Model Fusion: Weighted N-Version Programming for Resilient Autonomous Vehicle Steering Control

2018 ◽  
Author(s):  
Ailec Wu ◽  
Abu Hasnat Mohammad Rubaiyat ◽  
Chris Anton ◽  
Homa Alemzadeh
Keyword(s):  
Autonomous Vehicle ◽  
Steering Control ◽  
Model Fusion

Active steering control for autonomous vehicles based on a driver-in-the-loop platform: A case study of collision avoidance

2019 ◽  
Vol 233 (10) ◽  
pp. 1422-1437
Author(s):  
Keji Chen ◽  
Xiaofei Pei ◽  
Daoyuan Sun ◽  
Zhenfu Chen ◽  
Xuexun Guo ◽  
...  
Keyword(s):  
Risk Assessment ◽  
Model Predictive Control ◽  
Collision Avoidance ◽  
Predictive Control ◽  
Autonomous Vehicle ◽  
Steering Control ◽  
Safe Driving ◽  
Active Steering ◽  
Active Steering Control

Leveraging the advancements in sensor and mapping technologies, the collision-free autonomous vehicle becomes possible in the future. In this article, a case study of collision avoidance by active steering control is presented and verified by a driver-in-the-loop platform. The proposed control system integrates a risk assessment algorithm and a hierarchical model predictive control approach to ensure a safe driving. First, a fuzzy logic is used to estimate the potential conflict. Besides, a nonlinear model predictive control is introduced in the upper layer of the model predictive controller to generate a collision-free trajectory. Furthermore, the lower layer determines the optimal steering angle based on the linear time-variant model predictive control to follow the replanning path. The performance of the controller has been evaluated in the real-time driver-in-the-loop test. The results show that the autonomous vehicle is able to avoid the collision with the surrounding vehicle that is operated by a real driver, and the performance of collision avoidance is improved by means of the risk assessment.

scholarly journals Simulated Sensor Based Strategies for Obstacle Avoidance Using Velocity Profiling for Autonomous Vehicle FURBOT

Electronics ◽  
2020 ◽  
Vol 9 (6) ◽  
pp. 883 ◽  
Author(s):  
Khayyam Masood ◽  
Rezia Molfino ◽  
Matteo Zoppi
Keyword(s):  
Mathematical Model ◽  
Velocity Profile ◽  
Obstacle Avoidance ◽  
Autonomous Vehicle ◽  
Velocity Control ◽  
Steering Control ◽  
Traffic Lights ◽  
Simulated Environment ◽  
Last Mile ◽  
Robotic Vehicle

Freight Urban Robotic Vehicle (FURBOT) is an autonomous vehicle designed to transport last mile freight to designated urban stations. It is a slow vehicle designed to tackle urban environment with complete autonomy. A slow vehicle may have slightly different strategies for avoiding obstacles. Unlike on a highway, it has to deal with pedestrians, traffic lights and slower vehicles while maintaining smoothness in its drive. To tackle obstacle avoidance for this vehicle, sensor feedback based strategies have been formulated for smooth drive and obstacle avoidance. A full mathematical model for the vehicle is formulated and simulated in MATLAB environment. The mathematical model uses velocity control for obstacle avoidance without steering control. The obstacle avoidance is attained through velocity control and strategies are formulated with velocity profiling. Innovative techniques are formulated in creating the simulated sensory feed-backs of the environment. Using these feed-backs, correct velocity profiling is autonomously created for giving velocity profile input to the velocity controller. Proximity measurements are assumed to be available for the vehicle in its given range of drive. Novelty is attained by manipulating velocity profile without prior knowledge of the environment. Four different type of obstacles are modeled for simulated environment of the vehicle. These obstacles are randomly placed in the path of the vehicle and autonomous velocity profiling is verified in simulated environment. The simulated results obtained show satisfactory velocity profiling for controller input. The current technique helps to tune the existing controller and in designing of a better velocity controller for the autonomous vehicle and bridges the gap between sensor feed-back and controller input. Moreover, accurate input profiling creates less strain on the system and brings smoothness in drive for an overall safer environment.

scholarly journals Spatial Based Deep Learning Autonomous Wheel Robot Using CNN

2020 ◽  
Vol 11 (3) ◽  
pp. 167
Author(s):  
Eko Wahyu Prasetyo ◽  
Nambo Hidetaka ◽  
Dwi Arman Prasetya ◽  
Wahyu Dirgantara ◽  
Hari Fitria Windi
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Autonomous Vehicle ◽  
Image Data ◽  
Steering Control ◽  
Accuracy Rate ◽  
Artificial Intelligent ◽  
The Third ◽  
Network Method ◽  
Wheeled Vehicles

The development of technology is growing rapidly; one of the most popular among the scientist is robotics technology. Recently, the robot was created to resemble the function of the human brain. Robots can make decisions without being helped by humans, known as AI (Artificial Intelligent). Now, this technology is being developed so that it can be used in wheeled vehicles, where these vehicles can run without any obstacles. Furthermore, of research, Nvidia introduced an autonomous vehicle named Nvidia Dave-2, which became popular. It showed an accuracy rate of 90%. The CNN (Convolutional Neural Network) method is used in the track recognition process with input in the form of a trajectory that has been taken from several angles. The data is trained using Jupiter's notebook, and then the training results can be used to automate the movement of the robot on the track where the data has been retrieved. The results obtained are then used by the robot to determine the path it will take. Many images that are taken as data, precise the results will be, but the time to train the image data will also be longer. From the data that has been obtained, the highest train loss on the first epoch is 1.829455, and the highest test loss on the third epoch is 30.90127. This indicates better steering control, which means better stability.

Sign in / Sign up

Export Citation Format

Share Document