Humanoid Robot Walking Control With ZMP Error Compensation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing humanoid robot walking control methods, such as ZMP-based and FSM-based methods, face challenges in achieving stable walking due to zero moment point (ZMP) errors, leading to inefficient energy use, high joint rigidity, and potential collisions with obstacles, while FSM-based methods struggle with balance and energy wastage during balancing operations.

Innovation Solution

A humanoid robot system that includes a ZMP calculation and measurement unit, a compensation value calculation unit, and a compensation unit to adjust joint position and torque commands based on the difference between calculated and measured ZMPs, ensuring stable walking by compensating for ZMP errors and optimizing joint movements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If ZMP-based walking control method is used to accurately control joint position and control ZMP, then manufacturing precision and reliability are improved, but energy consumption increases and joint rigidity increases causing high impact on obstacles

Engineering Contradiction:
Improvejoint position control accuracyVSAvoidenergy efficiency
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent changes the control parameter from position-based to torque-based control. Instead of controlling joint positions to follow predetermined trajectories, the system calculates torque commands that directly achieve the desired ZMP while allowing more natural joint movement, thereby reducing energy consumption and joint rigidity

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent replaces the position servo control mechanism with a torque control mechanism. By directly controlling joint torques rather than positions, the system achieves ZMP control with lower energy consumption and more natural robot movement patterns

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If ZMP-based walking control method is used to accurately control joint position, then reliability is improved, but joint rigidity increases causing great impact on obstacles

Engineering Contradiction:
Improvewalking stabilityVSAvoidimpact on obstacles
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent changes from position control to torque control, allowing joints to be more compliant. This reduces the rigidity that causes high impact forces when colliding with obstacles, while maintaining walking stability through accurate torque-based ZMP control

Inventive Principle:
Principle #35Parameter changes

3Adaptability or versatility

If FSM-based walking control method is used to take various poses by changing operation state, then adaptability is improved, but balance control becomes complex and energy is wasted due to separate balancing operations

Engineering Contradiction:
Improvepose variation capabilityVSAvoidenergy waste in balancing operations
Core Design Contradiction:
Adaptability or versatilityVSLoss of energy

Solution Approach 1:

The patent merges the balance control function with the walking control function. Instead of performing separate balancing operations (like stamping feet) in addition to walking, the torque-based control simultaneously achieves both walking and balance maintenance, eliminating redundant energy-consuming operations

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The robot's natural dynamics and torque control enable self-balancing during walking without requiring additional active balancing operations. The system uses the robot's own movement and dynamics to maintain balance, rather than requiring separate corrective actions

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP2347866B1Humanoid robot and walking control method thereof
Publication Date: 2013.01.16 SAMSUNG ELECTRONICS CO LTD
  • EP2347866B1 patent drawingFigure 1
  • EP2347866B1 patent drawingFigure 2
  • EP2347866B1 patent drawingFigure 3

AI summary

Disclosed herein are a humanoid robot that compensates for a zero moment point (ZMP) error during finite state machine (FSM)-based walking to achieve stable walking and a walking control method thereof. The humanoid robot compensates for a joint position trajectory command or a joint torque command using compensation values calculated based on situations divided according to the position of a calculated ZMP and the position of a measured ZMP in a stable region of the robot.