Abstract
This paper presents the architecture and implementation of a tele-presence wheelchair system based on tele-presence robot, intelligent wheelchair, and touch screen technologies. The tele-presence wheelchair system consists of a commercial electric wheelchair, an add-on tele-presence interaction module, and a touchable live video image based user interface (called TIUI). The tele-presence interaction module is used to provide video-chatting for an elderly or disabled person with the family members or caregivers, and also captures the live video of an environment for tele-operation and semi-autonomous navigation. The user interface developed in our lab allows an operator to access the system anywhere and directly touch the live video image of the wheelchair to push it as if he/she did it in the presence. This paper also discusses the evaluation of the user experience.
Introduction
Elderly and disabled individuals often face the dual challenges of limited physical mobility and social isolation, which can lead to geriatric depression. While Intelligent Wheelchairs (IWs) address mobility, they are often expensive and lack social connectivity features. Conversely, tele-presence robots facilitate social interaction but do not assist the local user with their own mobility. This research introduces a low-cost tele-presence wheelchair system designed to bridge this gap by providing both mobility and social presence through an intuitive, collaborative interface.
System Architecture and Methods
The system is built upon a standard commercial electric wheelchair enhanced with a custom tele-presence interaction module. This modular approach ensures the system remains affordable and adaptable.
Imaging and the TIUI Interface
To provide the remote operator with sufficient situational awareness, the system utilizes two cameras: a Forward-Facing (FF) camera for social interaction and a Down-Facing (DF) camera with a wide-angle lens for navigation. Using Speeded Up Robust Features (SURF), these two feeds are stitched into a single "Forward-Down-Facing" (FDF) live stream. This prevents the cognitive load associated with switching between multiple video windows.
The Touchable Live Video Image based User Interface (TIUI) allows the operator to control the wheelchair via a tablet. By touching the lower portion of the video (the ground), the operator can "push" the wheelchair forward. This "Virtual Track-Point" creates a natural mapping of touch-to-motion, simulating the physical act of pushing a wheelchair from behind.
Semi-Autonomous Navigation
Navigating narrow spaces like doorways is challenging for remote operators due to network latency and perspective distortion. The system employs a "human-in-the-loop" collaborative control strategy. Using edge and corner detection, the system identifies doorways and generates safe trajectories using Bézier Curves. When an operator indicates intent to move through a door, the system smooths the path to avoid collisions, reducing the operator's burden.
Evaluation and Results
The system was evaluated through a user experience study with 20 volunteers and a navigation benchmark involving various doorway configurations.
- Efficiency: Remote operation was highly efficient; operators completed a test course in an average of 34.8 seconds, nearly matching the 33.1 seconds required for in-person assistance.
- Usability: 90% of participants found the TIUI interface intuitive and easy to learn.
- Navigation Success: The semi-autonomous system achieved an 86.6% success rate in simple scenes. While performance decreased in cluttered environments with multiple doors, the assistance was significantly faster than manual control during high-offset maneuvers.
Discussion and Future Work
While the system demonstrates a successful integration of tele-presence and mobility, limitations remain. Visual clutter in complex environments can challenge the detection algorithms, and the current reliance on high-quality video stitching necessitates robust bandwidth. Future work will focus on testing with the target demographic in real-world home environments and refining navigation algorithms to handle increased environmental complexity.
Summary updated