Inter-row Robot Navigation using 1D Ranging Sensors

In this paper a fuzzy logic navigation controller for an inter-row agricultural robot is developed and evaluated in laboratory settings. The controller receives input from one-dimensional (1D) ranging sensors on the robotic platform, and operated on ten fuzzy rules for basic row-following behavior. The control system was implemented on basic hardware for proof of concept and operated on a commonly available microcontroller development platform and open source software libraries. The robot platform used for experimentation was a small tracked vehicle with differential steering control. Fuzzy inferencing and defuzzification, step response and cross track error were obtained from the test conducted to characterize the transient and steady state response of the controller. Controller settling times were within 4 seconds. Steady state centering errors for smooth barrier navigation were found to be within 3.5% of center for 61 cm wide solid barrier tests, and within 38% for simulated 61 cm corn row tests.