Advertisement

Binary Classification of Terrains Using Energy Consumption of Hexapod Robots

  • Valeriia IegorovaEmail author
  • Sebastián Basterrech
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 554)

Abstract

The terrain classification problem is a relevant task in autonomous robots, which can help in the control locomotion and motion planning of autonomous robots. We conduct several experiments in different environments, where a hexapod walking robot covers some specific terrains. In this paper, we present an experimental analysis of the binary terrain classification problem using the most important variable (current signal) related to the energy consumption of the robot. The current signal is a sequential data that evolves in time, therefore our problem is limited to develop a machine learning method for classifying this signal according to the terrain. We analyze the problem using the Long Short-Term Memory (LSTM) model, which is a Recurrent Neural Networks that has obtained good performance for time-series classification. We evaluated several binary scenarios, where each scenario presents two different types of terrains. Our results show that the LSTM model trained only with information related to the current signal is able to distinguish binary situations of terrain.

Keywords

Control locomotion Terrain classification Time-series classification Neural networks Long short-term memory 

Notes

Acknowledgment

This work has been supported by the Czech Science Foundation (GAČR) under research project No. 18-18858S, and the authors acknowledge the support of the OP VVV MEYS funded project CZ.02.1.01/0.0/0.0/16_019/0000765 “Research Center for Informatics”.

References

  1. 1.
    DuPont, E.M., Moore, C.A., Collins, E.G., Coyle, E.: Frequency response method for terrain classification in autonomous ground vehicles. Auton. Robots 24(4), 337–347 (2008).  https://doi.org/10.1007/s10514-007-9077-0CrossRefGoogle Scholar
  2. 2.
    Lauro, O., Johann, B., Gary, W., Robert, K.: Terrain characterization and classification with a mobile robot. J, Field Robot. 23(2), 103–122 (2006). https://onlinelibrary.wiley.com/doi/abs/10.1002/rob.20113CrossRefGoogle Scholar
  3. 3.
    Otte, S., Laible, S., Hanten, R., Zell, A.: Robust visual terrain classification with recurrent neural networks. In: European Symposium in Artificial Neural Networks (ESANN), pp. 451–457, Jan 2015Google Scholar
  4. 4.
    Valada, A., Spinello, L., Burgard, W.: Deep feature learning for acoustics-based terrain classification. In: International Symposium on Robotics Research (ISRR) (2015)Google Scholar
  5. 5.
    Falck, R.H., Čižek, P., Basterrech, S.: Recurrence plot and convolutional neural networks for terrain classification using energy consumption of multi-legged robots. In: International Conference on Soft Computing MENDEL 2018. Czech Republic, Brno (2018)Google Scholar
  6. 6.
    Gers, F.A., Schmidhuber, J.A., Cummins, F.A.: Learning to forget: continual prediction with LSTM. Neural Comput. 12(10), 2451–2471 (2000).  https://doi.org/10.1162/089976600300015015CrossRefGoogle Scholar
  7. 7.
    Greff, K., Srivastava, R.K., Koutník, J., Steunebrink, B.R., Schmidhuber, J.: LSTM: A search space odyssey (2015). CoRR, vol. abs/1503.04069. http://arxiv.org/abs/1503.04069
  8. 8.
    Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning. MIT Press (2016). http://www.deeplearningbook.org
  9. 9.
    Bengio, Y., Frasconi, P., Simard, P.: The problem of learning long-term dependencies in recurrent networks. In: IEEE International Conference on Neural Networks, vol. 3(2), pp. 1183–1188 (1993)Google Scholar
  10. 10.
    Hochreiter, S., Schmidhuber, J.: Long short-term memory. Neural Comput. 9(8), 1735–1780 (1997)CrossRefGoogle Scholar
  11. 11.
    Čížek, P., Faigl, J.: On localization and mapping with RGB-D sensor and hexapod walking robot in rough terrains. In: IEEE International Conference on Conference Systems, Man, and Cybernetics (SMC), pp. 2273–2278 ( 2016)Google Scholar
  12. 12.
    Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.: Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res. 15, 1929–1958 (2014). http://jmlr.org/papers/v15/srivastava14a.htmlMathSciNetzbMATHGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Department of Computer Science, Faculty of Electrical Engineering and Computer ScienceCzech Technical UniversityPragueCzech Republic

Personalised recommendations