Autonomous AI Robotic Shopping Cart System

Abstract:

This research focuses on developing a fully autonomous following cart robot that transports items by following individuals, especially the physically challenged; this uses a systematic research methodology and a decentralized top-down approach, structured into four modules: Sensor Testing, Image Recognition, Obstacle Detection and Avoidance, and Object Tracking. The methodology involves a step-by-step progression, commencing with fundamental sensor testing and advancing to more intricate tasks, including object detection and tracking. The approach is characterized by a well-organized and logical sequence, breaking down the project into manageable components, testing each module, and integrating them into a cohesive system. This robot is designed using a Goal-Based Learning agent with a combination of colour-based image segmentation and path-planning algorithms to achieve accurate and efficient navigation in various environments, successfully identifying and following its target. This work contributes to the field of autonomous following robots demonstrating a modular design for improved development and maintainability. The project achieved remarkable progress across the four core modules. The system showcases accurate and efficient navigation capabilities through various environments, identifying and following its target, utilizing the four-module combination.

Keywords: Arduino, Autonomous Robot, Picxy2 CAM, Image Recognition, Following Robot, IR sensors, Ultrasonic sensors, DC motors, Obstacle detection and avoidance, and Path tracking.