Humanoid Robot Balance and Motion Tracking Control Framework
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Programming humanoid robots to perform natural, human-like motions is challenging due to difficulties in mapping human motion capture data to robotic kinematics and dynamics, with existing methods failing to simultaneously track motion capture data and maintain balance, especially under disturbances.
Innovation Solution
A control framework comprising a balance controller and a tracking controller that uses a simplified model, such as an inverted pendulum, to provide balancing control signals and generate motion commands that track reference motion while maintaining balance, allowing all joints to contribute to both tasks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If motion capture data is used to program robot motions, then natural human-like motions can be achieved, but the robot cannot simultaneously track the motion data and maintain balance
Solution Approach 1:
The control system is segmented into two independent controllers: a balance controller that maintains stability using a simplified inverted pendulum model, and a tracking controller that follows motion capture data. This segmentation allows each controller to specialize in one function while working together through coordinated joint torques, resolving the contradiction between tracking natural motions and maintaining balance.
Solution Approach 2:
The system changes parameters by using a simplified model for balance control (inverted pendulum) rather than the full complex model, and separately handles full-body dynamics for tracking. This parameter separation enables the robot to maintain balance while tracking motion capture data by adjusting torques based on both simplified and full dynamics considerations.
2Use of energy by moving object
If manual programming or numerical optimization is used to minimize energy consumption, then energy efficiency can be improved, but programming natural motions becomes more difficult
Solution Approach 1:
The tracking controller acts as an intermediary that takes motion capture data and transforms it into joint torques that satisfy both tracking requirements and energy efficiency considerations. By using the full-body dynamics model in the tracking controller, the system can optimize energy consumption while maintaining natural motions, avoiding the need for complex manual programming.
3Stability of the object's composition
If a simplified model is used for balance control, then balance stability can be improved, but tracking precision of motion capture data may deteriorate
Solution Approach 1:
The control system segments the modeling approach: a simplified inverted pendulum model is used specifically for balance control to ensure stability, while a full-body dynamics model is used for tracking motion capture data to maintain precision. This segmentation allows each model to be optimized for its specific purpose without compromising the other.
Solution Approach 2:
The system merges the outputs of two controllers: the balance controller using the simplified model and the tracking controller using the full model. By combining their torque commands, the system achieves both balance stability and motion tracking precision simultaneously, resolving the contradiction between using simplified and detailed models.
4Measurement precision
If all joints are used for tracking motion, then tracking accuracy improves, but balance maintenance becomes more difficult
Solution Approach 1:
The control system segments the function of joints between two controllers: the balance controller manages balance stability using a simplified model that considers overall body dynamics, while the tracking controller uses all joints to track motion capture data. This segmentation allows all joints to contribute to tracking accuracy while the balance controller ensures stability is maintained through coordinated torque adjustment.
Solution Approach 2:
The balance controller provides feedback on the robot's balance state and adjusts joint torques accordingly. This feedback mechanism allows all joints to be used for tracking while the balance controller continuously monitors and corrects any balance deviations, maintaining both tracking accuracy and balance stability.
Data Source
AI summary
Various embodiments of the invention provide a control framework for robots such that a robot can use all joints simultaneously to track motion capture data and maintain balance. Embodiments of the invention provide a framework enabling complex reference movements to be automatically tracked, for example reference movements derived from a motion capture data system.