Tensegrity Robot Joint With Learning-Based Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current collaborative robotic systems face challenges with bulky robot arms that have poor haptic and ergonomic properties due to rigid links and rotary joints, leading to resistive forces and safety issues, while tensegrity structures offer promising characteristics but require advanced modeling and control to achieve effective dexterous end-effector manipulation.
Innovation Solution
A compact, low-inertia, and reduced-friction soft tensegrity robot joint with two degrees of freedom, combined with a high-level data-driven machine-learning-based controller using tendon length encoding and skin shape sensing for real-time closed-loop control, allowing for safe interaction and robustness against disturbances.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Strength
If conventional rigid links and rotary joints with motors and gearboxes are used, then the robot can provide sufficient force and structural stability, but the robot arms become bulky with high inertia and actuation friction, leading to poor haptic properties and safety issues
Solution Approach 1:
The patent replaces conventional rotary joints with motors, gearboxes, and brakes with a tendon-driven tensegrity mechanism. This substitution eliminates heavy mechanical components while maintaining force capability through tension-based actuation, directly resolving the contradiction between strength and weight.
Solution Approach 2:
The patent extracts and removes the motor, brake, harmonic drive, and gearbox from the joint structure. By taking out these heavy components and replacing them with tendon actuators and tension-based constraints, the robot arm achieves reduced mass and inertia while preserving force generation capability.
2Strength
If tendon preloading is applied to increase structural stiffness, then the robot can maintain shape and resist deformation, but axial compression on backbone links and contact joints increases, requiring larger pulling forces to drive robot motion
Solution Approach 1:
The patent changes the fundamental actuation parameter from tendon preloading (compression-based) to active tendon tensioning (tension-based). This parameter change allows the system to achieve structural stiffness through controlled tension rather than passive preloading, eliminating the need to overcome axial compression forces during motion.
Solution Approach 2:
The patent inverts the conventional tendon actuation approach by using active tensioning instead of passive preloading. Rather than relying on preloaded tendons that create compressive forces, the system uses actively controlled tension to achieve both stiffness and motion, reversing the traditional force application method.
3Weight of moving object
If tensegrity structures are used to eliminate friction and reduce inertia, then the robot achieves compact size and low-inertia operation, but the system requires complex kinematics modeling and sophisticated actuation sensing with low tolerance to actuation errors
Solution Approach 1:
The patent implements self-service through the learning-based controller that automatically adapts to the tensegrity joint's nonlinear dynamics. The system uses real-time sensing feedback and machine learning to compensate for modeling uncertainties and actuation errors, eliminating the need for complex manual kinematics modeling and sophisticated sensing systems.
Solution Approach 2:
The patent employs feedback through the learning-based controller that continuously monitors joint configuration and adjusts tendon actuation accordingly. This closed-loop control with machine learning compensates for the inherent sensitivity to actuation errors, reducing the complexity of kinematics modeling and sensing requirements.
4Measurement precision
If learning-based control with tendon length encoding and skin shape sensing is implemented, then the robot achieves precise path following and disturbance rejection, but the system requires sophisticated sensing and computational resources for real-time control
Solution Approach 1:
The patent applies preliminary action through offline training of the learning-based controller using simulated physics data. By pre-training the neural network model with extensive simulation data before deployment, the system achieves precise path following and disturbance rejection without requiring complex real-time sensing and computation during actual operation.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The tensegrity robot joint achieves precise path following with low RMSE and resistance to external disturbances, demonstrating improved manipulation repeatability and safety without preloading tendons, and can vary stiffness to handle varying loads and impulsive events.
Implementation Method 1
an outer elastic protection sheath configured to wrap on a cylindrical surface between the first and second inner structures
Implementation Method 2
a central tendon connecting an apex of the first strut and an apex of the second strut such that the central tendon is in balance with the pulling forces exerted by the at least three actuation tendons
Data Source
AI summary
The present invention provides a tensegrity joint comprising: a first inner structure having a first frame and a first strut extending orthogonally from the first frame; a second inner structure having a second frame and a second strut extending orthogonally from the second frame; at least three actuation tendons, each being fixed at a respective hole at the first frame and being guided and allowed to slide through a respective hole at the second frame; a central tendon connecting an apex of the first strut and an apex of the second strut such that the central tendon is in balance with the pulling forces exerted by the at least three actuation tendons; and an outer elastic protection sheath configured to wrap on a cylindrical surface between the first and second inner structures.