Simultaneous localization and mapping via Multi-vehicle Map Merging

Abstract

Nowadays the research about the area of the autonomous ground vehicle is more popular than before Simultaneous Localization And Mapping always the first challenge we meet However constructing a trusty map cause a lot of time and labor consumptions Hence it is necessary to figure out a method to speed up the mapping process In fact one of the significant methods is that uses multiple vehicles to explore different parts of the map Once a vehicle can explore and merge maps autonomously we can save lots of time This thesis proposes a complete structure for autonomous driving including SLAM exploration map merging planning and control The vehicle can construct maps through the Rao-Blackwellized Particle filter and explores the unknown region according to the frontier points After that plan a trajectory from Timed-Elastic-Band planner and use the Model Predict Control to track the trajectory Finally use feature extraction to find out the relationship between each map and merge the maps together After integrates every module in ROS we can have the experiment in the laboratory Eventually the vehicles can really explore and merge the maps autonomously

Date of Award	2019
Original language	English
Supervisor	Jiun-Haur Tarn (Supervisor)