Overview
The main scientific goal of our project is to allow autonomous vehicles to understand not only how the environment ‘looks like’ but also ‘what is going on’ and use the ‘what is going on’ knowledge to predict the future actions of the other dynamic objects in vehicle’s vicinity. To do so reliably in real world scenarios, autonomous vehicles need to be aware of the other traffic actors in their area of operation.
To achieve the goal, we will research and develop methods that allow to detect and track dynamic objects in the IAV’s proximity in adverse weather conditions and dense traffic situations. These methods will be based on principles of life-long learning and adaptation, which will allow them to improve their efficiency over time, leading to vehicles capable of long-term autonomous navigation in realistic situations. The main practical objective of this project is to develop research cooperation between Czechia and France in autonomous vehicles, in particular between CTU and UTBM.
The principal investigators on both sides have a good basis for collaboration. When they worked together at the University of Lincoln in the UK, Dr. Tomas Krajnik mainly studied the life-long spatio-temporal mapping [8] and persistent self-localisation of mobile robots [18], while Dr. Zhi Yan researched and developed methods for dynamic object detection and tracking [5]. Dr. Zhi Yan is now joining Prof. Yassine Ruichek’s research team at UTBM working on environment perception and autonomous navigation. The research team has several intelligent and autonomous vehicle platforms equipped with various sensors such as RGB/RGB-D cameras, 2D/3D LiDARs, GPS, etc. Prof. Yassine Ruichek is working on environment perception and localization based on multi-source data fusion [19, 20, 21, 22]. These methods are in focus of the project’s principal investigators and constitute the main competencies of autonomous vehicles.