus into a joint space locus, inserting a locus intermediate point in the joint space locus, planning a locus equation, and obtaining the locus equation of a joint space; carrying out time interval interpolation calculating, and obtaining a robot control sequence point; and sending the obtained robot control sequence point to a robot controller through an RS-485 serial port. According to the invention, a demonstration system and a motion planning system are integrally combined, a robot motion planning function is provided, and the demonstration working efficiency is high.

The invention discloses a robot motion planning method, a path planning method, a grabbing method and a grabbing device. The method comprises the following steps: 1, when the robot operates the to-be-operated object, the robot operates the to-be-operated object, combining a space model of a real scene where the object is located; performing collision detection on the object and a collision objectin the space model; according to the collision sensitivity of the object and the collision sensitivity of the collision object, determining a motion planning scheme of the robot formed by operating the object by the robot corresponding to a corresponding collision detection result; therefore, when the robot operates the corresponding object, the robot can move based on the motion planning scheme,the phenomenon that in the process that the robot operates the object, the robot gets stuck in the motion process is avoided, the motion of the robot is smoothened, and the intelligent degree of industrial operation of the robot is improved.

The invention discloses an extension and solution set dual neural network solution for robot motion planning. The extension and solution set dual neural network solution for robot motion planning comprises the steps of acquiring a current state of a robot through a sensor, and adopting a quadratic optimization scheme for carrying out inverse kinematics analysis on a robot track on a speed layer; converting a quadratic optimization scheme of a minimum speed two-norm index to a standard quadratic planning problem; converting the quadratic planning problem to a solution of a Karush-Kuhn-Tucker optimality condition; utilizing an extension and solution set dual neural network solver for solving; and transferring a solved result to a robot controller, and driving a robot body to carry out trajectory tracking. According to the extension and solution set dual neural network solution for robot motion planning provided by the invention, through designing nonlinear equality constraint, convex setconstraint and non-convex set constraint can be compatible, a preliminary test error problem occurred in robot control is eliminated, and an error accumulation problem during a robot control processis solved.

The invention discloses a robot motion planning method for searching to identify multi-objective tasks by facing man-machine collaboration. The robot motion planning method includes that judgment about whether maximum working load is exceeded or not is performed when a task set is changed, a fixed time interval is not adopted, unnecessary waiting time is reduced, and efficiency is improved; when the judgment about whether the maximum working load is exceeded or not is performed, judgment about whether working load is exceeded or not is performed according to time that moving robots reach suspected target observation points planned by the robots, and the judgment mode is quite simple and efficient. When paths are re-planned for the robots, time for an operator to complete target identification is needed; the time for the operator to complete target identification task is uncertain, so that target identification task completion time of the operator is created as random models, probability distribution function of the target identification task completion time of the operator is established in advance, and rationality of path planning is guaranteed.

The invention discloses a quadruped robot motion planning method based on hierarchical reinforcement learning. According to the method, an upper layer behavior decision controller based on deep reinforcement learning and a lower layer motion execution controller based on model prediction control are constructed; for a deep reinforcement learning network at an upper layer, a state updating network, an action execution network, a reward function and the like of the quadruped robot are designed based on an SAC algorithm, and corresponding behavior control parameters are output in combination with height map information of the environment around the robot; and for a model prediction controller in the lower layer execution controller, a corresponding control instruction is obtained by solving a control parameter quadratic programming equation set input based on the upper layer. A hierarchical control framework combines the advantages of the deep reinforcement learning and model prediction control, so that the quadruped robot can make a safer and more reliable behavior mode according to the current fuselage state and the terrain environment, a fuselage posture is kept stable, risks are effectively avoided, and effective motion track planning under the complex terrain is achieved.

The invention relates to the technical field of robot motion planning, in particular to an end path planning algorithm for a mortar spraying and plastering construction robot. The end path planning algorithm comprises the following steps of 1) task trajectory planning: determining task trajectory points of a mechanical arm execution end moving from a task initial position to a task ending position, solving a dynamics inverse solution set aiming to each trajectory point, traversing all inverse solutions, and finding a path with the highest traversing success rate as a final task trajectory planning; and 2) free trajectory planning: calculating a joint displacement of a current posture of a mechanical arm, performing calculation by utilizing the task trajectory planning to obtain a joint displacement of an initial posture, and establishing a path planning of the mechanical arm execution end moving from a current position to the task initial position. By adopting the technical scheme, themotion trajectory from an arbitrary current position of the mechanical arm to a task operation path of the mechanical arm can be quickly and accurately planned, the execution precision is high, and the requirements for cement mortar spraying and plastering tasks can be met.

The invention provides a multi-robot motion planning method and system and a storage medium. The method comprises the following steps: acquiring the motion state and environment information of robots in motion; determining a motion strategy of the robot through reinforcement learning according to the discretized motion state and environment information, wherein the continuous action in the motion strategy is obtained by determining a reinforcement learning state space by using a fuzzy neural network and outputting the obtained continuous action; determining a basic behavior of the robot according to the motion strategy, and performing cluster motion, wherein the basic behaviors comprise at least one of the following behaviors: advancing to a target, obstacle avoidance movement, collision avoidance movement and movement along a wall. The method solves the problems that the convergence speed is too low and the predictive ability is poor when optimal behavior strategy learning is carried out on the intelligent body in a huge state space and a dynamic change environment, and can be widely applied to the technical field of robot control.

The invention discloses a robot motion planning method based on DMPs and a corrected obstacle avoidance algorithm. The obstacle avoidance effect can be improved, the DMPs can be assisted in better finishing a motion planning task, and according to the technical scheme, the method comprises the following step that a dynamic motion element DMPs method is adopted for planning a barrier-free path froma starting position to a target position for a mechanical arm of a robot as an expected trajectory. The corrected obstacle avoidance algorithm serves as a coupling term to be added into a DMPs second-order system to be used for generating an obstacle avoidance trajectory, the generated obstacle avoidance trajectory is used for serving as the motion trajectory of the robot, and preferably, the corrected obstacle avoidance algorithm comprises one or combination of multiple of an improved steering behavior method, an improved dynamic approach method and an improved dynamic obstacle avoidance method.

The invention relates to a method for avoiding kinematics singularity points in robot motion planning. The method comprises the following steps that a kinematic model of an industrial robot is established; a kinematics inverse solution of path points of a pre-planning path is obtained according to the kinematics model, and a kinematics track of a joint space of the industrial robot is establishedaccording to the obtained kinematics inverse solution; when it is determined that the kinematics track passes through spherical wrist kinematics singularity points, tiny deflection is carried out on the angle of a starting point of the pre-planning path to obtain a new planning path, and then a new kinematics track of the joint space of the industrial robot is established according to the new kinematics inverse solution; when it is determined that the new kinematics track still passes through the spherical wrist kinematics singularity points, the pre-planning path is abandoned to execute; andwhen it is determined that the new kinematics track does not pass through any spherical wrist kinematics singularity point, the new planning path is executed. The invention further relates to a devicefor avoiding the kinematics singularity points in the robot motion planning.

The invention discloses a redundant robot repeated motion planning method adopting a rapid double-power final state neural network. The method comprises the following steps that a desired trajectory r-d(t) of a robot end effector is given in cartesian space, an expected return angle theta -d(0) of each joint is given; for the repeated motion of a robot, an asymptotic convergence performance indexis adopted, a problem is solved by converting an quadratic optimization problem of redundant robot trajectory planning into a time-varying matrix equation, and the rapid double-power final state neural network is used as a solver. Under the condition of initial position offset, a repeated motion planning task for rapid finite-time convergence of the redundant robot is achieved. According to the provided redundant robot motion planning method adopting the rapid double-power final state neural network, rapid finite-time convergence is achieved, calculation precision is high, and implementation is easy.

The invention provides a robot motion planning method and system, computer equipment and a storage medium. The method comprises the steps that a time-based gait plan is obtained, and gait informationand phase information of a robot are detected when the robot performs the time-based gait plan; the time-based gait plan is input into a Kalman filtering detector, and an estimated contact state is predicted after the robot performs the time-based gait plan; the gait information and the phase information are input into the Kalman filtering detector for detection to obtain an actual contact state of the robot; and an event-based gait plan for robot motion is triggered according to the estimated contact state and the actual contact state. According to the method, an unknown complex environment is sensed by a robot body, the event-based gait plan is triggered according to the actual perception and the prediction amount after perception, and gaits are adjusted and corrected in time, so that the robot walks autonomously in a complex and unknown terrain by self-adaptive adjustment based on ontology perception, and stable walking of the quadruped robot in the unknown terrain is guaranteed.

The invention discloses an indoor following robot system. The indoor following robot system comprises a structure subsystem, a voice interaction subsystem, a sensing subsystem, an intelligent planningsubsystem and an omnidirectional maneuvering control subsystem, wherein the voice interaction subsystem, the sensing subsystem, the intelligent planning subsystem and the omnidirectional maneuveringcontrol subsystem are constructed on the structure subsystem; the man-machine voice interaction subsystem is used for acquiring voice information; the sensing subsystem is used for acquiring sensing information including RGB-D images and obstacle position information; the intelligent planning subsystem is used for determining a working mode of the robot and performing follow-up motion control planning by combining the voice information and the sensing information; and the omnidirectional maneuvering control subsystem is used for realizing maneuvering control of the robot according to the following motion control plan. According to the system, deep learning and Kalman filtering are adopted to realize human body target detection and tracking, robot motion planning under obstacle avoidance isrealized based on an artificial potential field method, and good following performance of the robot for human bodies with various postures is ensured; the system is simple in structure, low in cost and high in expansibility, and can serve as a universal basic platform for development of other multi-purpose indoor following robots.

The invention discloses a motion control system for a oblique wrist spray robot and a method thereof. The motion control method includes the steps that: reading each joint value of the oblique wrist spray robot, and recording the Cartesian space information of the oblique wrist spray robot through a positive kinematics solution, and storing the Cartesian space information; during the execution ofthe work process, solving the trajectory by analyzing the stored Cartesian space information and obtaining a feasible solution of the joint value of each joint by an inverse solution algorithm, and further according to the stored Cartesian space information, the solution is filtered to obtain the only suitable inverse solution result, and then the oblique wrist spray robot is controlled to move tothe target position corresponding to the inverse solution result. According to the motion control method, the problem of the joint space and the Cartesian space vector mapping of the oblique wrist spray robot can be solved, the high precision and real-time performance of the robot motion planning process can be ensured, and the inverse kinematics of the end oblique wrist without closed solution can be solved.

The invention discloses a quadruped robot motion planning method for a narrow environment. The method comprises the following steps: constructing a local point cloud map according to surrounding environment information detected by an airborne sensor of a quadruped robot, and converting the local point cloud map into a trafficability cost map containing environmental terrain information; determining a local target point of motion planning in a local range according to a global guide path input by an external global planner; based on a trafficability cost map and a local target point, planning a rough optimal collision-free path on the cost map by using a mass point model of the quadruped robot and using an A * algorithm; performing trajectory optimization on a motion dynamics model of the quadruped robot by using the rough optimal collision-free path to generate an effective motion trajectory meeting the motion dynamics constraint of the robot; a rolling time domain control strategy is used, and a motion control instruction of the quadruped robot is calculated through the effective motion trail; and judging whether the robot ends the execution track to reach the target position, if not, returning to the step 1 to repeat the whole process until completion.

The invention discloses a robot curve path generation method, system and device and a storage medium. The robot curve path generation method comprises the following steps that data points are acquired, and the data points are subjected to sequential marking; two data point groups are acquired according to the marked sequence; after the two data point groups are subjected to curve fitting, and twocurves are obtained; after the obtained two curves are spliced according to a preset mode, and a splicing curve is obtained; the next group of data points are acquired according to the marked sequence, a new curve is obtained after the next group of data points are subjected to curve fitting, and the return operation is carried out to execute the step S4, until all the data points are fitted, anda curve path of a robot is obtained. According to the robot curve path generation method, system and device and the storage medium, the data points are divided into multiple segments, then the multi-segment curves are spliced in sequence, so that the curve path of the smoother robot is obtained, the robot curve path generation method, system and device and the storage medium are suitable for freecurve planning with known undetermined number of points or large number of points, the robot curve path generation method, system and device and the storage medium has the advantages of being highly reusable, and the robot curve path generation method, system and device and the storage medium can be widely applied to the technical field of robot motion planning and control.

The invention discloses a robot motion planning method and system based on a graph Wasserstein self-encoding network. The method comprises the following steps of S1, constructing the graph Wasserstein self-encoding network (GraphWAE); S2, conducting non-uniform sampling distribution representation learning based on the GraphWAE; and S3, conducting robot motion planning based on the GraphWAE. Compared with the prior art, the robot motion planning method and system have the beneficial effects that the non-obstacle area in the robot configuration space is represented and learned through the GraphWAE, and serves as a sample generator of a mainstream sampling planning algorithm, the planning exploration process is guided to be carried out in the non-obstacle area, the planning time is shortened, and the planning path quality and the success rate are improved.