Abstract
An cloud computing platform based on B/S architecture and docker container technology for autonomous driving simulation has been established in this paper. The map editor module of the cloud platform lets users design 3D scenes for simulating and testing automated driving systems. When the customized roadway scene for simulation created, it would be saved as OpenDrive format both for the server of cloud platform and CarMaker’s TestRun which all parameters of the virtual environment (vehicle, road, tires, etc.) are sufficiently defined. Then, based on the application online (APO) communication protocol of CarMaker, the local APO agent service was created. When the 27 parameters of vehicle dynamics received from CarMaker server, they were sent to the cloud platform in real time through UPD protocol. The process of data communication is completed by APO agent. Through the work above, a co-simulation between cloud platform and CarMaker could be successfully established for autonomous driving with seventeen-degree-of-freedom. Through the co-simulation experiment, it is found that the real-time data sampling frequency of the co-simulation is 70Hz, which completes the synchronous simulation of carmaker and cloud platform.
Vision based hardware-software real-time control system for autonomous landing of an UAV