Noise gradient strategy for an enhanced hybrid convolutional-recurrent deep network to control a self-driving vehicle

Abstract In this paper a noise gradient strategy on the Adam optimizer is introduced, in order to reduce the training time of our enhanced Chauffeur hybrid deep model. This neural network was modified to take into account the time dependence of the input visual information from a time-distributed convolution, with the aim of increasing the autonomy of a self-driving vehicle. The effectiveness of the proposed optimizer and model was evaluated and quantified during training and validation with a higher performance than the original Chauffeur model in combination with the comparative optimizers. In terms of the autonomy, it can be seen that our enhanced Hybrid Convolutional-Recurrent Deep Network was better trained, achieving autonomy greater than 95% with a minimum number of human interventions.

[1]  Y. Nesterov.  A method for unconstrained convex minimization problem with the rate of convergence o(1/k^2) , 1983 .

[2]  Fei Xiao,et al.  Predictions of short-term driving intention using recurrent neural network on sequential data , 2018, ArXiv.

[3]  Torsten Sattler,et al.  3D visual perception for self-driving cars using a multi-camera system: Calibration, mapping, localization, and obstacle detection , 2017, Image Vis. Comput..

[4]  Quoc V. Le,et al.  Adding Gradient Noise Improves Learning for Very Deep Networks , 2015, ArXiv.

[5]  Xin Zhang,et al.  End to End Learning for Self-Driving Cars , 2016, ArXiv.

[6]  Robert Jenssen,et al.  Recurrent Neural Networks for Short-Term Load Forecasting , 2017, SpringerBriefs in Computer Science.

[7]  Suman Jana,et al.  DeepTest: Automated Testing of Deep-Neural-Network-Driven Autonomous Cars , 2017, 2018 IEEE/ACM 40th International Conference on Software Engineering (ICSE).

[8]  Nanning Zheng,et al.  Brain-Inspired Cognitive Model With Attention for Self-Driving Cars , 2017, IEEE Transactions on Cognitive and Developmental Systems.

[9]  Sebastian Ruder,et al.  An overview of gradient descent optimization algorithms , 2016, Vestnik komp'iuternykh i informatsionnykh tekhnologii.

[10]  Jitendra Malik,et al.  Region-Based Convolutional Networks for Accurate Object Detection and Segmentation , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[11]  Jonathan Sprinkle,et al.  Fuzzy control of an autonomous car using a smart phone , 2017, 2017 CHILEAN Conference on Electrical, Electronics Engineering, Information and Communication Technologies (CHILECON).

[12]  Matthew D. Zeiler.  ADADELTA: An Adaptive Learning Rate Method , 2012, ArXiv.

[13]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[14]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[15]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[16]  Abdulrahman H. A. Widaa,et al.  Design of Fuzzy-based autonomous car control system , 2017, 2017 International Conference on Communication, Control, Computing and Electronics Engineering (ICCCCEE).

[17]  Andreas Stafylopatis,et al.  Autonomous vehicle navigation using evolutionary reinforcement learning , 1998, Eur. J. Oper. Res..

[18]  Yoram Singer,et al.  Adaptive Subgradient Methods for Online Learning and Stochastic Optimization , 2011, J. Mach. Learn. Res..

[19]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[20]  Timothy Dozat,et al.  Incorporating Nesterov Momentum into Adam , 2016 .

[21]  Dante Mújica-Vargas,et al.  Non-parametric Brain Tissues Segmentation via a Parallel Architecture of CNNs , 2019, MCPR.

[22]  Quoc V. Le,et al.  Sequence to Sequence Learning with Neural Networks , 2014, NIPS.

[23]  Ning Qian,et al.  On the momentum term in gradient descent learning algorithms , 1999, Neural Networks.

[24]  Simon Andermatt,et al.  Multi-dimensional Gated Recurrent Units for Automated Anatomical Landmark Localization , 2017, ArXiv.

[25]  Didrik Nielsen,et al.  Fast and Scalable Bayesian Deep Learning by Weight-Perturbation in Adam , 2018, ICML.

[26]  Yoshua Bengio,et al.  Neural Machine Translation by Jointly Learning to Align and Translate , 2014, ICLR.

[27]  Phi Vu Tran,et al.  A Fully Convolutional Neural Network for Cardiac Segmentation in Short-Axis MRI , 2016, ArXiv.

[28]  Heechul Yun,et al.  DeepPicar: A Low-Cost Deep Neural Network-Based Autonomous Car , 2017, 2018 IEEE 24th International Conference on Embedded and Real-Time Computing Systems and Applications (RTCSA).

[29]  Xiaodong Cui,et al.  English Conversational Telephone Speech Recognition by Humans and Machines , 2017, INTERSPEECH.

[30]  Lawrence D. Jackel,et al.  Explaining How a Deep Neural Network Trained with End-to-End Learning Steers a Car , 2017, ArXiv.

[31]  Pascal Fua,et al.  Efficient large-scale multi-view stereo for ultra high-resolution image sets , 2011, Machine Vision and Applications.

[32]  Koen E. A. van de Sande,et al.  Selective Search for Object Recognition , 2013, International Journal of Computer Vision.

[33]  Misha Denil,et al.  Noisy Activation Functions , 2016, ICML.

[34]  Andrew Gordon Wilson,et al.  Learning Scalable Deep Kernels with Recurrent Structure , 2016, J. Mach. Learn. Res..