Real-Time Data Processing in Autonomous Vehicles Based on Distributed Architecture: A Case Study

This work aims to evaluate the real-time processing system in the context of an autonomous vehicle with limited hardware and software capabilities. We elaborate algorithm implemented in 1/10 scale electric car using a line scan camera, speed sensors, and embedded electronic control system. The vehicle navigates in an arbitrary one-lane circuit using an edge detection algorithm. The challenge was to make a complete one loop of the arbitrary circuit in the shortest time with various lighting conditions. The experiments show that several issues were revealed in each step of data sensors processing and need a robust algorithm to handle exceptions caused by multiple disturbances. Furthermore, we paid particular attention to time constraints in embedded processor calculation and actuators response time to achieve reliable critical software control algorithms.