Undergraduate Research 2012:
During the 2012-2013 academic year I worked with Dr. Steven Frank Barrett, of the University of Wyoming, on his autonomous wheelchair project. The aim of the project is create an fully functional autonomous wheelchair with a price point substantially lower than existing solutions. To achieve this the system does not use laptops or LiDAR sensors, prominent features of other designs. Instead the system utilizes and array of low cost microcontrollers for processing, and ultrasonic range detectors for collision avoidance and odometric drift cancellation. Input to the system will be achieved through an full color 640x480 LCD display and resistive touch screen.
To the Right is an overall diagram of the design of the electrical system. Various modules are in various states of development. It was my job to wrap up the disparate portions and integrate the entire system. I have also included my proposal as well as the final report of the last group to work on the project.
My work on the system revealed that the original design work for the system had not adequately anticipated the computational complexities of localization and navigation algorithms. The selected microcontroller architecture had too limited a machine precision to maintain a workable world map, and inadequate processing capabilities to handle both graphical display and sensor processing.
I began working on a second architecture for the system. This architecture would instead use a Beaglebone embedded Linux computer for processing and a Microsoft Kinect for sensing. I performed the initial design of the new system before transitioning to other work.
To the Right is an overall diagram of the design of the electrical system. Various modules are in various states of development. It was my job to wrap up the disparate portions and integrate the entire system. I have also included my proposal as well as the final report of the last group to work on the project.
My work on the system revealed that the original design work for the system had not adequately anticipated the computational complexities of localization and navigation algorithms. The selected microcontroller architecture had too limited a machine precision to maintain a workable world map, and inadequate processing capabilities to handle both graphical display and sensor processing.
I began working on a second architecture for the system. This architecture would instead use a Beaglebone embedded Linux computer for processing and a Microsoft Kinect for sensing. I performed the initial design of the new system before transitioning to other work.
Overall Electrical System
|
| ||||
