Robot Motion Trajectory Prediction Without Proprietary RCS Modules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic motion simulation systems face challenges in accurately predicting the motion trajectory of robots without Realistic Robot Simulation (RRS) modules, especially for medium and small-sized industrial robot vendors, and are hindered by high licensing costs and slow external client-server communication of proprietary RCS modules.
Innovation Solution
A method involving machine learning to generate a motion prediction module using training data from various sources, including physical and simulated robotic motions, to accurately predict the motion trajectory of specific robots without the need for RCS modules, reducing licensing costs and improving data communication speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If proprietary RCS modules with RRS protocol are used for accurate robotic motion simulation, then motion prediction accuracy is improved, but licensing costs increase and external client-server communication speed decreases
Solution Approach 1:
The patent extracts the essential motion prediction functionality from the proprietary RCS module and implements it as a standalone machine learning model. The system processes training data to obtain x tuples (robotic location pairs) and y tuples (intermediate locations with time stamps), then learns a function mapping that can predict robot motion independently of the original RCS module, eliminating the need for slow external client-server communication.
Solution Approach 2:
The patent creates a copy of the motion prediction capability by training a machine learning model on motion trajectory data. Instead of relying on the proprietary RCS module, the system learns from training data to replicate the motion prediction function, achieving both accuracy and independence from the original expensive system.
2Measurement precision
If proprietary RCS modules with RRS protocol are used for accurate robotic motion simulation, then motion prediction accuracy is improved, but licensing costs increase
Solution Approach 1:
The patent replaces the expensive proprietary RCS module with a cost-effective machine learning-based motion prediction system. By training on available motion trajectory data and creating an independent prediction model, the system eliminates licensing costs while maintaining the core functionality of accurate robot motion simulation.
3Reliability
If generic motion planners are used for robots without RCS modules, then licensing costs are reduced, but motion prediction accuracy deteriorates
Solution Approach 1:
The patent applies preliminary action by training a machine learning model on motion trajectory data before actual motion prediction is needed. The system processes training data to obtain tuples and learns the motion function in advance, so that when the robot needs motion prediction, the already-trained model can provide accurate results without requiring expensive proprietary modules during operation.
4Adaptability or versatility
If training data from multiple sources is processed to generate motion prediction module, then support for broad range of robot vendors is improved, but data processing complexity increases
Solution Approach 1:
The patent achieves universality by creating a machine learning-based motion prediction system that can handle data from multiple sources including physical robot motions and simulated robotic motions. The system processes training data from various vendors and robot types, learning a generalizable function mapping that works across different robot platforms without requiring vendor-specific proprietary modules.
Data Source
AI summary
Systems and a method for predicting a motion trajectory of a robot moving between a given pair of robotic locations. Training data of motion trajectories of the robot are received for a plurality of robotic location pairs. The training data are processed so as to obtain x tuples and y tuples for machine learning purposes; wherein the x tuples describe the robotic location pair and the y tuples describe one or more intermediate robotic locations at specific time stamps during the motion of the robot between the locations of the location pair. From the processed data, a function is learned for mapping the x tuples into the y tuples so as to generate a motion prediction module for the robot. For a given robotic location pair, the robotic motion between the given pair is predicted by obtaining the corresponding intermediate locations resulting from the motion prediction module.