Reference trajectory planning under constraints and path tracking using linear time-varying model predictive control for agricultural machines

• Published in: Biosystems engineering Vol. 153; pp. 28 - 41
• Main Authors: Graf Plessen, Mogens M., Bemporad, Alberto
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2017
• Subjects: Actuators; Agricultural equipment; agricultural machinery and equipment; Autonomous navigation; Design engineering; geometry; Linear time-varying model predictive control; Mathematical models; Optimization; planning; Predictive control; Reference trajectory planning; spraying; Tracking; Trajectories; Reference trajectory planning; Autonomous navigation; Linear time-varying model predictive control
• ISSN: 1537-5110; 1537-5129
• DOI: 10.1016/j.biosystemseng.2016.10.019

A method for the control of autonomously and slowly moving agricultural machinery is presented. Special emphasis is on offline reference trajectory generation tailored for high-precision closed-loop tracking within agricultural fields using linear time-varying model predictive control. When optimisation is carried out, high-level logistical processing can result in edgy reference paths for field coverage. Subsequent trajectory smoothing can consider specific actuator rate constraints and field geometry. The latter step is the subject of this paper. Focussing on forward motion only, the role of non-convexly shaped field geometry, repressed area minimisation and spraying gap avoidance is analysed. Three design methods for generating smooth reference trajectories are discussed: circle-segments, generalised elementary paths, and bi-elementary paths. •Three design tools for smooth path generation: circle segments, generalized elementary and bi-elementary paths.•Trade-off between repressed area minimization and spraying gap avoidance.•Achievable closed-loop tracking accuracies under nominal conditions (noise-free and full state-feedback).