Model predictive path-following controller for Generalised N-Trailer vehicles with noisy sensors and disturbances

Abstract

Generalised N-Trailer (GNT) vehicles, commonly used in agriculture, mining, and industry, consist of a tractor pulling multiple passive trailers. This paper presents a nonlinear moving horizon estimator (NMHE) and nonlinear model predictive controller (NMPC) for GNT robotic vehicles. The NMHE accurately estimates the vehicle’s state under challenging conditions such as noisy measurements, disturbances, and different soil conditions, without relying on assumptions about disturbances, making it more effective than techniques like the Extended Kalman Filter (EKF). Similarly, the NMPC successfully steers the GNT along a predefined path with lower control effort compared to existing algorithms, thanks to smoother tractor manoeuvres that minimise energy in the objective function. Moreover, unlike other methods, this approach computes the tractors’ velocities instead of those of the last trailer, eliminating the need for kinematic inversion and making it suitable for various vehicle configurations. Efficient solvers for the NMHE and NMPC problems were generated using two different open-source frameworks. The proposed framework was evaluated through challenging simulated studies and field experiments. Additional resources, including code implementation and field experiment videos, can be found at https://drive.google.com/drive/folders/1n-qVWJA40cZn-WiDpwXhhF9AdK54LTfX?usp=sharing.