Many people have difficulty walking, and the percentage of people with this challenge increases with age. The "mobility challenge,"which we address in this project, centers on one's ability to independently move through the world. Enabling individuals to maintain their independence of mobility has significant social importance for society as a whole. While research in sensing and autonomous technology has made great strides in recent years, affordable fully autonomous systems are still a distant goal, primarily because of a lack of sensing accuracy and robustness based on off-the-shelf low-cost sensors. Self-driving vehicles being tested by companies rely heavily on expensive 3D LiDAR to locate themselves on the detailed maps they need to get around, and to identify things like pedestrians and other vehicles. In this project, we investigate an efficient sensing and perception hardware and software system design for autonomous and intelligent wheelchairs. The goal is to develop an experimental testbed with multi-modal sensors, computing systems, control and mobility systems. This affordable testbed will be a full-fledged modular platform to test and deploy latest deep learning-based algorithm without expensive hardware components.