Selection of a Navigation Strategy According to Agricultural Scenarios and Sensor Data Integrity

In digital farming, the use of technology to increase agricultural production through automated tasks has recently integrated development AgBots for more reliable data collection using autonomous navigation. These are equipped with various sensors such as GNSS, cameras, and LiDAR, but these can be prone limitations low accuracy under-canopy navigation sensitivity outdoor lighting platform vibration LiDAR occlusion issues. order address ensure robust navigation, this paper presents a sensor selection methodology based on identification environmental conditions data. Through extraction features from images, point clouds, we able determine feasibility each create vector indicating its viability. Our results demonstrate that proposed effectively selects between cameras or within crops GNSS outside crops, at least 87% time. The main problem found is that, in transition inside crop, take 20 s adapt. We compare variety classification algorithms terms performance computational cost show our method higher lower cost. Overall, allows low-cost most suitable given environment.