Safe Configurable Maps for Off-Road Sites: Proposed methods for safe and efficient map updates for autonomous trucks
2019 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE credits
Student thesis
Abstract [en]
Autonomous vehicle technology is advancing at a very high pace and self-driving trucks on control-tower operated work sites is already a reality. These autonomous trucks need a highly accurate map of the surroundings for operation and navigation, and it is of great importance to be able to update that map with the ever-changing off-road work site. The autonomous fleet examined have to stop for every update of the site map, which induces unnecessary downtime when updating the site map frequently. The purpose of this work is to contribute to the development of safe configurable maps for autonomous vehicles on off-road sites by identifying and analyzing different map updating methods, proposing the best one, and suggesting how to implement it for this project's case. The result was five different map updating methods, which were evaluated with respect to efficiency and safety. Efficiency was evaluated by comparing total fleet downtime of the proposed solutions with the existing situation. Safety was evaluated by doing a fault tree analysis (FTA) for each proposed solution and comparing the relative size of the fault trees. Proposed Solution III using map tiles was chosen as the most appropriate method to implement for this project's case because it is both efficient and relatively simple. It divides the site map with a grid into smaller rectangular maps and only needs to stop vehicles which are inside the updated tile. The rest of the fleet is able to replace that tile parallel to operation and, therefore, total fleet downtime is significantly reduced. By reaching the stated goal, this work is in line with its original purpose and has contributed to the development of safe configurable maps for autonomous vehicles on off-road sites.
Place, publisher, year, edition, pages
2019.
Keywords [en]
Maps, updates, autonomous
National Category
Vehicle Engineering
Identifiers
URN: urn:nbn:se:ltu:diva-74178OAI: oai:DiVA.org:ltu-74178DiVA, id: diva2:1320338
External cooperation
Scania CV AB
Subject / course
Student thesis, at least 30 credits
Educational program
Mechanical Engineering, master's level
Supervisors
Examiners
2019-06-052019-06-042019-06-05Bibliographically approved