A journal of IEEE and CAA , publishes high-quality papers in English on original theoretical/experimental research and development in all areas of automation
Volume 7 Issue 1
Jan.  2020

IEEE/CAA Journal of Automatica Sinica

  • JCR Impact Factor: 11.8, Top 4% (SCI Q1)
    CiteScore: 17.6, Top 3% (Q1)
    Google Scholar h5-index: 77, TOP 5
Turn off MathJax
Article Contents
Parham M. Kebria, Abbas Khosravi, Syed Moshfeq Salaken and Saeid Nahavandi, "Deep Imitation Learning for Autonomous Vehicles Based on Convolutional Neural Networks," IEEE/CAA J. Autom. Sinica, vol. 7, no. 1, pp. 82-95, Jan. 2020. doi: 10.1109/JAS.2019.1911825
Citation: Parham M. Kebria, Abbas Khosravi, Syed Moshfeq Salaken and Saeid Nahavandi, "Deep Imitation Learning for Autonomous Vehicles Based on Convolutional Neural Networks," IEEE/CAA J. Autom. Sinica, vol. 7, no. 1, pp. 82-95, Jan. 2020. doi: 10.1109/JAS.2019.1911825

Deep Imitation Learning for Autonomous Vehicles Based on Convolutional Neural Networks

doi: 10.1109/JAS.2019.1911825
More Information
  • Providing autonomous systems with an effective quantity and quality of information from a desired task is challenging. In particular, autonomous vehicles, must have a reliable vision of their workspace to robustly accomplish driving functions. Speaking of machine vision, deep learning techniques, and specifically convolutional neural networks, have been proven to be the state of the art technology in the field. As these networks typically involve millions of parameters and elements, designing an optimal architecture for deep learning structures is a difficult task which is globally under investigation by researchers. This study experimentally evaluates the impact of three major architectural properties of convolutional networks, including the number of layers, filters, and filter size on their performance. In this study, several models with different properties are developed, equally trained, and then applied to an autonomous car in a realistic simulation environment. A new ensemble approach is also proposed to calculate and update weights for the models regarding their mean squared error values. Based on design properties, performance results are reported and compared for further investigations. Surprisingly, the number of filters itself does not largely affect the performance efficiency. As a result, proper allocation of filters with different kernel sizes through the layers introduces a considerable improvement in the performance. Achievements of this study will provide the researchers with a clear clue and direction in designing optimal network architectures for deep learning purposes.

     

  • loading
  • [1]
    Y. LeCun, Y. Bengio, and G. Hinton, " Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, May 2015. doi: 10.1038/nature14539
    [2]
    M. Wainberg, D. Merico, A. Delong, and B. J. Frey, " Deep learning in biomedicine,” Nat. Biotechnol., vol. 36, no. 9, pp. 829–838, Oct. 2018. doi: 10.1038/nbt.4233
    [3]
    P. Sermanet, D. Eigen, X. Zhang, M. Mathieu, R. Fergus, and Y. LeCun, " OverFeat: integrated recognition, localization and detection using convolutional networks,” arXiv preprint arXiv: 1312.6229, Dec. 2013.
    [4]
    K. Simonyan and A. Zisserman, " Two-stream convolutional networks for action recognition in videos,” in Proc. 27th Int. Conf. Neural Information Processing Systems, Montreal, Canada, 2014, pp. 568–576.
    [5]
    K. Simonyan and A. Zisserman, " Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv: 1409.1556, Sept. 2014.
    [6]
    M. D. Zeiler and R. Fergus, " Visualizing and understanding convolutional networks,” in Proc. 13th European Conf. Computer Vision, Zurich, Switzerland, 2014, pp. 818–833.
    [7]
    A. Krizhevsky, I. Sutskever, and G. E. Hinton, " ImageNet classification with deep convolutional neural networks,” in Proc. 25th Int. Conf. Neural Information Processing Systems, Lake Tahoe, Nevada, USA, 2012, pp. 1097–1105.
    [8]
    L. Chen, X. M. Hu, T. Xu, H. L. Kuang, and Q. Q. Li, " Turn signal detection during nighttime by CNN detector and perceptual hashing tracking,” IEEE Trans. Intell. Transp. Syst., vol. 18, no. 12, pp. 3303–3314, Dec. 2017. doi: 10.1109/TITS.2017.2683641
    [9]
    Q. Wang, J. Y. Gao, and Y. Yuan, " Embedding structured contour and location prior in siamesed fully convolutional networks for road detection,” IEEE Trans. Intell. Transp. Syst., vol. 19, no. 1, pp. 230–241, Jan. 2018. doi: 10.1109/TITS.2017.2749964
    [10]
    S. P. Zhang, Y. K. Qi, F. Jiang, X. Y. Lan, P. C. Yuen, and H. Y. Zhou, " Point-to-set distance metric learning on deep representations for visual tracking,” IEEE Trans. Intell. Transp. Syst., vol. 19, no. 1, pp. 187–198, Jan. 2018. doi: 10.1109/TITS.2017.2766093
    [11]
    P. M. Kebria, A. Khosravi, S. Nahavandi, Z. Najdovski, and S. J. Hilton, " Neural network adaptive control of teleoperation systems with uncertainties and time-varying delay,” in Proc. 2018 IEEE 14th Int. Conf. Automation Science and Engineering, Munich, Germany, 2018, pp. 252–257.
    [12]
    P. M. Kebria, A. Khosravi, S. Nahavandi, D. R. Wu, and F. Bello, " Adaptive type-2 fuzzy neural-network control for teleoperation systems with delay and uncertainties,” IEEE Trans. Fuzzy Syst.
    [13]
    M. Kuderer, S. Gulati, and W. Burgard, " Learning driving styles for autonomous vehicles from demonstration,” in Proc. 2015 IEEE Int. Conf. Robotics and Automation, Seattle, WA, USA, 2015, pp. 2641–2646.
    [14]
    S. X. Gu, E. Holly, T. Lillicrap, and S. Levine, " Deep reinforcement learning for robotic manipulation with asynchronous off-policy updates,” in Proc. 2017 IEEE Int. Conf. Robotics and Automation, Singapore, 2017, pp. 3389–3396.
    [15]
    B. D. Argall, S. Chernova, M. Veloso, and B. Browning, " A survey of robot learning from demonstration,” Rob. Auton. Syst., vol. 57, no. 5, pp. 469–483, May 2009. doi: 10.1016/j.robot.2008.10.024
    [16]
    A. Hussein, M. M. Gaber, E. Elyan, and C. Jayne, " Imitation learning: a survey of learning methods,” ACM Comput. Surv., vol. 50, no. 2, pp. 21, Jun. 2017.
    [17]
    D. Silver, J. A. Bagnell, and A. Stentz, " Applied imitation learning for autonomous navigation in complex natural terrain,” in Field and Service Robotics, A. Howard, K. Iagnemma, and A. Kelly, Eds. Berlin, Heidelberg, Germany: Springer, 2010, pp. 249–259.
    [18]
    A. Martínez-Tenor, J. A. Fernández-Madrigal, A. Cruz-Martín, and J. González-Jiménez, " Towards a common implementation of reinforcement learning for multiple robotic tasks,” Expert Syst. Appl., vol. 100, pp. 246–259, Jun. 2018. doi: 10.1016/j.eswa.2017.11.011
    [19]
    K. Menda, Y. C. Chen, J. Grana, J. W. Bono, B. D. Tracey, M. J. Kochenderfer, and D. Wolpert, " Deep reinforcement learning for event-driven multi-agent decision processes,” IEEE Trans. Intell. Transp. Syst., vol. 20, no. 4, pp. 1259–1268, Apr. 2019. doi: 10.1109/TITS.2018.2848264
    [20]
    B. Ghazanfari and N. Mozayani, " Extracting bottlenecks for reinforcement learning agent by holonic concept clustering and attentional functions,” Expert Syst. Appl., vol. 54, pp. 61–77, Jul. 2016. doi: 10.1016/j.eswa.2016.01.030
    [21]
    J. Courbon, Y. Mezouar, and P. Martinet, " Autonomous navigation of vehicles from a visual memory using a generic camera model,” IEEE Trans. Intell. Transp. Syst., vol. 10, no. 3, pp. 392–402, Sep. 2009. doi: 10.1109/TITS.2008.2012375
    [22]
    T. H. Zhang, Z. McCarthy, O. Jow, D. Lee, X. Chen, K. Goldberg, and P. Abbeel, " Deep imitation learning for complex manipulation tasks from virtual reality teleoperation,” in Proc. 2018 IEEE Int. Conf. Robotics and Automation, Brisbane, QLD, Australia, 2018, pp. 5628–5635.
    [23]
    W. Sun, A. Venkatraman, G. J. Gordon, B. Boots, and J. A. Bagnell, " Deeply AggreVaTed: Differentiable imitation learning for sequential prediction,” in Proc. 34th Int. Conf. Machine Learning, Sydney, Australia, 2017, pp. 3309–3318.
    [24]
    B. K. Chen, C. Gong, and J. Yang, " Importance-aware semantic segmentation for autonomous vehicles,” IEEE Trans. Intell. Transp. Syst., vol. 20, no. 1, pp. 137–148, Jan. 2019. doi: 10.1109/TITS.2018.2801309
    [25]
    W. Sun, J. A. Bagnell, and B. Boots, " Truncated horizon policy search: Combining reinforcement learning & imitation learning,” in Proc. ICLR 2018 Conf. Acceptance Decision, Vancouver, BC, Canada, 2018.
    [26]
    J. Merel, Y. Tassa, T. B. Dhruva, S. Srinivasan, J. Lemmon, Z. Y. Wang, G. Wayne, and N. Heess, " Learning human behaviors from motion capture by adversarial imitation,” arXiv preprint arXiv: 1707.02201, Jul. 2017.
    [27]
    T. Liu, B. Tian, Y. F. Ai, L. Li, D. P. Cao, and F. Y. Wang, " Parallel reinforcement learning: a framework and case study,” IEEE/CAA J. Autom. Sinica, vol. 5, no. 4, pp. 827–835, Jul. 2018. doi: 10.1109/JAS.2018.7511144
    [28]
    L. Cardamone, D. Loiacono, and P. L. Lanzi, " Learning drivers for torcs through imitation using supervised methods,” Proc. 2009 IEEE Symp. Computational Intelligence and Games Milano,Italy, pp. 148–155, 2009.
    [29]
    J. Ho and S. Ermon, " Generative adversarial imitation learning,” in Proc. 30th Conf. Neural Information Processing Systems, Barcelona, Spain, 2016, pp. 4565–4573.
    [30]
    Y. Duan, M. Andrychowicz, B. Stadie, O. J. Ho, J. Schneider, I. Sutskever, P. Abbeel, and W. Zaremba, " One-shot imitation learning,” in Proc. 31th Conf. Neural Information Processing Systems, Long Beach, CA, USA, 2017, pp. 1087–1098.
    [31]
    B. C. Stadie, P. Abbeel, and I. Sutskever, " Third-person imitation learning,” arXiv preprint arXiv: 1703.01703, Mar. 2017.
    [32]
    J. Saunders, C. L. Nehaniv, and K. Dautenhahn, " Teaching robots by moulding behavior and scaffolding the environment,” in Proc. 1st ACM SIGCHI/SIGART Conf. Human-robot Interaction, Salt Lake City, Utah, USA, 2006, pp. 118–125.
    [33]
    S. Nahavandi, " Trusted autonomy between humans and robots: toward human-on-the-loop in robotics and autonomous systems,” IEEE Syst.,Man,Cybern. Mag., vol. 3, no. 1, pp. 10–17, Jan. 2017. doi: 10.1109/MSMC.2016.2623867
    [34]
    D. Gandhi, L. Pinto, and A. Gupta, " Learning to fly by crashing,” in Proc. IEEE/RSJ Int. Conf. Intelligent Robots and Systems, Vancouver, BC, Canada, 2017, pp. 3948–3955.
    [35]
    M. Mueller, V. Casser, N. Smith, and B. Ghanem, " Teaching UAVs to race using UE4Sim,” arXiv preprint arXiv: 1708.05884, Aug. 2017.
    [36]
    C. Innocenti, H. Lindén, G. Panahandeh, L. Svensson, and N. Mohammadiha, " Imitation learning for vision-based lane keeping assistance,” in Proc. 20th Int. IEEE Conf. Intelligent Transportation Systems, Yokohama, Japan, 2017, pp. 3948–3955.
    [37]
    S. Priesterjahn, O. Kramer, A. Weimer, and A. Goebels, " Evolution of reactive rules in multi player computer games based on imitation,” in Proc. 1st Int. Conf. Natural Computation, Changsha, China, 2005, pp. 744–755.
    [38]
    P. M. Kebria, A. Khosravi, S. M. Salaken, I. Hossain, H. M. D. Kabir, A. Koohestani, R. Alizadehsani, and S. Nahavandi, " Deep imitation learning: the impact of depth on policy performance,” in Proc. 25th Int. Conf. Neural Information Processing, Siem Reap, Cambodia, 2018, pp. 172–181.
    [39]
    A. Diosi, S. Segvic, A. Remazeilles, and F. Chaumette, " Experimental evaluation of autonomous driving based on visual memory and image-based visual servoing,” IEEE Trans. Intell. Transp. Syst., vol. 12, no. 3, pp. 870–883, Sep. 2011. doi: 10.1109/TITS.2011.2122334
    [40]
    H. B. Gao, B. Cheng, J. Q. Wang, K. Q. Li, J. H. Zhao, and D. Y. Li, " Object classification using CNN-based fusion of vision and LIDAR in autonomous vehicle environment,” IEEE Trans. Ind. Inform., vol. 14, no. 9, pp. 4224–4231, Sep. 2018. doi: 10.1109/TII.2018.2822828
    [41]
    P. M. Kebria, R. Alizadehsani, S. M. Salaken, I. Hossain, A. Khosravi, D. Kabir, A. Koohestani, H. Asadi, S. Nahavandi, E. Tunsel, and M. Saif, " Evaluating architecture impacts on deep imitation learning performance for autonomous driving,” in Proc. IEEE Int. Conf. Industrial Technology, Melbourne, Australia, 2019, pp. 865–870.
    [42]
    J. H. Kim, G. Batchuluun, and K. R. Park, " Pedestrian detection based on faster R-CNN in nighttime by fusing deep convolutional features of successive images,” Expert Syst. Appl., vol. 114, pp. 15–33, Dec. 2018. doi: 10.1016/j.eswa.2018.07.020
    [43]
    Y. W. Seo, J. Lee, W. D. Zhang, and D. Wettergreen, " Recognition of highway workzones for reliable autonomous driving,” IEEE Trans. Intell. Transp. Syst., vol. 16, no. 2, pp. 708–718, Apr. 2015.
    [44]
    A. Dominguez-Sanchez, M. Cazorla, and S. Orts-Escolano, " Pedestrian movement direction recognition using convolutional neural networks,” IEEE Trans. Intell. Transp. Syst., vol. 18, no. 12, pp. 3540–3548, Dec. 2017. doi: 10.1109/TITS.2017.2726140
    [45]
    S. Di, H. G. Zhang, C. G. Li, X. Mei, D. Prokhorov, and H. B. Ling, " Cross-domain traffic scene understanding: a dense correspondence-based transfer learning approach,” IEEE Trans. Intell. Transp. Syst., vol. 19, no. 3, pp. 745–757, Mar. 2018. doi: 10.1109/TITS.2017.2702012
    [46]
    Y. J. Zeng, X. Xu, D. Y. Shen, Y. Q. Fang, and Z. P. Xiao, " Traffic sign recognition using kernel extreme learning machines with deep perceptual features,” IEEE Trans. Intell. Transp. Syst., vol. 18, no. 6, pp. 1647–1653, Jun. 2017.
    [47]
    Q. Wang, J. Y. Gao, and Y. Yuan, " A joint convolutional neural networks and context transfer for street scenes labeling,” IEEE Trans. Intell. Transp. Syst., vol. 19, no. 5, pp. 1457–1470, May 2018. doi: 10.1109/TITS.2017.2726546
    [48]
    J. Yosinski, J. Clune, Y. Bengio, and H. Lipson, " How transferable are features in deep neural networks?,” in Proc. 27th Int. Conf. Neural Information Processing Systems, Montreal, Canada, 2014, 3320–3328.
    [49]
    Z. H. Zhou, J. X. Wu, and W. Tang, " Ensembling neural networks: many could be better than all,” Artif. Intell., vol. 137, no. 1-2, pp. 239–263, May 2002. doi: 10.1016/S0004-3702(02)00190-X
    [50]
    L. Breiman, " Bagging predictors,” Mach. Learn., vol. 24, no. 2, pp. 123–140, Aug. 1996.
    [51]
    T. G. Dietterich, " An experimental comparison of three methods for constructing ensembles of decision trees: bagging, boosting, and randomization,” Mach. Learn., vol. 40, no. 2, pp. 139–157, Aug. 2000. doi: 10.1023/A:1007607513941
    [52]
    C. Strobl, J. Malley, and G. Tutz, " An introduction to recursive partitioning: rationale, application and characteristics of classification and regression trees, bagging, and random forests,” Psychol. Methods, vol. 14, no. 4, pp. 323–348, Dec. 2009. doi: 10.1037/a0016973
    [53]
    J. Mendes-Moreira, C. Soares, A. M. Jorge, and J. F. De Sousa, " Ensemble approaches for regression: a survey,” ACM Comput. Surv., vol. 14, no. 1, pp. 10, Nov. 2012.
    [54]
    Udacity. Self-drivign-car. (2018) [Online]. Available: https://www.udacity.com/
    [55]
    V. Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusu, J. Veness, M. G. Bellemare, A. Graves, M. Riedmiller, A. K. Fidjeland, G. Ostrovski, S. Petersen, C. Beattie, A. Sadik, I. Antonoglou, H. King, D. Kumaran, D. Wierstra, S. Legg, and D. Hassabis, " Human-level control through deep reinforcement learning,” Nature, vol. 518, no. 7540, pp. 529–533, 2015. doi: 10.1038/nature14236
    [56]
    T. Mareda, L. Gaudard, and F. Romerio, " A parametric genetic algorithm approach to assess complementary options of large scale windsolar coupling,” IEEE/CAA J. Autom. Sinica, vol. 4, no. 2, pp. 260–272, Apr. 2017. doi: 10.1109/JAS.2017.7510523
    [57]
    Q. Kang, X. Y. Song, M. C. Zhou, and L. Li, " A collaborative resource allocation strategy for decomposition-based multiobjective evolutionary algorithms,” IEEE Trans. Syst.,Man,Cybern.:Syst., vol. 49, no. 12, pp. 2416–2423, Dec. 2019.

Catalog

    通讯作者: 陈斌, bchen63@163.com
    • 1. 

      沈阳化工大学材料科学与工程学院 沈阳 110142

    1. 本站搜索
    2. 百度学术搜索
    3. 万方数据库搜索
    4. CNKI搜索

    Figures(11)  / Tables(3)

    Article Metrics

    Article views (14273) PDF downloads(341) Cited by()

    Highlights

    • A comprehensive evaluation and comparison of the three major architectural parameters, including the number of layers, filters, and kernel size in the design of a CNN, and their impact on the network’s overall performance. This comparison gives the researchers an overview of the most effective way to optimally design their deep networks to achieve the best possible performance.
    • A new MSE-based ensemble methodology for regression problems that improves the performance according to the average performance of each model throughout the previous observation samples.
    • As a popular ensemble approach, Bagging method is also considered to comparatively illustrate the superiority of the proposed ensemble approach.
    • Demonstrative comparison between the developed models provides the information about the impact of design parameters on the overall performance which leads to optimal structures for better performances.

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return