Development and evaluation of demonstration information recording approach for wheelchair mounted robotic arm

Mingshan Chi,Yaxin Liu,Yufeng Yao,Yan Liu,Shouqiang Li,Chao Zeng,Ming Zhong

DOI: https://doi.org/10.1007/s40747-021-00350-9

2021-04-27

Abstract:Abstract To offer simple and convenient assistance for the elderly and disabled, researchers focus on programming by demonstration approach to improve the intelligence and adaptability of wheelchair mounted robotic arm assistive robot. But how to easily and quickly obtain the demonstration information is still an urgent problem to be solved. Based on the systematic analysis of the daily living tasks in need of robot assistance, this paper proposes the key-point-based programming by demonstration recording approach to quickly obtain the demonstration information and develops a specified demonstration interface to simplify the operation process. A corresponding evaluation approach is also proposed from the demonstration trajectories and demonstration process two aspects. Additionally, tasks of “holding water glass task”, “eating task”, and “opening door task” are carried out and experimental results, as well as comparative evaluations confirm the validity of the proposed approach with high efficiency. This study can not only offer a convenient and feasible way to obtain the demonstration information of daily living tasks, but also lay a good foundation for the assistive robot to learn relative motion skills, especially for the demonstrated dexterous manipulation skills, and semi-autonomously accomplish complex, multi-step tasks following the user’s instructions in the daily home environment.

computer science, artificial intelligence

What problem does this paper attempt to address?

The paper primarily addresses the challenges faced by wheelchair-mounted robotic arms (WMRA) in assisting the elderly and disabled in home environments, particularly the issue of how to quickly and easily obtain demonstration information.

The core contribution of the paper is the proposal of a keypoint-based Programming by Demonstration (PbD) recording method. This method effectively captures high-quality demonstration trajectories and significantly reduces the operational difficulty and psychological burden on the demonstrator. Specifically, compared to traditional direct recording methods, the keypoint-based approach not only improves the quality of the obtained demonstration trajectories but also ensures that the robot can effectively complete related household tasks, avoiding task replication failures due to excessive misoperations.

To validate the effectiveness of the proposed keypoint-based PbD recording method, the authors conducted the following work:

1. **Analyzed tasks requiring assistive robots in daily life**: Identified and categorized daily tasks including object grasping, eating/drinking, door/drawer opening, button pressing, and personal care.

2. **Constructed the WMRA assistive robot**: Selected the JACO robotic arm from Kinova as the core component and provided a detailed introduction to its performance.

3. **Designed the demonstration interface**: Developed a dedicated demonstration interface to simplify the entire operation process, allowing users to obtain specific task demonstration information through simple button clicks and input of relevant information.

4. **Proposed an evaluation method for the demonstration information recording method**: Compared the keypoint-based PbD recording method with traditional direct recording methods from the aspects of demonstration trajectories and demonstration processes, and experimentally verified the effectiveness and efficiency of the new method.

In summary, this study aims to provide users with a convenient and feasible way to quickly obtain demonstration information for daily life tasks, laying a solid foundation for assistive robots to learn related motor skills, especially for dexterous operation skills demonstrated and the ability to semi-autonomously complete complex, multi-step tasks under user guidance.