Joint-Space MPC for Smooth Robot Handover Trajectories
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic systems face challenges in performing smooth, intuitive, and reliable handover actions with humans, often resulting in unexpected contact or obstruction of cameras due to jerky movements and inadequate path planning.
Innovation Solution
A fast, reactive, and safe predictive robot control system using joint-space model-predictive control (MPC) with parallel processing, incorporating constraints to optimize motion paths over multiple time steps, and utilizing learning-based grasp planners to ensure smooth and predictable handovers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional robot control systems are used for handover actions, then the system structure is simple, but the motion is jerky and unreliable causing unexpected contact with humans
Solution Approach 1:
The system performs preliminary planning of the entire motion trajectory and computes future motions in advance using model predictive control. By planning the sequence of motions before execution and preparing for potential human movements ahead of time, the robot achieves smoother and more reliable handover actions without reactive delays.
Solution Approach 2:
The control system is designed to be dynamic by continuously updating the motion plan based on predicted human movements and current system state. The model predictive control framework allows the robot to adapt its trajectory in real-time while maintaining overall plan coherence, resolving the contradiction between reliability and complexity.
2Ease of operation
If traditional motion control is used, then the control algorithm is simple, but the motion smoothness and intuitiveness are poor
Solution Approach 1:
The system computes the entire sequence of future motions in advance using model predictive control, optimizing for smoothness and intuitiveness before execution. This preliminary optimization of the motion sequence ensures smooth transitions and natural-looking movements without requiring complex real-time control algorithms during execution.
Solution Approach 2:
The motion planning maintains continuous and smooth trajectories by optimizing the entire motion sequence rather than discrete steps. The model predictive control ensures continuous adjustment of motion parameters while maintaining smoothness, eliminating jerky movements without excessive algorithmic complexity during runtime.
3Speed
If reactive control is implemented, then the response to human movement is fast, but the computational load increases significantly
Solution Approach 1:
The system performs computationally intensive motion planning and prediction in advance, preparing multiple potential response trajectories before human movement occurs. When human movement is detected, the system only needs to select from pre-computed options rather than performing full optimization, achieving fast response with reduced real-time computational energy.
Solution Approach 2:
The computational load is dynamically managed by performing heavy calculations offline or in advance, then using lighter real-time updates based on predicted human behavior. This dynamic allocation of computational resources maintains fast response speed while controlling energy consumption during actual handover execution.
4Object-affected harmful factors
If simple path planning is used, then the computational requirements are low, but camera occlusion and contact with humans cannot be avoided
Solution Approach 1:
The system performs preliminary collision and occlusion analysis during the motion planning phase, identifying and avoiding trajectories that would cause harmful contacts or camera occlusions before execution. By pre-computing safe paths that account for human movement predictions, the system eliminates harmful factors without requiring complex real-time detection and avoidance algorithms.
Solution Approach 2:
The model predictive control framework incorporates feedback from predicted human states and sensor data to continuously refine the motion plan. This feedback mechanism ensures that the robot maintains safe distances and avoids camera occlusions by adjusting the pre-planned trajectory based on actual human movement, resolving the contradiction between safety and planning complexity.
Data Source
AI summary
Approaches presented herein provide for predictive control of a robot or automated assembly in performing a specific task. A task to be performed may depend on the location and orientation of the robot performing that task. A predictive control system can determine a state of a physical environment at each of a series of time steps, and can select an appropriate location and orientation at each of those time steps. At individual time steps, an optimization process can determine a sequence of future motions or accelerations to be taken that comply with one or more constraints on that motion. For example, at individual time steps, a respective action in the sequence may be performed, then another motion sequence predicted for a next time step, which can help drive robot motion based upon predicted future motion and allow for quick reactions.