Skip to main content
SpringerLink
Log in

Self-driving model car acquiring three-point turn motion by using improved LSTM model

  • Original Article
  • Published:
Artificial Life and Robotics Aims and scope Submit manuscript

Abstract

This paper considers to make a small-scale car composed of low-cost hardware components acquire autonomous driving tasks by using deep learning. The tasks to be acquired are assumed to be used for indoor patrols of mobile robots, and are given as a series of tasks, where the car proceeds along the walls of a narrow corridor and stops in front of the target object. The tasks include a three-point turn as a difficult task. The difficulties in the tasks are that different actions are required even for similar input camera images and that the acceleration value cannot be kept at zero when stopped. To deal with these difficulties, we propose models outputting the state of the car in addition to steering and throttle values, models with throttle value of one step before in addition to an image, and improved long short-term memory (LSTM) models with/without additional input and output. Our experiments demonstrate that the proposed LSTM models with three output and additional input achieve a 100% success rate until just before the goal.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12

Similar content being viewed by others

References

  1. Badue C et al (2019) Self-driving cars: a survey. arXiv:1901.04407v2

  2. Bojarski M et al. (2016) End to end learning for self-driving cars. arXiv:1604.07316v1

  3. Du S, Guo H, Simpson A (2019) Self-driving car steering angle prediction based on image recognition. arXiv:1912.05440v1

  4. Tzafestas SG (2018) Mobile robot control and navigation: a global overview. J Intell Robot Syst 91:35–58

    Article  Google Scholar 

  5. Huang BQ, Cao GY, Guo M (2005) Reinforcement learning neural network to the problem of autonomous mobile robot obstacle avoidance. In: International conference on machine learning and cybernetics, pp 85–89

  6. Ganapathy V, Yun SC, Joe HK (2009) Neural Q-learning controller for mobile robot. In: 2009 IEEE/ASME international conference on advanced intelligent mechatronics, pp 863–868

  7. Jing Y, Chen Y, Jiao M, Huang J, Niu B, Zheng W (2019) Mobile robot path planning based on improved reinforcement learning optimization. In: International conference on robotics systems and vehicle technology, pp 138–143

  8. Donkey Car—An opensource DIY self driving platform for small scale cars. https://www.donkeycar.com/. Accessed 13 Dec 2020

  9. Bechtel MG et al (2018) Deeppicar: a low-cost deep neural network-based autonomous car. In: IEEE international conference on embedded and real-time computing systems and applications, pp 11–21

  10. Thomas A, Hedley J (2019) FumeBot: a deep convolutional neural network controlled robot. Robotics 8(3):62

    Article  Google Scholar 

  11. Brooks RA (1986) A robust layered control system for a mobile robot. IEEE J Robot Autom 2(1):14–23

    Article  Google Scholar 

  12. Kurt TE (2006) Hacking Roomba: ExtremeTech. Wiley Publishing, Indianapolis

    Google Scholar 

  13. Colledanchise M, Ögren P (2018) Behavior trees in robotics and AI: an introduction. CRC Press, Boston

    Book  Google Scholar 

  14. Fukuoka R et al (2020) Proposal of CNN models for acquiring stop and turn motion of self-driving model car. In: The twenty-fifth international symposium on artificial life and robotics 2020, pp 371–375

  15. Lecun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-based learning applied to document recognition. Proc IEEE 86(11):2278–2324

    Article  Google Scholar 

  16. Kingma DP, Ba J (2014) Adam: a method for stochastic optimization. arXiv:1412.6980v9

Download references

Acknowledgements

This work was supported by JSPS KAKENHI Grant Number 20K11994.

Author information

Authors and Affiliations

  1. Graduate School of Science and Engineering, Kagoshima University, Kagoshima, Japan

    Ryo Fukuoka, Noritaka Shigei, Yoshihiro Nakamura & Hiromi Miyajima

  2. School of Information and Data Sciences, Nagasaki University, Nagasaki, Japan

    Hirofumi Miyajima

Authors
  1. Ryo Fukuoka
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Noritaka Shigei
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Hirofumi Miyajima
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Yoshihiro Nakamura
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Hiromi Miyajima
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Noritaka Shigei.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This work was presented in part at the 26th International Symposium on Artificial Life and Robotics (Online, January 21–23, 2021).

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Fukuoka, R., Shigei, N., Miyajima, H. et al. Self-driving model car acquiring three-point turn motion by using improved LSTM model. Artif Life Robotics 26, 423–431 (2021). https://doi.org/10.1007/s10015-021-00697-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10015-021-00697-9

Keywords

Search

Navigation