A Comprehensive Framework for an Autonomous Vehicle System Utilizing GPS, Google Maps, and Real-Time Sensor Fusion

This paper describes the design, development, and testing of a complete autonomous vehicle system. The main aim of this project is to develop a driverless transport system that can reach any destination provided by the user using global path planning through GPS and Google Maps and local path planning through sensor fusion. The system architecture consists of a global path planner, which takes input high-level routes and produces a series of waypoints for the local control unit. The local control unit is responsible for avoiding obstacles and keeping the vehicle within its lane boundaries and is also capable of altering the global path if conditions change along that path. A simple Graphical User Interface (GUI) allows users to input the destination and display the status of the system. The vehicle has a GPS module to know where it is, an IMU to know where it is orientated, ultrasonic or LiDAR sensors to sense its environment and more. According to experimental aggregation on a scaled model, the system was able to successfully reach destination from arbitrary source with smooth avoidance of static and dynamic obstacles along with maximum path adherence. This paper is a basic proof-of-concept showing that cloud-based navigation and real-time robot control can work together. Furthermore, this work establishes a roadmap toward further autonomy and reliability.