Robot Trajectory Retargeting for Real-Time Whole-Body Motion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robotic devices face challenges in navigating complex environments with difficult terrain and unforeseen obstacles, requiring improved agility and dynamic whole-body control while conserving computational resources.
Innovation Solution
A system that pre-computes 'template behaviors' offline and adapts them in real-time using kinematic and environmental data to determine 'retargeted trajectories', allowing robots to efficiently navigate obstacles by combining pre-computed movements with real-time adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the robot performs detailed nonlinear trajectory optimizations with densely sampled data online, then the movement precision and adaptability improve, but the computational resource consumption increases significantly
Solution Approach 1:
The system pre-computes template trajectories offline with detailed nonlinear optimizations and dense sampling, storing them for later use. During online operation, the robot selects and adapts pre-computed templates rather than performing full optimizations, dramatically reducing computational energy while maintaining trajectory precision through the use of high-quality pre-prepared templates
Solution Approach 2:
The system changes the sampling density parameter from dense (offline) to sparse (online), and adjusts the time horizon from long (offline) to short (online). This parameter adaptation allows the robot to use computationally intensive methods when energy is abundant (offline) and efficient methods when energy constraints apply (online), resolving the contradiction between precision and energy consumption
2Adaptability or versatility
If the robot computes full retargeted trajectories online with real-time adaptations, then the adaptability to environmental changes improves, but the computational time and resource requirements increase
Solution Approach 1:
The robot pre-computes a library of template trajectories covering various movement types (walking, running, jumping, parkour) offline. During real-time operation, it selects the most appropriate template and applies minimal adaptations, achieving fast real-time responsiveness without the computational burden of full online trajectory optimization
Solution Approach 2:
The system dynamically adjusts the level of computation based on operational context: using pre-computed templates for standard situations and applying incremental adaptations only when necessary. This dynamic approach balances adaptability requirements with computational time constraints, allowing the robot to respond quickly to environmental changes without excessive computation
3Productivity
If the robot uses sparsely sampled data and short time horizons for online trajectory computation, then the computational efficiency improves, but the accuracy and detail of movement trajectories decrease
Solution Approach 1:
The system performs the computationally intensive task of generating accurate, densely sampled trajectories offline before operation. The pre-computed templates serve as high-precision references that guide online decision-making, allowing the robot to maintain trajectory accuracy while using efficient sparse sampling and short horizons during real-time operation
Solution Approach 2:
The pre-computed template trajectories act as an intermediary between offline planning and online execution. They transfer the computational burden from online to offline operation while providing accurate reference paths that guide real-time adaptations, effectively decoupling the requirements for high accuracy (satisfied by dense offline computation) and high efficiency (satisfied by sparse online computation)
Data Source
AI summary
Systems and methods for determining movement of a robot about an environment are provided. A computing system of the robot (i) receives information including a navigation target for the robot and a kinematic state of the robot; (ii) determines, based on the information and a trajectory target for the robot, a retargeted trajectory for the robot; (iii) determines, based on the retargeted trajectory, a centroidal trajectory for the robot and a kinematic trajectory for the robot consistent with the centroidal trajectory; and (iv) determines, based on the centroidal trajectory and the kinematic trajectory, a set of vectors having a vector for each of one or more joints of the robot.