Exploring Machine Learning Techniques for Identification of Cues for Robot Navigation with a LIDAR Scanner

Aj Bieszczad

Abstract

In this paper, we report on our explorations of machine learning techniques based on backpropagation neural networks and support vector machines in building a cue identifier for mobile robot navigation using a LIDAR scanner. We use synthetic 2D laser data to identify a technique that is most promising for actual implementation in a robot, and then validate the model using realistic data. While we explore data preprocessing applicable to machine learning, we do not apply any specific extraction of features from the raw data; instead, our feature vectors are the raw data. Each LIDAR scan represents a sequence of values for measurements taken from progressive scans (with angles vary from 0° to 180°); i.e., a curve plotting distances as a functions of angles. Such curves are different for each cue, and so can be the basis for identification. We apply varied grades of noise to the ideal scanner measurement to test the capability of the generated models to accommodate for both laser inaccuracy and robot motion. Our results indicate that good models can be built with both back-propagation neural network applying Broyden–Fletcher–Goldfarb–Shannon (BFGS) optimization, and with Support Vector Machines (SVM) assuming that data shaping took place with a [-0.5, 0.5] normalization followed by a principal component analysis (PCA). Furthermore, we show that SVM can create models much faster and more resilient to noise, so that is what we will be using in our further research and can recommend for similar applications.

References

  1. Hilde, L., 2009. Control Software for the Autonomous Interoffice Delivery Robot. Master Thesis, Channel Islands, California State University, Camarillo, CA.
  2. Rodrigues, D., 2009. Autonomous Interoffice Delivery Robot (AIDeR) Software Development of the Control Task. Master Thesis, Channel Island, California State University, Camarillo, CA.
  3. Thrun, S., 1998. Learning metric-topological maps for indoor mobile robot navigation. In Artificial Intelligence, Vol. 99, No. 1, pp. 21-71.
  4. Hinkel, R, and Knieriemen, T., 1988. Environment Perception with a Laser Radar in a Fast Moving Robot. In Robot Control 1988 (SYROCO'88): Selected Papers from the 2nd IFAC Symposium, Karlsruhe, FRG.
  5. Henderson, A. M., 1012. Autonomous Interoffice Delivery Robot (AIDeR) Environmental Cue Detection. Master Thesis, Channel Island, California State University, Camarillo, CA.
  6. Leonard, J., et al., 2008. A Perception-Driven Autonomous Urban Vehicle. In Journal of Field Robotics, 1-48.
  7. Tan, F., Yang, J., Huang, J., Jia, T., Chen, W. and Wang, J., 2010. A Navigation System for Family Indoor Monitor Mobile Robot. In The 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems, October 18-22, 2010, Taipei, Taiwan.
  8. Roman, M., Miller, D., and White, Z., 2007. Roving Faster Farther Cheaper. In 6th International Conference on Field and Service Robotics - FSR 2007, Jul 2007, Chamonix, France. Springer, 42, Springer Tracts in Advanced Robotics; Field and Service Robotics.
  9. Vilasis-Cardona, X., Luengo, S., Solsona, J., Maraschini, A., Apicella, G. and Balsi, M., 2002. Guiding a mobile robot with cellular neural networks. In International Journal of Circuit Theory and Applications; 30:611- 624.
  10. Shu, L., Xu, H., and Huang, M., 2013. High-speed and accurate laser scan matching using classified features. In IEEE International Symposium on Robotic and Sensors Environments (ROSE), Page(s): 61 - 66.
  11. Nunez, P. Vazquez-Martin, R. del Toro, J. C. Bandera, A. and Sandoval, F., 2006. Feature extraction from laser scan data based on curvature estimation for mobile robotics. In IEEE International Conference Robotics and Automation (ICRA),, pp. 1167-1172.
  12. Zhang, L. and Ghosh, B. K., 2000. Line segment based map building and localization using 2d laser rangefinder. In IEEE International Conference on Robotics and Automation (ICRA), pp. 2538-2543.
  13. Harbl M., Abielmona, R., Naji, K.,and Petriu, E., 2010. Neural Networks for Environmental Recognition and Navigation of a Mobile Robot. In Proceedings of IEEE International Instrumentation and Measurement Technology Conference, Victoria, Vancouver Island, Canada.
  14. Kubota, S., Ando, Y., and Mizukawa, M., 2007. Navigation of the Autonomous Mobile Robot Using Laser Range Finder Based on the Non Quantity Map. In International Conference on Control, Automation and Systems, COEX, Seoul, Korea.
  15. Bieszczad, A. and Pagurek, B. (1998), Neurosolver: Neuromorphic General Problem Solver. In Information Sciences: An International Journal 105 (1998), pp. 239-277, Elsevier North-Holland, New York, NY.
  16. Bieszczad, A., 2015. Identifying Landmark Cues with LIDAR Laser Scanner Data Taken from Multiple Viewpoints. In Proceedings of International Conference on Informatics in Control, Automation and Robotics (ICINCO), Scitepress Digital Library.
Download


Paper Citation


in Harvard Style

Bieszczad A. (2015). Exploring Machine Learning Techniques for Identification of Cues for Robot Navigation with a LIDAR Scanner . In Proceedings of the 12th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ANNIIP, (ICINCO 2015) ISBN 978-989-758-122-9, pages 645-652. DOI: 10.5220/0005569006450652


in Bibtex Style

@conference{anniip15,
author={Aj Bieszczad},
title={Exploring Machine Learning Techniques for Identification of Cues for Robot Navigation with a LIDAR Scanner},
booktitle={Proceedings of the 12th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ANNIIP, (ICINCO 2015)},
year={2015},
pages={645-652},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005569006450652},
isbn={978-989-758-122-9},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 12th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ANNIIP, (ICINCO 2015)
TI - Exploring Machine Learning Techniques for Identification of Cues for Robot Navigation with a LIDAR Scanner
SN - 978-989-758-122-9
AU - Bieszczad A.
PY - 2015
SP - 645
EP - 652
DO - 10.5220/0005569006450652


in Harvard Style

Bieszczad A. (2015). Exploring Machine Learning Techniques for Identification of Cues for Robot Navigation with a LIDAR Scanner . In Proceedings of the 12th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ANNIIP, (ICINCO 2015) ISBN 978-989-758-122-9, pages 645-652. DOI: 10.5220/0005569006450652


in Bibtex Style

@conference{anniip15,
author={Aj Bieszczad},
title={Exploring Machine Learning Techniques for Identification of Cues for Robot Navigation with a LIDAR Scanner},
booktitle={Proceedings of the 12th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ANNIIP, (ICINCO 2015)},
year={2015},
pages={645-652},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005569006450652},
isbn={978-989-758-122-9},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 12th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ANNIIP, (ICINCO 2015)
TI - Exploring Machine Learning Techniques for Identification of Cues for Robot Navigation with a LIDAR Scanner
SN - 978-989-758-122-9
AU - Bieszczad A.
PY - 2015
SP - 645
EP - 652
DO - 10.5220/0005569006450652