Development of Behavior-Based Control Systems for Autonomous Vehicle Navigation in Urban Environment
Abstract
In this research, we developed new motion planning system for autonomous vehicles navigationin real-time, capable to maneuver around moving and stationary objects and avoid hitting pedestrian. The developed system consisted of three components 1) pedestrian's trajectories prediction using generalized regression neural network, 2) safe zone navigation controller using fuzzy logic, and 3) Adaptive model predictive controller. Following the system engineering approach, the proposed system was composed into two subsystems: Perception subsystem, and Planning and Control subsystem. All subsystems were developed, verified, then integrated, tested, and validated in this dissertation. To evaluate the proposed system, different driving scenarios were conducted, including different vehicle speeds, different pedestrian's speeds, and directions. The simulation results showed the system’s ability to predict the pedestrian’s trajectories, and the ability to maneuver the vehicle and avoid collision while maintaining a safe distance and speed.
Subject Area
Computer science|Transportation
Recommended Citation
Abdalla Hasan Al-Salah,
"Development of Behavior-Based Control Systems for Autonomous Vehicle Navigation in Urban Environment"
(2022).
ETD Collection for Tennessee State University.
Paper AAI29260009.
https://digitalscholarship.tnstate.edu/dissertations/AAI29260009