Robot Sensitivity via Virtual Torque and Jacobian Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wearable muscle assistive robots face challenges in accurately transmitting the wearer's intention without using specific sensors, as EMG sensors can distort signals and torque sensors reduce robot durability and are costly.
Innovation Solution
A method that calculates angular velocities of joints using encoders, converts these into velocities and accelerations using a Jacobian matrix, and amplifies forces to reflect the wearer's intention as torque at the joints without separate sensors, employing a virtual spring-damper model and low-pass filtering to enhance sensitivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If EMG sensors are used to detect wearer intention, then the robot can transmit force, but the signals are distorted when the wearer moves
Solution Approach 1:
The patent extracts the intention detection function from EMG sensors and implements it through software-based processing of encoder signals. By removing the physical EMG sensors and using mathematical models to interpret joint angle data, the system eliminates signal distortion issues while maintaining intention detection capability
Solution Approach 2:
The patent replaces the biological signal detection system (EMG sensors on muscle) with a mechanical measurement system (encoders on joints). This substitution uses mechanical angle measurements combined with mathematical modeling to infer intention, avoiding the signal distortion problems of electrical sensors during movement
2Measurement precision
If torque sensors are mounted on driving joints to measure wearer intention, then the intention can be reflected accurately, but the robot durability decreases and cost increases
Solution Approach 1:
The patent extracts the torque measurement function from physical torque sensors and implements it through software calculation. By removing torque sensors and using mathematical models that compute torque from joint angle and acceleration data, the system maintains accurate intention detection while eliminating the durability issues associated with physical sensors
Solution Approach 2:
The patent creates a virtual model that copies the function of torque sensors. Instead of measuring torque directly with physical sensors, the system calculates equivalent torque values from encoder data through mathematical modeling, providing the same information without the physical hardware
3Measurement precision
If torque sensors are used to measure driving joint torque, then the wearer intention can be discriminated, but expensive sensors are required
Solution Approach 1:
The patent replaces expensive torque sensors with inexpensive encoder sensors. The encoder data is processed through mathematical models to provide torque information, using cheap hardware to achieve the same functional result as expensive dedicated sensors
Solution Approach 2:
The patent replaces the expensive mechanical torque sensing system with a cheaper optical or magnetic encoder system combined with computational algorithms. This substitution maintains measurement precision while significantly reducing hardware cost
Data Source
AI summary
A method of improving sensitivity of a robot which includes: a calculation step, an induction step and a conversion step. The calculation step calculates angular velocities of joints of a robot. The induction step determines induced accelerations at the end of the robot by converting the angular velocities of the joints into a velocity at the end of the robot, using a Jacobian matrix, and by differentiating the velocity. The conversion step determines forces at a middle portion of the robot by multiplying the induced accelerations at the middle portion of the robot by a weight of the robot, multiplies the forces by an enhancement ratio, and then converts results of the multiplication into necessary torque at the joints, using a Jacobian matrix.