ROS middle-layer integration to Unity 3D as an interface option for propulsion drive simulations of autonomous vehicles

Abstract

Abstract As autonomous vehicle development continues at growing speeds, so does the need for optimization, diagnosis, and testing of various autonomous systems elements, under different conditions. However, since such processes should be carried out in parallel, it may result in bottlenecks in development and increased complexity. The trend for Digital Twins brings a promising option for the diagnosis and testing to be carried out separately from the physical devices, incl. Autonomous vehicles, in the virtual world. The idea of intercommunication between virtual and physical twins provides possibilities to estimate risks, drawbacks, physical damages to the vehicle’s drive systems, and the physical one’s critical conditions. Although the problem of providing communications between these systems arises, at the speed that will be adequate to represent the physical vehicle in the virtual world correctly, it is still a trending topic. The paper aims to demonstrate a way to solve this problem-by using ROS as a middleware interface between two twining systems on the autonomous vehicle propulsion drive example. Data gathered from the physical and virtual world can be exchanged in the middle to allow continuous training and optimization of the propulsion drive model, leading to more efficient path planning and energy-efficient drive of the autonomous vehicle itself.