Loop closure detection in SLAM by combining visual CNN features and submaps

Using simultaneous localization and mapping (SLAM) with 2D LIDAR is an efficient approach for robots to build a floor plan, but it is sensitive to the environment. For improving the accuracy, we match LIDAR data with sub-maps. Furthermore, we convert LIDAR data to images and merge with camera data for image matching. Combining the two approaches, we achieve robust and accurate loop closure detection. The descriptors generated by CNN model will be used as features to image matching for accuracy improvement.

[1]  Wolfgang Hess,et al.  Real-time loop closure in 2D LIDAR SLAM , 2016, 2016 IEEE International Conference on Robotics and Automation (ICRA).

[2]  Sven Hellbach,et al.  Large scale place recognition in 2D LIDAR scans using Geometrical Landmark Relations , 2014, 2014 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[3]  Ruofei Zhong,et al.  Feature-Based Laser Scan Matching and Its Application for Indoor Mapping , 2016, Sensors.

[4]  Matthew A. Brown,et al.  Learning Local Image Descriptors , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[5]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[6]  Andrea Vedaldi,et al.  HPatches: A Benchmark and Evaluation of Handcrafted and Learned Local Descriptors , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[7]  Wolfram Burgard,et al.  Monte Carlo Localization: Efficient Position Estimation for Mobile Robots , 1999, AAAI/IAAI.

[8]  Edwin Olson,et al.  M3RSM: Many-to-many multi-resolution scan matching , 2015, 2015 IEEE International Conference on Robotics and Automation (ICRA).

[9]  Nicholas Roy,et al.  A Linear Approximation for Graph-Based Simultaneous Localization and Mapping , 2012 .

[10]  John J. Leonard,et al.  Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age , 2016, IEEE Transactions on Robotics.

[11]  Jeffrey K. Uhlmann,et al.  Unscented filtering and nonlinear estimation , 2004, Proceedings of the IEEE.

[12]  Evangelos E. Milios,et al.  Globally Consistent Range Scan Alignment for Environment Mapping , 1997, Auton. Robots.

[13]  Krystian Mikolajczyk,et al.  Learning local feature descriptors with triplets and shallow convolutional neural networks , 2016, BMVC.

[14]  Stefan Kohlbrecher,et al.  A flexible and scalable SLAM system with full 3D motion estimation , 2011, 2011 IEEE International Symposium on Safety, Security, and Rescue Robotics.

[15]  Wolfram Burgard,et al.  Improved Techniques for Grid Mapping With Rao-Blackwellized Particle Filters , 2007, IEEE Transactions on Robotics.

[16]  Sebastian Thrun,et al.  The Graph SLAM Algorithm with Applications to Large-Scale Mapping of Urban Structures , 2006, Int. J. Robotics Res..

[17]  Benson Limketkai,et al.  Comparison of indoor robot localization techniques in the absence of GPS , 2010, Defense + Commercial Sensing.

[18]  Nikos Komodakis,et al.  Learning to compare image patches via convolutional neural networks , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).