Abstract
Simultaneous localization and mapping (SLAM) plays an important role in fully autonomous systems when a GNSS (global navigation satellite system) is not available. Studies in both 2D indoor and 3D outdoor SLAM are based on the appearance of environments and utilize scan-matching methods to find rigid body transformation parameters between two consecutive scans. In this study, a fast and robust scan-matching method based on feature extraction is introduced. Since the method is based on the matching of certain geometric structures, like plane segments, the outliers and noise in the point cloud are considerably eliminated. Therefore, the proposed scan-matching algorithm is more robust than conventional methods. Besides, the registration time and the number of iterations are significantly reduced, since the number of matching points is efficiently decreased. As a scan-matching framework, an improved version of the normal distribution transform (NDT) is used. The probability density functions (PDFs) of the reference scan are generated as in the traditional NDT, and the feature extraction - based on stochastic plane detection - is applied to the only input scan. By using experimental dataset belongs to an outdoor environment like a university campus, we obtained satisfactory performance results. Moreover, the feature extraction part of the algorithm is considered as a special sampling strategy for scan-matching and compared to other sampling strategies, such as random sampling and grid-based sampling, the latter of which is first used in the NDT. Thus, this study also shows the effect of the subsampling on the performance of the NDT.
Highlights
Simultaneous localization and mapping (SLAM) is of a great importance for autonomous systems in an environment where a global navigation satellite system (GNSS) is not reachable
In feature-based SLAM (Fb-SLAM), the state vector - holding the robot pose and landmark positions - and the uncertainty covariance matrix are estimated with well-known probabilistic filtering methods, such as a Kalman filter (KF), an information filter (IF) or a particle filter (PF)
Since the artificial landmarks can only be used in already-known areas, the feature extraction methods have a vital importance for Fb-SLAM in unknown environments
Summary
Simultaneous localization and mapping (SLAM) is of a great importance for autonomous systems in an environment where a global navigation satellite system (GNSS) is not reachable. For the dense data mapping, scan-matching or image registration is the essential part of SLAM methods For this purpose, most of researchers prefer to use the iterative closest point (ICP) scan-matching method [3]. The reason is that the objects with certain geometric shapes - like lines or planes - are matched and irregular points - like bushes and tree leaves - are filtered out in the registration process Another advantage is that the method does not need high precision odometry data since scan-matching is based on the features. We propose a fast and robust scan registration method using feature points extracted from the input scan for 3D outdoor SLAM.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
