The application of site-specific herbicides requires rapid and stable pressure regulation under intermittent nozzle switching and time-varying hydraulic demand. This article presents an embedded hierarchical multirate pressure-control strategy for a robotic smart weed sprayer. The proposed architecture combines a fast fuzzy generalized predictive control (FGPC) loop acting on four diaphragm pumps with a slower control proportional integral (PI)-based loop acting on a proportional valve. The pump loop provides the main pressure-tracking action, while the valve regulation assists pressure decay during downward transients and shapes the operating point to reduce excessive pump effort. The strategy was experimentally validated on an embedded robotic platform under pressurized recirculation and repeated nozzle-switching disturbances. In a mixed pressure reference-tracking test, the controller achieved a mean absolute error of 1.10 psi and a root-mean-square error of 1.63 psi. Under repeated nozzle switching at 10 psi, it maintained bounded regulation with a mean absolute error of 2.08 psi, a root-mean-square error of 3.01 psi, and worst-case recovery to the ± 1 psi band in 1.26 s. The results demonstrate that the proposed multirate architecture provides effective pressure regulation, bounded actuation, and controlled electrical demand for the embedded hydraulic subsystem of a robotic smart weed sprayer. These findings support the integration of the proposed controller into full site-specific robotic spraying workflows.
Schutz et al. (Mon,) studied this question.