Human-Robot Collaboration System Setup for Weed Harvesting Scenarios in Aquatic Lakes

Abstract

Artificial Water Bodies (AWBs) are human-made and require continuous monitoring due to their artificial biological processes. These systems necessitate regular maintenance to manage their ecosystems effectively. Unmanned Surface Vehicle (USV) offers a collaborative approach for monitoring these environments, working alongside human operators such as boat skippers to identify specific locations. This paper discusses a weed harvesting scenario, demonstrating how human-robot collaboration can be achieved, supported by preliminary results. The USV mainly utilises multibeam SOund NAvigation and Ranging (SONAR) for underwater weed monitoring, showing promising outcomes in these scenarios.

Figures (5)

• Figure 1: Weed Harvester boat and USV collaborating in Maschsee Lake.
• Figure 2: Georeferenced map of Maschsee Lake with RGB image captured by the USV's camera overlaid. The SONAR data, visualising detected underwater features, is displayed along the predefined path of the USV. The map shows the surveyed areas and the nature of the detected features - whether underwater vegetation, the seabed, or other objects - requiring further analysis
• Figure 3: 3D view of SONAR point cloud data from an inspection area, showing the height difference in weed distribution detected by multibeam SONAR [Top] before mowing and [Bottom] after mowing.
• Figure 4: Underwater image showing weed at Maschsee Lake captured using GoPro camera.
• Figure 5: Submerged Ladder [a] Camera image above water [b] Multibeam SONAR 3D point cloud data [c] SONAR Backscatter data.