Walking Robot Balance Control via Ankle Torque

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing walking robot control methods, such as ZMP and FSM, face challenges in maintaining balance efficiently, with ZMP requiring high energy and causing unnatural walking due to high servo gain, and FSM being inaccurate and requiring complex dynamic equations, especially for robots with 6 degrees of freedom.

Innovation Solution

A walking robot equipped with sensors to detect ground contact and pose, using a walking controller to generate an ankle reference angle and compensate for pose to maintain balance without solving complex dynamic equations, allowing stable walking on inclined surfaces.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If ZMP control method is used to achieve accurate position control, then position control accuracy is improved, but energy consumption increases and joint rigidity increases causing shock to surroundings

Engineering Contradiction:
Improveposition control accuracyVSAvoidenergy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent replaces the traditional ZMP position-based control method with a torque-based control method. Instead of controlling joint positions to satisfy ZMP constraints, the system controls joint torques directly based on operation states, substituting the mechanical position control system with a torque control system that achieves balance without high servo gain and energy consumption.

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

Solution Approach 2:

The patent changes the control parameter from joint position (ZMP method) to joint torque (FSM method). By switching to torque-based control, the system can maintain balance and stability without requiring high servo gain, thereby reducing energy consumption and joint rigidity while avoiding shock to surroundings.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If ZMP control method is used to achieve accurate position control, then position control accuracy is improved, but joint rigidity increases giving surroundings a big shock

Engineering Contradiction:
Improveposition control accuracyVSAvoidjoint rigidity
Core Design Contradiction:
Measurement precisionVSStrength

Solution Approach 1:

The patent replaces the traditional ZMP position-based control method with a torque-based control method. Instead of controlling joint positions to satisfy ZMP constraints, the system controls joint torques directly based on operation states, substituting the mechanical position control system with a torque control system that achieves balance without high servo gain and energy consumption.

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

Solution Approach 2:

The patent changes the control parameter from joint position (ZMP method) to joint torque (FSM method). By switching to torque-based control, the system can maintain balance and stability without requiring high servo gain, thereby reducing energy consumption and joint rigidity while avoiding shock to surroundings.

Inventive Principle:
Principle #35Parameter changes

3Use of energy by moving object

If FSM control method is used to achieve low energy consumption and high safety, then energy efficiency is improved, but walking control accuracy decreases causing the robot to lose balance

Engineering Contradiction:
Improveenergy efficiencyVSAvoidwalking control accuracy
Core Design Contradiction:
Use of energy by moving objectVSMeasurement precision

Solution Approach 1:

The patent incorporates feedback mechanisms to monitor the robot's actual walking state and compares it with the desired operation state. Based on the deviation detected by feedback, the control system adjusts the torque commands in real-time, ensuring accurate walking control while maintaining the energy efficiency and safety benefits of torque-based control.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent defines operation states in advance and prepares corresponding torque commands before execution. By pre-defining the desired operation states and calculating the necessary torque commands in advance, the system can execute smooth and accurate transitions between states, improving walking control accuracy while maintaining energy efficiency.

Inventive Principle:
Principle #10Preliminary action

4Strength

If FSM control method is used to achieve low servo gain and high safety, then joint rigidity is reduced providing safety to surroundings, but walking control accuracy decreases causing the robot to lose balance

Engineering Contradiction:
Improvejoint rigidityVSAvoidwalking control accuracy
Core Design Contradiction:
StrengthVSMeasurement precision

Solution Approach 1:

The patent incorporates feedback mechanisms to monitor the robot's actual walking state and compares it with the desired operation state. Based on the deviation detected by feedback, the control system adjusts the torque commands in real-time, ensuring accurate walking control while maintaining the energy efficiency and safety benefits of torque-based control.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent defines operation states in advance and prepares corresponding torque commands before execution. By pre-defining the desired operation states and calculating the necessary torque commands in advance, the system can execute smooth and accurate transitions between states, improving walking control accuracy while maintaining energy efficiency.

Inventive Principle:
Principle #10Preliminary action

5Measurement precision

If complicated dynamic equation is solved to obtain command torque for balancing, then balance control accuracy is improved, but device complexity increases

Engineering Contradiction:
Improvebalance control accuracyVSAvoidcontrol system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts the balance control function from the complex dynamic equation solving process. By separating balance control into a distinct module that uses simplified principles (ankle joint control based on ground contact detection and pose information), the system achieves accurate balance control without requiring solution of complicated dynamic equations, thereby reducing overall system complexity.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the control system into distinct functional modules: walking control module, balance control module, and sensor processing module. The balance control is handled separately through ankle joint control based on ground contact detection, independent from the complex dynamic equation solving, thereby simplifying the overall control architecture while maintaining balance accuracy.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP2343163B1Walking robot and method of controlling balance thereof
Publication Date: 2017.08.09 SAMSUNG ELECTRONICS CO LTD
  • EP2343163B1 patent drawingFigure 1
  • EP2343163B1 patent drawingFigure 2
  • EP2343163B1 patent drawingFigure 3

AI summary

Disclosed herein are a walking robot which controls balance using an ankle when the robot walks, and a method of controlling balance thereof. In a method of determining an angle of an ankle joint without solving a complicated dynamic equation such that the robot stays balanced so as not to fall, an angle of the ground is fixed as a reference angle for balance control of the robot such that the robot stably walks while maintaining the same balance control performance even when the ground is inclined. When the robot moves slowly or quickly, the robot may maintain balance. Since the robot stays balanced using the ankle of a stance leg even when the ground is inclined, the method is simple and is applied to a robot having joints with 6 degrees of freedom.