Achieving sufficient safety measures is among the major challenges in developing automated vehicles that can operate safely in an urban environment. Data fusion between an in-vehicle camera and a LiDAR sensor can be used for detection and tracking of other road users in an automated vehicle. In addition, simulated environments together with high-level deterministic, supervised and reinforcement learning-based autonomous control could provide traffic safety benefits in the future. These AI-based technologies have been studied in the AI4DI project to enable the Mobility as a Service (MaaS) operators fleet management of automated vehicles. The development and testing of these methods are presented in this chapter with the first promising results. The Camera - LiDAR fusion algorithm provided very good results with the accuracy evaluation using the KITTI dataset.The real-time applicability of the fusion algorithm was also successfully verified.
|Title of host publication||Artificial Intelligence for Digitising Industry|
|Publication status||Published - 27 Sep 2021|
|MoE publication type||B2 Part of a book or another research book|
- automated driving
- sensor data fusion
- 3D object detection and tracking
- reinforcement learning