College - Author 1

College of Engineering

Department - Author 1

Electrical Engineering Department

Degree Name - Author 1

BS in Electrical Engineering

College - Author 2

College of Engineering

Department - Author 2

Electrical Engineering Department

Degree - Author 2

BS in Electrical Engineering

College - Author 3

College of Engineering

Department - Author 3

Electrical Engineering Department

Degree - Author 3

BS in Electrical Engineering

Date

6-2026

Primary Advisor

Clay McKell, College of Engineering, Electrical Engineering Department

Abstract/Summary

This report documents the design, implementation, and testing of an autonomous litter-collection rover developed as a Senior Project Design Lab (EE 460/463/464) at California Polytechnic State University. The rover integrates autonomy, computer vision, embedded real-time control, mecanum-wheel omnidirectional mobility, and a two-degree-of-freedom robotic arm to detect, approach, and collect small ground-level litter such as bottles, wrappers, and paper fragments.

The system uses a two-layer compute architecture: an NVIDIA Jetson Orin Nano running ROS 2 for perception, SLAM, and path planning, paired with an STM32L4A6ZG microcontroller for real-time motor control and odometry. The robot is built on a multi-level aluminum frame with a 3D-printed collection ramp and custom electronics housings.

The completed system demonstrated autonomous litter detection, target localization, path planning, obstacle avoidance, and robotic collection capabilities. Testing validated battery-powered teleoperation, real-time perception, and autonomous navigation using a custom Theta* planning framework and ROS 2 autonomy stack. Path-following experiments achieved mean cross-track errors as low as 0.033 m during nominal operation and 0.059 m in obstacle-avoidance scenarios. Scoop testing achieved success rates of up to 100% for plastic bottle collection and 90% for several paper-based litter configurations. The final platform weighed approximately 42 lb and integrated a YOLO-based detector trained on approximately 12,700 images. This report presents the system architecture, design methodology, implementation, testing procedures, results, and lessons learned.

Share

COinS