GNSS‐stereo‐inertial SLAM for arable farming

The accelerating pace in the automation of agricultural tasks demands highly accurate and robust localization systems for field robots. Simultaneous Localization Mapping (SLAM) methods inevitably accumulate drift on exploratory trajectories primarily rely place revisiting loop closing to keep a bounded global error. Loop closure techniques are significantly challenging fields, as local visual appearance different views is very similar might change easily due weather effects. A suitable alternative practice employ sensor positioning jointly with rest robot sensors. In this paper we propose implement fusion navigation satellite system (GNSS), stereo views, inertial measurements purposes. Specifically, incorporate, tightly coupled manner, GNSS into stereo-inertial ORB-SLAM3 pipeline. We thoroughly evaluate our implementation sequences Rosario data set, recorded by an autonomous soybean own in-house data. Our includes from conventional GNSS, rarely included evaluations state-of-the-art approaches. characterize performance GNSS-stereo-inertial SLAM application case, reporting pose error reductions between 10% 30% compared visual-inertial loosely baselines. addition such analysis, also release code open source.