Synchronizing Multiple Data Streams by Time For Vehicle Control

New research on autonomous driving is beginning to incorporate remote, relative-to-vehicle, environmental data. Few protocols exist to deliver remote data to autonomous vehicles - for example observation meta-data or sensor streams such as camera, LIDAR, etc. Researchers bypass this problem by using a single Robot OS controller or rudimentary streaming pipes. However, with recent advancements in vehicle to everything (V2X) protocols, including better LTE and millimeter wave (MMwave) communications in mobile environments, a dedicated multiple-source streaming protocol is increasingly necessary. For this reason, we developed a protocol that allows multiple and distinct transmitters of data to stream sensor information to a single receiver with features including session management, encryption, and jitter regulation. The Synchronized Buffer proposed herein seeks to synchronize data frames by their time stamps, in the spirit of the Real-Time Transport Protocol (RTP) used in VoIP. We validated the proposed time synchronization algorithm by streaming data at different simulated data frames per second. Quality of service (QoS) and potential security flaws are also explored. Such a dedicated streaming protocol addresses current problems by allowing vehicles to integrate previously unavailable remote data sources into critical autonomous decision making.