Overview
Senior design project at Cedarville University. A golf cart had been partially automated by previous student teams using a hand-coded robot framework. Our team took it the rest of the way — full ROS migration, Lidar-based SLAM, and autonomous campus navigation.
Led a 6-person team with two faculty advisors over two semesters.
What I Built
ROS migration. Replaced the existing hand-coded framework with a full ROS stack, enabling use of the standard navigation stack and community tooling.
Lidar costmap pipeline. Built costmaps from raw Lidar point clouds using ground plane estimation and surface normal analysis via the C++ Point Cloud Library (PCL).
Multi-sensor SLAM. Implemented SLAM fusing Lidar, IMU, and GPS. Applied Kalman filtering for IMU/GPS state estimation to support reliable autonomous campus navigation.
Embedded integration. Interfaced with vehicle hardware through CAN, SPI, and I2C communication protocols.
Team & Process
- 6 engineers, 2 faculty advisors
- Weekly progress reports throughout both semesters
- Formal project proposal and final presentation at completion