Towards Context-Aware Navigation for Long-Term Autonomy in Agricultural Environments

Autonomous surveying systems for agricultural applications are becoming increasingly important. Currently, most systems are remote-controlled or relying on a single global map representation. Over the last years, several use-case-specific representations for path and action planning in different contexts have been proposed. However, solely relying on fixed representations and action schemes limits the flexibility of autonomous systems. Especially in agriculture, the surroundings in which autonomous systems are deployed, may change rapidly during vegetation periods, and the complexity of the environment may vary depending on farm size and season. In this paper, we propose a context-aware system implemented in ROS that allows to change the representation, planning strategy and execution logics based on a spatially grounded semantic context. Our vision is to build up an autonomous system called Autonomous Robotic Experimental Platform (AROX) that is able to generate crop maps over a whole vegetation period without any user interference. To this end, we built up the hardware infrastructure for storing and charging the robot as well as the needed software to realize context-awareness using available ROS packages.