Humanoid Robot Joint Acceleration Compensation for Stable Fast Walking

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing walking gait control methods for humanoid robots face challenges in trajectory tracking at high speeds due to increased angular acceleration, leading to significant tracking errors and balance issues.

Innovation Solution

An acceleration compensation method that calculates angular acceleration and torque for each joint based on angular velocity and six-dimensional acceleration of the centroid, using a Jacobi matrix and Newton-Euler inverse kinematics, and superimposes a feedforward current value on the PID control signal to enhance trajectory tracking.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If walking speed is increased, then productivity is improved, but trajectory tracking precision deteriorates due to larger angular acceleration

Engineering Contradiction:
Improvewalking speedVSAvoidtrajectory tracking precision
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent calculates and compensates for angular acceleration in advance before it causes trajectory tracking errors. By pre-computing the acceleration compensation amount based on the relationship between angular velocity and acceleration, the system proactively counteracts the expected tracking deviations that would occur at higher walking speeds, thus maintaining precision while improving productivity

Inventive Principle:
Principle #10Preliminary action

2Productivity

If angular acceleration is increased to achieve higher walking speed, then productivity is improved, but control reliability deteriorates as PID controller can no longer realize required trajectory tracking

Engineering Contradiction:
Improvewalking speedVSAvoidtrajectory tracking reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent implements a feedback mechanism where the actual angular velocity is continuously measured, and the acceleration compensation amount is dynamically adjusted based on the relationship between angular velocity and acceleration. This closed-loop approach ensures that even at high walking speeds with large angular accelerations, the trajectory tracking remains reliable by continuously compensating for acceleration effects

Inventive Principle:
Principle #23Feedback

3Productivity

If PID control parameters are adjusted to handle higher acceleration, then trajectory tracking at high speed is improved, but balance ability deteriorates due to large trajectory tracking errors

Engineering Contradiction:
Improvewalking speedVSAvoidbalance ability
Core Design Contradiction:
ProductivityVSStability of the object's composition

Solution Approach 1:

The patent changes the control approach by introducing acceleration compensation as an additional control parameter. Instead of merely adjusting PID parameters, the system computes a compensation amount based on the angular acceleration (derived from angular velocity) and adds it to the PID control output. This parameter enhancement allows the controller to maintain both trajectory accuracy and balance at higher walking speeds

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS11179855B2Acceleration compensation method for humanoid robot and apparatus and humanoid robot using the same
Publication Date: 2021.11.23 UBTECH ROBOTICS CORP LTD
  • US11179855B2 patent drawing
  • US11179855B2 patent drawing
  • US11179855B2 patent drawing

AI summary

The present disclosure provides an acceleration compensation method for a humanoid robot as well as an apparatus and a humanoid robot using the same. The method includes: calculating an angular acceleration of each joint and calculating a six-dimensional acceleration of a centroid of a connecting rod corresponding to the joint in an absolute world coordinate system, if the humanoid robot is in a single leg supporting state; calculating a torque required by the angular acceleration of each joint of the humanoid robot; determining a feedforward current value corresponding to the torque of each joint; and superimposing the feedforward current value on a control signal of each joint to control the humanoid robot. In this manner, the influence of the acceleration can be effectively suppressed, the rigidity of the PID controller of the humanoid robot can be reduced, thereby improving the stability of the entire humanoid robot.