loading page

Nonlinear Networked Predictive Control for Unmanned Ground Vehicles
  • +3
  • Subhan Khan ,
  • Yonghui Li ,
  • Jose Guivant ,
  • Dusit Niyato ,
  • Xuesong Li ,
  • Wanchun Liu
Subhan Khan
University of Sydney

Corresponding Author:[email protected]

Author Profile
Yonghui Li
Author Profile
Jose Guivant
Author Profile
Dusit Niyato
Author Profile
Xuesong Li
Author Profile
Wanchun Liu
Author Profile

Abstract

In this paper, we explore the application of nonlinear networked predictive control (NNPC) to the unmanned ground vehicle (UGV) that experiences skid-slip, static and moving obstacles, and communication delays in both sensor-to-controller (SC) and controller-to-actuator (CA) channels. Our approach utilizes adaptive dynamic programming (ADP) based model predictive control (MPC) and introduces a delay compensation strategy, effectively mitigating the impact of communication delays on control performance. Specifically, our proposed method generates a control sequence that accounts for delayed control inputs and transmits it to the actuator when the controller node receives a new state measurement. To ensure robustness, our method explicitly considers the presence of collision avoidance and skid-slip and employs an ADP process to develop the control law, which receives state estimation from the Extended Kalman filter (EKF). By integrating these elements, our approach provides an effective and reliable solution for the path-tracking control of the UGV in the presence of communication delays and external disturbances. Finally, comprehensive simulations are performed on a large map and traditional circular trajectory to test and validate the performance of the proposed control scheme.