End-to-End Latency Optimization for Containerized ROS 2 Autonomous Driving Software

One of the most significant shifts in software development is the growing prevalence of microservice-oriented and containerized applications. However, it faces challenges regarding timing and latency optimization. We present an optimization approach that has been specifically tailored to the use of containerized ROS 2 architectures. The optimization approach was implemented and evaluated using three different configurations, differing by the weighting of individual callback sequences of the system during optimization. Our experimental evaluation demonstrated the optimization of an autonomous racing software, developed for high-speed racing applications. The optimization yields a notable improvement in the maximum end-to-end latency by 22.47 % compared to the baseline measurement.