Data Storage System Requirement for Autonomous Vehicle

Autonomous Vehicle Systems (AVS) requires large set of data to not only deliver its functionality but also guarantee safety and reliability of AVS. The main functions of AVS are perception, planning and control, and big data-based machine learning algorithm is often used to perceive surrounding automatically. Like eyes and ears of human to perceive our environment, sensors such as camera, lidar and radar play a critical role for AVS. Therefore, it is also essential to process its data real-time within cache memory in computing system. While developing AVS it is required to save as much data as possible to improve AVS and to prove its reliability for its homologation as defined in UN R157. After it is in production data storage system for automated driving (DSSAD) and event data recorder (EDR) are mandated to identify responsibility of accident and to recover accident respectively. In this paper, data storage requirement of AVS is investigated by reviewing regulations and standards of AVS and its capacity requirement is estimated based on sensor data size and its requirement defined in UN R157 and UN R160 to deliver safe and reliable AVS.