Upper Limb Rehab Robot Control Using sEMG and Game Theory
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current rehabilitation robots lack flexible and adaptive control methods to accurately estimate human movement intention, leading to discontinuous control inputs and potential safety issues during mode switching, especially in resistance training, which is crucial for progressive muscle strengthening.
Innovation Solution
An adaptive control method for upper limb rehabilitation robots based on game theory and surface Electromyography (sEMG) that uses a Back Propagation Neural Network (BPNN) to establish a nonlinear dynamic relationship between sEMG signals and end force, allowing for adaptive mode switching and real-time adjustment of training intensity based on human movement intention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional force sensors are used to measure force during rehabilitation training, then the measurement is straightforward, but the signal-to-noise ratio is low and the time delay is high compared to sEMG signals
Solution Approach 1:
The patent replaces the mechanical force sensor system with a biological signal-based estimation system. Specifically, it uses sEMG signals combined with a musculoskeletal model and neural network to estimate joint torque and end force, substituting direct mechanical measurement with a computational estimation approach that leverages the temporal advantage of electrical muscle signals.
2Adaptability or versatility
If hard switching between multiple controllers is performed during rehabilitation training, then different training modes can be implemented, but the control input becomes discontinuous causing violent jitter and safety issues
Solution Approach 1:
The patent implements dynamic mode switching where the robot transitions between auxiliary training mode and resistance training mode based on real-time sEMG signal analysis and estimated human joint torque. The switching is governed by a state machine that continuously evaluates muscle activation levels and adjusts the control mode smoothly, avoiding abrupt transitions and ensuring control continuity throughout the rehabilitation process.
Solution Approach 2:
The system employs continuous feedback from sEMG sensors to monitor muscle activation in real-time. This feedback drives the adaptive control algorithm that determines when to switch between training modes, creating a closed-loop system that responds to the patient's physiological state and maintains safe, continuous control input during mode transitions.
3Extent of automation
If sEMG signals are used to estimate joint torque for adaptive control, then the movement intention can be accurately identified, but the system complexity increases due to signal processing and model requirements
Solution Approach 1:
The patent introduces a musculoskeletal model as an intermediary between the sEMG signals and the control system. The model includes muscle force-length-velocity relationships and moment arms to translate electrical muscle signals into estimated joint torque. This intermediary layer processes the complex physiological relationships while presenting a simplified interface to the control algorithm, managing system complexity through structured modeling.
Solution Approach 2:
The system creates a computational copy of the human musculoskeletal system through a virtual model that replicates muscle mechanics, joint dynamics, and force generation characteristics. This digital twin allows the control system to estimate internal physiological states from external sEMG measurements without requiring direct access to joint torque sensors, simplifying the physical system while maintaining computational accuracy.
Data Source
AI summary
An adaptive control method and system for an upper limb rehabilitation robot based on a game theory and surface Electromyography (sEMG) is disclosed. A movement trajectory that a robot is controlled to run within a training time is designed during subject operation. An sEMG-based Back Propagation Neural Network (BPNN) muscle force estimation model establishes a nonlinear dynamic relationship between an sEMG signal and end force by constructing a three-layer neural network. A human-computer interaction system is analyzed by the game theory principle, and a role of the robot is deduced. The control rate between the robot and a subject is updated by Nash equilibrium, and adaptive weight factors of the robot and the subject are determined. The robot adaptively adjusts the training mode thereof according to a movement intention of the subject during operation and a weight coefficient obtained by the game theory principle.