A new intelligent walker with the integration of a six-axis force sensor, LiDar sensors, depth camera, and motorized wheels was developed to bridge the gap between passive walkers and other robotic walkers. Using previous designs as a baseline for improvement, an updated walking frame has been developed with an emphasis on human-centered design, finite element analysis of nominal loading conditions, weight reduction, adjustability, usability, safety, and manufacturability. 

The design further differentiates from version I of the smart walker by incorporating only two rear motorized with improved wheel choice and two front passive caster wheels to provide better traction, reduce overall weight, usability, and safety during operation, and the frame is fully adjustable to improve ergonomics and user adaptability. The design will eventually utilize a suspension mechanism to minimize vibration and help in navigating rougher terrains.

Other additional features included a suspension system and horizontally expandable legs, significantly improving mobility and stability across various terrains whereas traditional walkers often fall short on uneven surfaces. The integration of horizontally expandable legs allows for footprint adjustment and more postural stability. The walker utilizes a control system with LiDar, camera, and force sensors with machine learning (ML) algorithms for environment mapping, obstacle detection, and user intention estimation to enhance the safety and usability of the smart walker.

Team members: Tiffany Chau, Quynh Tran, Harshal Makwana, Avadh Vanparia

This intelligent motorized walker features a 4-wheel structurally stable design, supported by the fundamental implementation of an optimized Walker Tipping Index (WTI). Its control system includes motion planning based on the user’s intention, collision detection, and fall prevention, all to enhance the walking experience for individuals with motor impairments. The design of the walker chassis prioritized postural stability in all directions, utilizing a WTI calculator to determine overall dimensions while ensuring the index remained below 0.5 in all directions (front-back, left-right).  

The components were manufactured in a metal tubular frame with handles 3D printed with thermoplastic polyurethane (TPU). The wheel used an omnidirectional design attached to four in-hub individual high torque DC motors using velocity control with LiDar sensor object detection that either decreases speed or stops the motors when an object is near or present.

Team members: Tiffany Chau, Jerry Lopez, Anubhav Sharma