Multi-Purpose Robot Work Primitives for Simulated Autonomous Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robots require complex tele-operation systems and significant training through real-world wear and tear, limiting their accessibility and efficiency in completing multiple work objectives autonomously.
Innovation Solution
A robot system equipped with a library of reusable work primitives and processor-executable instructions that allows it to autonomously initiate and execute workflows for various work objectives, using a combination of reusable work primitives and a telecommunication interface for instruction receipt, enabling semi-autonomous operation and reducing training needs through simulated environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If robots are trained through repeated real-world physical tasks, then the robot can autonomously perform work objectives, but significant wear and tear occurs on robot components before deployment
Solution Approach 1:
The patent creates virtual copies of the robot and its work environment in a simulated training environment. The robot's physical actions, sensor inputs, and task sequences are replicated in virtual form, allowing extensive training to occur without physical wear. The virtual training data is then used to program the actual robot for autonomous operation.
2Ease of operation
If tele-operation systems are used to control robots, then the robot can perform physical actions emulating human operators, but the interface becomes elaborate and complicated requiring full operator attention
Solution Approach 1:
The patent extracts the complex control interface from the final deployment scenario and concentrates it in the training phase. Operators perform tasks in the virtual environment during training, and their actions are recorded and converted into autonomous control sequences. This eliminates the need for complex real-time tele-operation interfaces during actual work execution.
Solution Approach 2:
The robot is programmed to serve itself by executing pre-recorded task sequences autonomously. Once trained, the robot independently performs work objectives without requiring continuous human intervention or complex operator interfaces, achieving self-service operation.
3Adaptability or versatility
If robots are programmed to operate semi-autonomously or fully autonomously, then accessibility improves, but training through real-world repetition causes wear and tear
Solution Approach 1:
The patent creates virtual replicas of training scenarios where the robot learns autonomous operation sequences. These virtual copies allow unlimited repetition of training tasks without physical degradation, enabling the robot to achieve high-level autonomous capabilities while preserving component integrity.
Data Source
AI summary
Systems, devices, and methods for training and operating (semi-)autonomous robots to complete multiple different work objectives are described. A robot control system stores a library of reusable work primitives each corresponding to a respective basic sub-task or sub-action that the robot is operative to autonomously perform. A work objective is analyzed to determine a sequence (i.e., a combination and/or permutation) of reusable work primitives that, when executed by the robot, will complete the work objective. The robot executes the sequence of reusable work primitives to complete the work objective. The reusable work primitives may include one or more reusable grasp primitives that enable(s) a robot's end effector to grasp objects. Simulated instances of real physical robots may be trained in simulated environments to develop control instructions that, once uploaded to the real physical robots, enable such real physical robots to autonomously perform reusable work primitives.