38 results about "Robot planning" patented technology

Systems, methods, and user interfaces are used for controlling a robot. An environment map and a robot designator are presented to a user. The user may place, move, and modify task designators on the environment map. The task designators indicate a position in the environment map and indicate a task for the robot to achieve. A control intermediary links task designators with robot instructions issued to the robot. The control intermediary analyzes a relative position between the task designators and the robot. The control intermediary uses the analysis to determine a task-oriented autonomy level for the robot and communicates target achievement information to the robot. The target achievement information may include instructions for directly guiding the robot if the autonomy level indicates low robot initiative and may include instructions for directing the robot to determine a robot plan for achieving the task if the autonomy level indicates high robot initiative.

Embodiments of the invention comprise a system and method that enable robotic harvesting of agricultural crops. One approach for automating the harvesting of fresh fruits and vegetables is to use a robot comprising a machine-vision system containing rugged solid-state digital cameras to identify and locate the fruit on each tree, coupled with a picking system to perform the picking. In one embodiment of the invention a robot moves through a field first to “map” the field to determine plant locations, the number and size of fruit on the plants and the approximate positions of the fruit on each plant. A robot employed in this embodiment may comprise a GPS sensor to simplify the mapping process. At least one camera on at least one arm of a robot may be mounted in appropriately shaped protective enclosure so that a camera can be physically moved into the canopy of the plant if necessary to map fruit locations from inside the canopy. Once the map of the fruit is complete for a field, the robot can plan and implement an efficient picking plan for itself or another robot. In one embodiment of the invention, a scout robot or harvest robot determines a picking plan in advance of picking a tree. This may be done if the map is finished hours, days or weeks before a robot is scheduled to harvest, or if the picking plan algorithm selected requires significant computational time and cannot be implemented in “real time” by the harvesting robot as it is picking the field. If the picking algorithm selected is less computationally intense, the harvester may calculate the plan as it is harvesting. The system harvests according to the selected picking plan. The picking plan may be generated in the scout robot, harvest robot or on a server. Each of the elements in the system may be configured to communicate with each other using wireless communications technologies.

A robot control apparatus allows a plurality of mobile robots to carry out tasks at a reduced or minimized cost as a whole with consideration given to costs derived from an encounter with an obstacle is provided. An action optimization controller provided in the robot control apparatus generates an instruction for optimizing actions of the plurality of mobile robots so that the plurality of mobile robots carry out the tasks at a minimized cost, based upon locomotion plan information indicative of locomotion plans of the plurality of mobile robots. A possibility that any robots have an encounter with an obstacle is determined by comparing distances from the robots to the obstacles. A locomotion plan implementation cost is calculated with consideration given to the possibility of encounter, and thus an optimum route is selected based upon the locomotion plan implementation cost with the encounter-derived cost. In accordance with the optimum route, the locomotion plan for the robot is modified.

The invention provides a mobile robot path planning method based on an improved ant colony algorithm. The improved ant colony algorithm with sixteen search directions and twenty-four neighbor nodes is proposed, so that ants have more direction selections and wider search range in a robot planning path searching process, and two heuristic information computing methods based on a vector angle are designed; moreover, the effective search range of the ants in a raster map is controlled by introducing control parameters of transition probability in a transition probability part, an optimal global shortest path can be searched, the search effectiveness of the algorithm can be improved, and the randomness of search is reduced. According to the method, the optimal global shortest path can be searched through the improved ant colony algorithm, the search effectiveness of the algorithm can be improved, the randomness of search is reduced, and a good condition is provided for robot path planning.

The invention discloses a mobile robot full-coverage traversal chaotic path planning method. The method comprises the following steps: carrying out basic design of a mobile robot full-coverage traversal chaotic path; designing a distance evaluation index of adjacent sub-target points in a robot path planning locus; performing operation sub-space division on a work space of a robot; performing affine transformation on a chaotic kinetic equation in operation subspaces; designing a mixed iteration strategy; and carrying out iterative computation based on a given initial point. First of all, a basic chaotic plan planning method is designed, in order to enable the designed chaotic path planning method to be more feasible, then the mixed iteration strategy combining iteration of a whole work area with iteration of the sub-spaces divided by the work area is employed, traversal of the work space is realized, and the chaotic feature of the robot planning locus is maintained basically unchanged, such that the iteration step pitch of the mobile robot is effectively reduced and effective tracking of a robot controller is realized.

The invention relates to a scene reconstruction method and device, computer equipment, and a computer storage medium. The method comprises the steps: obtaining a photographed depth image, and generating a current tensor field key frame according to the depth image; carrying out the spatial and temporal interpolation of the current tensor field key frame and a former tensor field key frame, and obtaining an interpolation tensor field key frame; obtaining a robot planning path through a particle drift algorithm; obtaining a camera planning path according to the robot planning path and an energyequation of a camera; and reconstructing a scene through the depth images collected when a robot walks along the robot planning path and the camera walks along the camera planning path. According to the invention, the camera planning path is obtained through the robot planning path and the energy equation of the camera, thereby achieving a purpose of enabling the robot which holds the camera to carry out the exploration and reconstruction of a complex scene at the same time.

The invention discloses a robot remote operating system and method based on VR. The system comprises an operation robot, a camera shooting device, a VR head display device, a robot planning device, aremote control device and a background server. The method comprises the following steps that firstly, a virtual robot and a VR virtual scene are constructed; secondly, the camera shooting device is utilized for obtaining video images of the operation environment and an operation target in real time; thirdly, the video images and the VR virtual scene are synchronously overlapped; fourthly, action planning is conducted on the virtual robot in the virtual scene; fifthly, the operation robot is controlled to execute the actions; sixthly, the robot actions are corrected according to sensor data; and seventhly, log data in the fourth step, the fifth step and the sixth step are stored, and deep learning is conducted. The system and method have higher flexibility, safety and precision, manual workcan be replaced for electrification work, electrification work intelligence is achieved, the electrification work efficiency is improved, and personal risks of related operators are reduced.

The invention discloses a cleaning robot optimal path planning method based on approximate model multi-step optimization. The method comprises steps: a model, hyperparameters and an environment are initialized; an exploration strategy is selected; the model is updated with the current sample; a value function, a strategy and the current state are updated; whether the current plot is ended is judged and further, the current sampling trajectory is used to update a trajectory pool; a reconstructed sampling trajectory is used to update the trajectory pool; all trajectories in the trajectory pool are used to update the model; the model is adopted for planning; whether to reach the maximum plot number is judged; and finally, according to a learnt optimal strategy, the optimal cleaning robot planning path is acquired. Through adopting the sampling trajectory and the single sample to update the model, the model learning precision is improved; the model is then used for planning, the learning speeds of the value function, the strategy and the whole algorithm are improved; and the sample utilization efficiency is also improved, and thus, the optimal cleaning robot planning path is acquired by adopting the fewer samples in the shorter time.

Parallel linguistic corpora used to train cognitive translation systems are built from robot plans. A collection of robot plans in various languages is analyzed in candidate pairings to see if any two plans are directed to the same task. If so, verbal instructions in different languages from the two plans (such as plan names) are assumed to have the same meaning, and are stored in association as part of the bilingual corpus. If each plan has multiple steps, the corresponding steps from the two plans can also be stored in association as part of the corpus. Robot plans are modeled as finite state automata having a plurality of links representing sensory states and a plurality of nodes representing motor actions, and the determination of plan similarity is based on a comparison of the two FSAs.

The invention belongs to the field of robot task planning, and provides a robot time sequence task planning method which comprises the following steps: acquiring environment information of a working area and performing discretization modeling on the working area; obtaining an individual time sequence task requirement and a cooperative time sequence task requirement of the robot; searching a collaborative task sequence according to collaborative time sequence task requirements; constructing a collaborative subtask allocation model according to the collaborative task sequence, and solving a collaborative subtask allocation result; and sending the collaborative sub-task allocation result to the robot, so that the robot individually calculates a task execution plan. According to the method, the problem of multi-robot planning under the constraint of individual and global complex time sequence tasks can be efficiently solved, the method has good expansibility, and the limitation of unstable regional communication can be overcome; in addition, the privacy information of the individual robot can be effectively protected while the calculation load is dispersed.

The invention discloses a robot control method based on ground socket detection, a chip and a robot. The control method specifically comprises the steps that ground socket distribution information is acquired; according to the ground insert distribution information, judging whether a ground insert exists in the robot planning motion route area or not; if the ground insert exists in the robot planning motion route area, obtaining first ground insert information corresponding to the nearest ground insert passed by the robot in the robot planning motion route area; and controlling the robot to re-plan the motion route according to the first ground insert information. The motion route is planned according to the first ground insert information obtained in real time, and the smoothness of the motion route of the robot is reasonably improved on the basis under the condition that it is guaranteed that the robot is not stuck by the ground insert.

The invention discloses an arbitration planning system and method, and belongs to the technical field of robot control. The arbitration planning method comprises the following steps of receiving, by the arbitration planning system, a plan command input from the external and converting to form an executing plan; simultaneously, receiving, by the arbitration planning system, external environment information sent by a sensing system; performing, by the arbitration planning system, conflict arbitration judgment on the executing plan and the external environment information; if a conflict is not generated, handing the executing plan and the external environment information to a control unit to generate a corresponding control command, issuing the control command to a control system of a robot,and controlling, by the control system, motor units of various movable joints of the robot to execute the control command; and if the conflict is generated, returning to the arbitration planning system and performing re-planning, wherein the re-planned executing plan is different from an originally executed plan. The technical scheme has the beneficial effects that an arbitration problem when theconflict is generated between a planned action of the robot and the external environment can be solved through the arbitration planning system and the arbitration planning method thereof, so that a predetermined executing plan is executed smoothly; the robot can complete a established action; and the work efficiency of the robot is improved.