• Proposed an NMPC trajectory tracking control method based on front–rear Ackermann kinematic modeling. • Identified the limitation of fixed-weight NMPC under complex trajectory variations and introduced a curvature-adaptive NMPC strategy. • Developed an Adaptive NMPC controller for 4WID-4WIS robots on unpaved terrain, achieving continuous trajectory tracking under large curvature changes. • Validated the proposed controller through experiments, demonstrating accurate and robust trajectory tracking under large curvature changes. This paper investigates the trajectory tracking control problem of four-wheel independently driven and steered (4WID-4WIS) mobile robots for operation in dwarf dense orchards. A trajectory tracking method based on Nonlinear Model Predictive Control (NMPC) is proposed for the unpaved and undulating terrain environment. However, the NMPC trajectory tracking controller with fixed parameters cannot achieve satisfactory tracking performance for trajectories with significant curvature differences. By analyzing the relationship between trajectory curvature and controller parameters, a curvature-adaptive NMPC controller is further designed to enhance the robot’s tracking performance under curvature-varying conditions. Field experimental results demonstrate that the adaptive NMPC trajectory tracking controller proposed in this paper can achieve high-precision tracking control of complex trajectories on typical unpaved surfaces including flat and undulating grassland, which effectively improves the execution success rate and operational efficiency of orchard operating robots.
Bi et al. (Wed,) studied this question.