Budget Tesla: An Autonomous Solution for Campus Transportation
Format
SOECS Senior Project
Faculty Mentor Name
Elizabeth Basha
Faculty Mentor Department
Electrical and Computer Engineering
Additional Faculty Mentor Name
Rahim Khoie
Additional Faculty Mentor Department
Electrical and Computer Engineering
Abstract/Artist Statement
Pacific's campus is just the right size to walk from one end to the other in about 15 minutes. However, for students with injuries or disabilities, this 15-minute trip is magnified, and it can make a trip from one class to another impossible if the time in-between is too small. The goal of this project was to create an autonomous system that would enable the transfer of people around campus, specifically targeted towards those for whom travel is more difficult and time-consuming. To fill this niche, we refurbished an abandoned electric golf cart and rigged it with the processing and perceptive hardware necessary for automation.
Designing and implementing the system took place in distinct blocks, assigned among the team members. To analyze camera data and detect obstacles, a convolutional neural network is paired with a short-range FMCW radar sensor. Based on the sensor input, a mini computer processes the data and computes what the motor control should do in order to get the golf cart from point A to B in a safe manner. The path between those points is calculated and tracked via an onboard GPS unit. These electronics are powered by a marine battery coupled with buck converters. A microcontroller-operated winch is used to activate the brake pedal and the throttle is controlled with a digital potentiometer. The cart's power is supplied by a refurbished array of batteries, sufficient for several hours of operation.
Location
School of Engineering & Computer Science
Start Date
7-5-2022 2:30 PM
End Date
7-5-2022 4:00 PM
Budget Tesla: An Autonomous Solution for Campus Transportation
School of Engineering & Computer Science
Pacific's campus is just the right size to walk from one end to the other in about 15 minutes. However, for students with injuries or disabilities, this 15-minute trip is magnified, and it can make a trip from one class to another impossible if the time in-between is too small. The goal of this project was to create an autonomous system that would enable the transfer of people around campus, specifically targeted towards those for whom travel is more difficult and time-consuming. To fill this niche, we refurbished an abandoned electric golf cart and rigged it with the processing and perceptive hardware necessary for automation.
Designing and implementing the system took place in distinct blocks, assigned among the team members. To analyze camera data and detect obstacles, a convolutional neural network is paired with a short-range FMCW radar sensor. Based on the sensor input, a mini computer processes the data and computes what the motor control should do in order to get the golf cart from point A to B in a safe manner. The path between those points is calculated and tracked via an onboard GPS unit. These electronics are powered by a marine battery coupled with buck converters. A microcontroller-operated winch is used to activate the brake pedal and the throttle is controlled with a digital potentiometer. The cart's power is supplied by a refurbished array of batteries, sufficient for several hours of operation.