Abstract—
This letter presents an automated roadside mowing
agent. A proof-of-concept testbed was developed based on a lowcost sensor array selected for quick, efficient, and effective roadside
mowing. Dead reckoning from shaft-mounted encoders is fused via
an extended Kalman filter with absolute positioning and heading
information provided by a dual-antenna differential (wide area
augmentation system-corrected) global positioning system (GPS)
unit, to obtain accurate localization. The dual GPS unit, which
includes an integrated gyro, provides accurate drift-free heading,
yielding a low-cost high-performance solution. Outdoor tests on
grassy and uneven terrain illustrate the success and practicality
of the proposed agent. Mowing performance was evaluated based
on consistency of consecutive mowing paths: 10.0 cm consistency
is attainable. Slow-moving drift inherent in GPS positioning does
not substantially affect automated mowing performance, which
depends more on positioning precision than accuracy.
Index Terms—Global positioning system (GPS), Kalman filtering, mobile robots, road transportation.