Generating and Processing of Appearance Maps for 3D Outdoor SLAM
Prof.Dr. Hakan Temeltaş, Dept. Of Control Engineering, Istanbul Technical University, Maslak
Summary: Simultaneous localization and mapping (SLAM) technique in 3D and large scale environment is one the challenging problem in autonomous navigation of unmanned vehicles or mobile robots. In this talk sensing the 3D environment and generating and processing appearance maps will be discussed as a part of large scale SLAM. On building appearance maps of 3D environments LIDAR provides precise 3D data in form of point clouds and a holistic 3D view of the environment. Camera on the other hand provides rich visual data which is very beneficial for extracting distinctive features to be used in data association and loop closure. In this talk, it is also introduced a fast and robust scan matching method that combines the Multi-Layered Normal Distributions Transform (ML-NDT) and a feature extraction algorithm into a single framework. This is achieved by first applying the conventional NDT generation process to the reference scan, and the plane segments are extracted with the help of Random Sample Consensus (RANSAC) algorithm for the input scan. As a result, while the method can be used for robust and fast scan matching, it can also be used for feature extraction method for SLAM problem with a little extra computation. The method is applied to real experimental data and the results are quite affirmative.
Hakan Temeltaş received BSc and MSc degrees from Istanbul Technical University in 1984 and 1987 in Electrical Engineering and Control and Computer Engineering respectively. He also received his PhD degree from The University of Nottingham, UK, at Department of Electrical and Electronics Engineering in 1993. Currently he works as a professor at Department of Control Engineering in Istanbul Technical University. His current research areas of interest include path planning, control and navigation of autonomous robots and robotics locomotion systems.