Browse > Article
http://dx.doi.org/10.7236/JIIBC.2022.22.6.75

A Comparative Analysis of Reinforcement Learning Activation Functions for Parking of Autonomous Vehicles  

Lee, Dongcheul (Hannam University)
Publication Information
The Journal of the Institute of Internet, Broadcasting and Communication / v.22, no.6, 2022 , pp. 75-81 More about this Journal
Abstract
Autonomous vehicles, which can dramatically solve the lack of parking spaces, are making great progress through deep reinforcement learning. Activation functions are used for deep reinforcement learning, and various activation functions have been proposed, but their performance deviations were large depending on the application environment. Therefore, finding the optimal activation function depending on the environment is important for effective learning. This paper analyzes 12 functions mainly used in reinforcement learning to compare and evaluate which activation function is most effective when autonomous vehicles use deep reinforcement learning to learn parking. To this end, a performance evaluation environment was established, and the average reward of each activation function was compared with the success rate, episode length, and vehicle speed. As a result, the highest reward was the case of using GELU, and the ELU was the lowest. The reward difference between the two activation functions was 35.2%.
Keywords
Autonomous Vehicle; Parking; Reinforcement Learning;
Citations & Related Records
Times Cited By KSCI : 1  (Citation Analysis)
연도 인용수 순위
1 T. Jiang, J. Cheng, "Target Recognition Based on CNN with LeakyReLU and PReLU Activation Functions," 2019 International Conference on Sensing, Diagnostics, Prognostics, and Control, pp. 718-722, 2019.DOI: https://doi.org/10.1109/SDPC.2019.00136   DOI
2 D. Jeon, D. Park, "Malware Detection in Encrypted TLS Traffic using Machine Learning Techniques," Journal of KIIT, Vol. 19, No. 10, pp. 125-136, 2021. DOI: http://dx.doi.org/10.14801/jkiit.2021.19.10.125   DOI
3 H. Cho, H. Shin, "Trading Strategies Using Reinforcement Learning," Journal of the Korea Academia-Industrial cooperation Society, Vol. 22, No. 1, pp. 123-130, 2021. DOI: https://doi.org/10.5762/KAIS.2021.22.1.123   DOI
4 M. Lau, K. Lim, "Review of Adaptive Activation Function in Deep Neural Network", 2018 IEEE-EMBS Conference on Biomedical Engineering and Sciences, pp. 686-690, 2018. DOI: https://doi.org/10.1109/IECBES.2018.8626714   DOI
5 M. Roodschild, J.G. Sardinas, A. Will, "A new approach for the vanishing gradient problem on sigmoid activation," Progress in Artificial Intelligence, Vol. 9, pp. 351-360, 2020. DOI: https://doi.org/10.1007/s13748-020-00218-y   DOI
6 A. Isac, C. J. Frederico, D. Kragic, J.A. Stork, "The effect of Target Normalization and Momentum on Dying ReLU," Proceedings of the 32nd annual workshop of the Swedish Artificial Intelligence Society, 2020. DOI: https://doi.org/10.48550/arXiv.2005.06195   DOI
7 X. Glorot, A. Bordes, Y. Bengio, "Deep sparse rectifier neural networks," Proceedings of the 14th International Conference on Artificial Intelligence and Statistics, PMLR, Vol. 15, pp. 315-323, 2011.
8 D. Clevert, T. Unterthiner, S. Hochreiter, "Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)," Proceedings of 4th International Conference on Learning Representations (ICLR), 2016. DOI: https://doi.org/10.48550/arXiv.1511.07289   DOI
9 K. He, X. Zhang, S. Ren, J. Sun, "Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification," Proceedings of IEEE International Conference on Computer Vision, 2015. DOI: https://doi.org/10.1109/iccv.2015.123   DOI
10 D. Hendrycks, K. Gimpel, "Gaussian Error Linear Units," arXiv:1606.08415, 2016. DOI: https://doi.org/10.48550/arXiv.1606.08415   DOI
11 G. Klambauer, T. Unterthiner, A. Mayr, S. Hochreiter, "Self-Normalizing Neural Networks," Proceedings of 31st Conference on Neural Information Processing Systems (NIPS), 2017. DOI: https://doi.org/10.48550/arXiv.1706.02515   DOI
12 B. Jonathan, "Continuously Differentiable Exponential Linear Units," arXiv:1704.07483, 2017. DOI: https://doi.org/10.48550/arXiv.1704.07483   DOI
13 S. Elfwing, E. Uchibe, K. Doya. "Sigmoid-weighted linear units for neural network function approximation in reinforcement learning," Neural Networks, Vol. 107, pp. 3-11, 2018. DOI: https://doi.org/10.1016/j.neunet.2017.12.012   DOI
14 D. Misra, "Mish: A Self Regularized Non-Monotonic Activation Function", arXiv:1908.08681, 2020. DOI: https://doi.org/10.48550/arXiv.1908.08681   DOI
15 H. Andrew, S. Mark, G. Chu, L. Chen, B. Chen, M. Tan, W. Wang, Y. Zhu, R. Pang, V. Vasudevan, Q.V. Le, H. Adam, "Searching for MobileNetV3," Proceedings of IEEE/CVF international conference on computer vision, 2019. DOI: https://doi.org/10.48550/arXiv.1905.02244   DOI
16 A. Kuznetsov, P. Shvechikov, A. Grishin, D. Vetrov, "Controlling Overestimation Bias with Truncated Mixture of Continuous Distributional Quantile Critics," Proceedings of the 37 th International Conference on Machine Learning (PMLR), 2020. DOI: https://doi.org/10.48550/arXiv.2005.04269   DOI
17 Leurent, Edouard, "An Environment for Autonomous Driving Decision-Making," GitHub repository: https://github.com/eleurent/highway-env, 2018.
18 D. Shim, J. Yang, J. Son, S. Han, H. Lee, "Smart Parking Guidance System based on IoT Car-stoppers," The Journal of The Institute of Internet, Broadcasting and Communication, Vol. 17, No. 3, pp. 137-143, 2017. DOI: https://doi.org/10.7236/JIIBC.2017.17.3.137   DOI