Electrical

  • Electronics designs the various printed circuit boards (PCBs) and integrating the low voltage systems throughout the vehicle. Projects include sensor data acquisition, driver interface systems, power distribution, and wiring to integrate all the electronic subsystems. Working with eTrain in Spring quarter, there is a large emphasis on producing a reliable vehicle that can consistently run as expected. This involves heavy debugging sessions, reviewing wiring, and other testing.

  • The Firmware team spearheads the design, implementation, and integration of all programmable systems within the vehicle. This subteam’s projects center around enabling the car’s Controller Area Network (CAN Bus), a communication protocol that connects systems such as the high-voltage battery, inverters/motors, driver dash, low-voltage power distribution, and much more. The Firmware subteam also collects and stores all vehicle data across the CAN Bus using a state-of-the-art telemetry system, which is used to tweak parts that shave seconds on race days. By being on the software team, members get real full-stack experience, from developing low-level STM32/RP2040 modules in C/C++ and Python to client/member-facing applications in JavaScript.

  • As the newest division of UW Formula Motorsports, Driverless is pioneering the development of a fully autonomous system for the team’s race car. In 2026, we competed in the inaugural Driverless Cup at FSAE Michigan, placing 3rd overall and 2nd in Driverless Design. Since then, Driverless has become fully integrated with both the team and the vehicle, bringing together disciplines ranging from low-level programming and embedded systems to computer vision, robotics, computer science, vehicle dynamics, and controls. The autonomous pipeline begins by processing thousands of CAN messages every second alongside more than two million LiDAR points per second. From this raw data, we use algorithms such as Extended Kalman Filters to estimate vehicle states like velocity and yaw rate, while perception algorithms extract and locate cones from the LiDAR point cloud. We then use this information to build a map of our surroundings, determine where the car is within that map, plan a trajectory through the track, and calculate the steering, braking, and torque commands needed to follow it before sending those commands back through CAN into the car’s electrical and mechanical systems. As if this wasn't complex enough, this entire pipeline must happen in real time under 100 milliseconds for each lidar frame, while meeting strict accuracy and reliability requirements despite sensor noise, model inaccuracies, latency, and probabilistic uncertainty throughout the system. It is an incredibly exciting engineering problem and this year for T38, we want to push the limits of our car with our autonomous driver.