Design and Evaluation of a Levelling System for a Weeding Robot

Lin Chen, Supod Kaewkorn, Long He, Qin Zhang, Manoj Karkee

Research output: Contribution to journalArticle

Abstract

Weed control is a critically important task in organic crop farming. Even though there are machines available for inter-row weeding, manual weeding is still the only choice for weed control in organic farms, especially in the narrow spaces between crop plants (intra-row weeds). Such an operation is highly labor intensive and costly in organic vegetable production. Automatic or robotic weeding could provide a potential solution for addressing labor related issues. In intra-row weed control, weeding end-effectors need to be positioned accurately to remove weeds growing very close to the plant while the robotic vehicle is continuously moving on a generally uneven and uncertain field surface. This study was aimed at assessing the performance of an end-effector auto-levelling system designed to accurately control the position of the end-effector during weeding operations in vegetable crops. The performance assessment was conducted via a set of laboratory experiments using a specifically designed and fabricated proof-of-concept prototype. To achieve the desired level of performance in actual field conditions, the prototype system required maintaining the end-effector base at horizontal position within a ± 0.25º angular error when the testbed (laboratory prototype) roll and pitch angles were varied from -8º to 8º. The test results verified that the developed end-effector base levelling system could maintain the drift of the end-effector tip position within 18 mm when input roll or pitch angle reached 8º. Meanwhile, the corresponding position error caused by angular error of the levelling plate was limited in 0.2 mm when the levelling plate at a height of 10 cm, which means the levelling system can efficiently reduce the effect of the rough field. The regularity of the end-effector tip position drift can also help us with end-effector control.

Original languageEnglish (US)
Pages (from-to)299-304
Number of pages6
JournalIFAC-PapersOnLine
Volume49
Issue number16
DOIs
StatePublished - Jan 1 2016

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering

Cite this