TL;DR
This paper introduces a novel wrist robotic exoskeleton interface (FC-WREI) that leverages a fully actuated coaxial spherical parallel mechanism (CSPM) to achieve a human-like range of motion, high torque output, and low moving part inertia. This design aims to overcome limitations of previous interfaces, enabling more natural and effective force interaction for applications like rehabilitation and teleoperation.
What Problem Does This Paper Address?
Robotic interfaces for physical interaction, particularly exoskeletons for the upper extremity, are crucial for applications like rehabilitation, training, and teleoperation. The wrist, being a highly dexterous and critical joint for daily living activities (ADLs), presents a significant challenge for robotic interface design. Existing wrist exoskeletons often struggle to balance several key requirements:
- Range of Motion (ROM): Mimicking the natural, multi-axis movement of the human wrist is essential for comfortable and effective interaction.
- Torque Output: Sufficient torque is needed to assist or resist human motion, especially for therapeutic or training purposes.
- Inertia: Low moving part inertia is critical for dynamic motion response, allowing the interface to keep pace with agile human movements and provide accurate force feedback.
- Ergonomics and Safety: Interfaces must avoid causing discomfort, pain, or injury, which can arise from misalignment with the human joint or kinematic singularities.
Previous parallel mechanisms often suffered from increased inertia due to actuators on moving links and limited workspaces. Serial mechanisms offered better workspace and alignment but could still have inertial issues. This paper aims to address these trade-offs by proposing a fully actuated, base-mounted parallel mechanism that inherently reduces moving part inertia while maintaining high torque and a wide ROM.
Methods
The core of this work is the development and analysis of a Fully Actuated Coaxial Spherical Parallel Mechanism (FC-WREI). Key methodological aspects include:
- Kinematic Analysis: Detailed forward and inverse kinematics were derived for the CSPM to understand its motion capabilities and relationships between joint angles and platform orientation. This involved analyzing the spherical linkage system and defining rotation matrices.
- Interference Safety Analysis: A metric called the Interference Safety Margin (ISM) was introduced to quantify the clearance between the robot's links and the user's forearm. This is crucial for ensuring safe operation and preventing collisions.
- Multi-Objective Optimization: A Non-dominated Sorting Genetic Algorithm II (NSGA-II) was employed to optimize the FC-WREI's design parameters (e.g., link angles, joint offsets). The optimization aimed to simultaneously maximize performance indices like the Global Condition Index (GCI) – which relates to dexterity and force isotropy – and the ISM, while adhering to constraints on local condition indices to avoid singularities.
- Actuator Capacity Estimation: Based on the optimized design and estimated human wrist torques and frequencies, the required actuator capacities (torque and power) were determined.
- Dynamic Simulation and Experimental Verification: The dynamic response was analyzed using MATLAB/SimMechanics. Experimental validation was conducted through motion control experiments (using PD control) and force control experiments (using an impedance controller). The performance of the FC-WREI was compared against similar devices.
Datasets
Not discussed in the paper. The paper focuses on the design, simulation, and experimental validation of a novel mechanism rather than training on large datasets.
Key Insights and Ideas
- Base-Mounted Actuators are Key to Low Inertia: By placing all actuators on the fixed base, the FC-WREI significantly reduces the inertia of its moving parts. This is a fundamental design choice that directly addresses the dynamic response limitations of many existing exoskeletons.
- CSPM for Natural Wrist Motion: The coaxial spherical parallel mechanism inherently provides pure rotational degrees of freedom, closely mimicking the human wrist's natural movement without translational components.
- Torque Amplification through Parallel Synthesis: The parallel nature of the mechanism allows for torque amplification, meaning the output torque can be greater than the torque generated by individual actuators, enhancing its capability for force interaction.
- Optimization Balances Performance and Safety: The multi-objective optimization successfully balanced competing goals: maximizing kinematic performance (GCI) and ensuring safety (ISM), while also considering the need to avoid singularities. This systematic approach is vital for robust robotic design.
- Compact Design Achieved via ISM Maximization: Maximizing the Interference Safety Margin allowed for a more compact design, further contributing to lower link inertia and a more streamlined interface.
Key Results
- Workspace: The FC-WREI achieved a workspace scope of 85.7–106.6% of the ADL range for the wrist, with specific ranges of ±30° (AA), ±50° (FE), and ±80° (PS).
- Torque Output: The device provides torque output that covers ADL torques and is 56.49–130.43% of the maximum isometric torque of the human wrist.
- Inertia: The FC-WREI exhibits significantly lower moving part inertia, up to 5.35 times lower than comparable devices (e.g., OpenWrist), with axial inertia values ranging from 0.0015 to 0.0057 kgm².
- Optimization Performance: The optimized design achieved a GCI of 0.701 and a maximized ISM of 0.43.
- Dynamic Response: The closed-loop bandwidth frequencies were found to be 5.42–11.83 Hz, exceeding the target of 5 Hz for dynamic motion.
- Actuator Torque Amplification: The system demonstrated torque amplification, with peak end-effector torque to actuator torque ratios of up to 3.63 times for a single actuator.
Limitations and Future Work
- User Acceptance: The paper acknowledges that the ring-shaped design might cause aversion in users due to fear of being unable to perform emergency evasion. Future work should investigate user acceptance.
- User-Test Evaluation: The force interaction interface needs to be evaluated through user testing to fully understand its practical application and effectiveness.
- Manufacturing Clearance: While kinematic analysis was verified with small errors, remaining kinematic errors are attributed to manufacturing clearance, suggesting areas for improvement in future versions.
- Further Optimization: While the study performed two optimization steps, a full multi-objective optimization considering all performance measures simultaneously could yield further improvements.
- Advanced Control: Exploring more advanced control strategies beyond PD and basic impedance control could further enhance the interface's capabilities.
Summary updated