Orchestrate performance metrics collection of an Automated Driving System software platform

Sammanfattning

In an autonomous vehicle software, the base software platform is responsible for managing hardware resources and providing fundamental services and abstractions for higher-level applications. In the current iterative development environment of the industry, it has become necessary to ensure the reliability and performance of the base software platform through robust testing conducted on target hardware and the collection of relevant metrics. This project, carried out on Scania’s physical machines, uses the Robot Framework to orchestrate relevant tests while creating a load on the system by replaying real Light Detection and Ranging (LiDAR) data. By collecting and evaluating the system’s performance metrics during testing, the performance and stability of the base software platform under various conditions are observed. To validate the feasibility and effectiveness of the orchestration system, User Datagram Protocol (UDP) buffer sizes were used as a case study for adjusting system parameter configurations. Additionally, another case study was conducted to monitor Central Processing Unit (CPU) usage under different configurations. Through the implementation and validation of this project, the feasibility of the orchestration system was confirmed. It can be used as a development tool to enhances the reliability and efficiency of Automated Driving Systems (ADS) software development.