Speaker Details

Speaker Company

Kris De Meester

Kris graduated as an electromechanical engineer at the University of Leuven and started working in the automotive sector in 1993. He began as a technical consultancy engineer on high-tech projects in NVH. During his career, he has acquired experience in engineering management and technical sales, was promoted to business unit manager and managing director within different industrial companies, and acquired worldwide experience in Europe, the USA and Asia. Strongly convinced of the innovative power of the XenomatiX technology and its inevitable breakthrough, he joined XenomatiX as vice president of sales and business development in 2016.


Particular AI considerations in autonomous off-road applications

Off-road applications require particular attention and understanding of the required operation and safety detection. Although the applications run in a confined area, training an NN on occasionally occurring events is difficult and requires manual manipulation. The consequences of false positives or false negatives translate less frequently into lethal accidents but always with very high financial damages. This paper presents an off-road strategy favoring rule-based AI algorithms, avoiding extensive training and validation, and allowing flexible tuning, modifications and extensions with simple, speedy and robust AI algorithms. Examples are presented to support this strategy.