scholarly journals End-to-End Autonomous Driving Through Dueling Double Deep Q-Network

2021 ◽  
Author(s):  
Baiyu Peng ◽  
Qi Sun ◽  
Shengbo Eben Li ◽  
Dongsuk Kum ◽  
Yuming Yin ◽  
...  
Keyword(s):  
Neural Network ◽  
State Space ◽  
Learning Algorithm ◽  
Rapid Development ◽  
Autonomous Driving ◽  
Saliency Map ◽  
Hierarchical Architecture ◽  
Link Type ◽  
The Neural Network ◽  
End To End

AbstractRecent years have seen the rapid development of autonomous driving systems, which are typically designed in a hierarchical architecture or an end-to-end architecture. The hierarchical architecture is always complicated and hard to design, while the end-to-end architecture is more promising due to its simple structure. This paper puts forward an end-to-end autonomous driving method through a deep reinforcement learning algorithm Dueling Double Deep Q-Network, making it possible for the vehicle to learn end-to-end driving by itself. This paper firstly proposes an architecture for the end-to-end lane-keeping task. Unlike the traditional image-only state space, the presented state space is composed of both camera images and vehicle motion information. Then corresponding dueling neural network structure is introduced, which reduces the variance and improves sampling efficiency. Thirdly, the proposed method is applied to The Open Racing Car Simulator (TORCS) to demonstrate its great performance, where it surpasses human drivers. Finally, the saliency map of the neural network is visualized, which indicates the trained network drives by observing the lane lines. A video for the presented work is available online, https://youtu.be/76ciJmIHMD8 or https://v.youku.com/v_show/id_XNDM4ODc0MTM4NA==.html.

scholarly journals An End-to-End Deep Neural Network for Autonomous Driving Designed for Embedded Automotive Platforms

Sensors ◽  
2019 ◽  
Vol 19 (9) ◽  
pp. 2064 ◽  
Author(s):  
Jelena Kocić ◽  
Nenad Jovičić ◽  
Vujo Drndarević
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Convolutional Neural Networks ◽  
Network Architecture ◽  
Deep Neural Network ◽  
Learning Algorithm ◽  
Autonomous Driving ◽  
Frame Rate ◽  
Power Cost ◽  
End To End

In this paper, one solution for an end-to-end deep neural network for autonomous driving is presented. The main objective of our work was to achieve autonomous driving with a light deep neural network suitable for deployment on embedded automotive platforms. There are several end-to-end deep neural networks used for autonomous driving, where the input to the machine learning algorithm are camera images and the output is the steering angle prediction, but those convolutional neural networks are significantly more complex than the network architecture we are proposing. The network architecture, computational complexity, and performance evaluation during autonomous driving using our network are compared with two other convolutional neural networks that we re-implemented with the aim to have an objective evaluation of the proposed network. The trained model of the proposed network is four times smaller than the PilotNet model and about 250 times smaller than AlexNet model. While complexity and size of the novel network are reduced in comparison to other models, which leads to lower latency and higher frame rate during inference, our network maintained the performance, achieving successful autonomous driving with similar efficiency compared to autonomous driving using two other models. Moreover, the proposed deep neural network downsized the needs for real-time inference hardware in terms of computational power, cost, and size.

scholarly journals Stock Market Prediction Using Artificial Neural Networks

2012 ◽  
Vol 6-7 ◽  
pp. 1055-1060 ◽  
Author(s):  
Yang Bing ◽  
Jian Kun Hao ◽  
Si Chang Zhang
Keyword(s):  
Neural Network ◽  
Prediction Models ◽  
Learning Algorithm ◽  
Stock Exchange ◽  
Back Propagation ◽  
Composite Index ◽  
Network Models ◽  
Back Propagation Neural Network ◽  
Neural Network Models ◽  
The Neural Network

In this study we apply back propagation Neural Network models to predict the daily Shanghai Stock Exchange Composite Index. The learning algorithm and gradient search technique are constructed in the models. We evaluate the prediction models and conclude that the Shanghai Stock Exchange Composite Index is predictable in the short term. Empirical study shows that the Neural Network models is successfully applied to predict the daily highest, lowest, and closing value of the Shanghai Stock Exchange Composite Index, but it can not predict the return rate of the Shanghai Stock Exchange Composite Index in short terms.

scholarly journals Robotic grasp detection using a novel two-stage approach

2021 ◽  
Vol 1 (1) ◽  
pp. 19-29
Author(s):  
Zhe Chu ◽  
Mengkai Hu ◽  
Xiangyu Chen
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Network Models ◽  
Particle Swarm Optimizer ◽  
Neural Network Models ◽  
Two Stage ◽  
The Neural Network ◽  
End To End ◽  
Small Change ◽  
Robotic Grasp

Recently, deep learning has been successfully applied to robotic grasp detection. Based on convolutional neural networks (CNNs), there have been lots of end-to-end detection approaches. But end-to-end approaches have strict requirements for the dataset used for training the neural network models and it’s hard to achieve in practical use. Therefore, we proposed a two-stage approach using particle swarm optimizer (PSO) candidate estimator and CNN to detect the most likely grasp. Our approach achieved an accuracy of 92.8% on the Cornell Grasp Dataset, which leaped into the front ranks of the existing approaches and is able to run at real-time speeds. After a small change of the approach, we can predict multiple grasps per object in the meantime so that an object can be grasped in a variety of ways.

Design and Realization of Experimental Autonomous Driving System Based on Neural Network Control

2012 ◽  
Vol 241-244 ◽  
pp. 1953-1958
Author(s):  
Qing Fu Kong ◽  
Fan Ming Zeng ◽  
Jie Chang Wu ◽  
Jia Ming Wu
Keyword(s):  
Neural Network ◽  
Autonomous Driving ◽  
Network Control ◽  
Neural Network Control ◽  
Test Results ◽  
Driving System ◽  
Neural Network Controller ◽  
The Neural Network ◽  
Network Controller ◽  
Autonomous Driving System

Intelligent vehicle is an attractive solution to the traffic problems caused by automobiles. An experimental autonomous driving system based on a slot car set is designed and realized in the paper. With the application of a wireless camera equipped on the slot car, the track information is acquired and sent to the controlling computer. A backpropogation (BP) neural network controller is built to imitate the way of player’s thinking. After being trained, the neural network controller can give the proper voltage instructions to the direct current (DC) motor of the slot car according to different track conditions. Test results prove that the development of the autonomous driving system is successful.

scholarly journals Use of a Multilayer Perceptron to Automate Terrain Assessment for the Needs of the Armed Forces

2018 ◽  
Vol 7 (11) ◽  
pp. 430 ◽  
Author(s):  
Krzysztof Pokonieczny
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Multilayer Perceptron ◽  
Learning Algorithm ◽  
Armed Forces ◽  
Shuttle Radar Topography Mission ◽  
Terrain Classification ◽  
The Neural Network ◽  
Data Configuration ◽  
Trained Neural Network

The classification of terrain in terms of passability plays a significant role in the process of military terrain assessment. It involves classifying selected terrain to specific classes (GO, SLOW-GO, NO-GO). In this article, the problem of terrain classification to the respective category of passability was solved by applying artificial neural networks (multilayer perceptron) to generate a continuous Index of Passability (IOP). The neural networks defined this factor for primary fields in two sizes (1000 × 1000 m and 100 × 100 m) based on the land cover elements obtained from Vector Smart Map (VMap) Level 2 and Shuttle Radar Topography Mission (SRTM). The work used a feedforward neural network consisting of three layers. The paper presents a comprehensive analysis of the reliability of the neural network parameters, taking into account the number of neurons, learning algorithm, activation functions and input data configuration. The studies and tests carried out have shown that a well-trained neural network can automate the process of terrain classification in terms of passability conditions.

MPID Control Tuning for a Flexible Manipulator Using a Neural Network

2010 ◽  
Vol 22 (1) ◽  
pp. 82-90 ◽  
Author(s):  
Tamer Mansour ◽  
◽  
Atsushi Konno ◽  
Masaru Uchiyama
Keyword(s):  
Neural Network ◽  
Vibration Control ◽  
Learning Algorithm ◽  
The Other ◽  
Gradient Algorithm ◽  
Flexible Manipulator ◽  
Computational Time ◽  
Speed Of Response ◽  
Control Gain ◽  
The Neural Network

This paper studies the use of neural networks as a tuning tool for the gain in Modified Proportional-Integral-Derivative (MPID) control used to control a flexible manipulator. The vibration control gain in the MPID controller has been determined in an empirical way so far. It is a considerable time consuming process because the vibration control performance depends not only on the vibration control gain but also on the other parameters such as the payload, references and PD joint servo gains. Hence, the vibration control gain must be tuned considering the other parameters. In order to find optimal vibration control gain for the MPID controller, a neural network based approach is proposed in this paper. The proposed neural network finds an optimum vibration control gain that minimizes a criteria function. The criteria function is selected to represent the effect of the vibration of the end effector in addition to the speed of response. The scaled conjugate gradient algorithm is used as a learning algorithm for the neural network. Tuned gain response results are compared to results for other types of gains. The effectiveness of using the neural network appears in the reduction of the computational time and the ability to tune the gain with different loading condition.

A Solution for Two-Dimensional Mazes with Use of Chaotic Dynamics in a Recurrent Neural Network Model

2004 ◽  
Vol 16 (9) ◽  
pp. 1943-1957 ◽  
Author(s):  
Yoshikazu Suemitsu ◽  
Shigetoshi Nara
Keyword(s):  
Neural Network ◽  
State Space ◽  
Network Model ◽  
Neural Network Model ◽  
Chaotic Dynamics ◽  
Dimensional Space ◽  
Dynamical Structure ◽  
Two Dimensional ◽  
Time Step ◽  
The Neural Network

Chaotic dynamics introduced into a neural network model is applied to solving two-dimensional mazes, which are ill-posed problems. A moving object moves from the position at t to t + 1 by simply defined motion function calculated from firing patterns of the neural network model at each time step t. We have embedded several prototype attractors that correspond to the simple motion of the object orienting toward several directions in two-dimensional space in our neural network model. Introducing chaotic dynamics into the network gives outputs sampled from intermediate state points between embedded attractors in a state space, and these dynamics enable the object to move in various directions. System parameter switching between a chaotic and an attractor regime in the state space of the neural network enables the object to move to a set target in a two-dimensional maze. Results of computer simulations show that the success rate for this method over 300 trials is higher than that of random walk. To investigate why the proposed method gives better performance, we calculate and discuss statistical data with respect to dynamical structure.

scholarly journals FORECASTING OF THE SURFACE ROUGHNESS IN FINISHING MILLING USING NEURAL NETWORKS

2019 ◽  
Vol 4 (10) ◽  
pp. 135-141
Author(s):  
Е. Ерыгин ◽  
E. Erygin ◽  
Т. Дуюн ◽  
T. Duyun
Keyword(s):  
Neural Network ◽  
Cutting Tool ◽  
Learning Algorithm ◽  
Network Modeling ◽  
Training Sample ◽  
Neural Network Modeling ◽  
Network Learning ◽  
The Neural Network ◽  
And Training ◽  
Positive Results

This article describes the task of predicting roughness when finishing milling using neural network modeling. As a basis for the creation and training of an artificial neural network, a progressive formu-la for determining the roughness during finishing milling is chosen. The thermoEMF of the processing and processed materials is used as one of the parameters for calculating the roughness. The use of thermoEMF allows to take into account the material of the workpiece and the cutting tool, which af-fects the accuracy of the results. A training sample is created with data for five inputs and one output. The architecture, features and network learning algorithm are described. A neural network that de-termines the roughness for finishing milling has been created and configured. The process of learning and debugging of the neural network by means of graphs is clearly displayed. The network operability is checked on the test data, which allows obtaining positive results.

Sign in / Sign up

Export Citation Format

Share Document