Robot-assisted measurement in data-sparse regions

This work investigated the use of low-cost robots, small unmanned aerial vehicles (UAVs) and small unmanned surface vehicles (USVs), to assist researchers in environmental data collection in the Arkavathy River Basin in Karnataka, India. In the late 20th century, river flows in the Arkavathy began to decline severely, and Bangalore’s dependence on the basin for local water supply shifted while the causes of drying remain unknown. Due to the lack of available data for the region, it is difficult for water management agencies to address the issue of declining surface flows; by collecting critical hydrologic data accurately and efficiently through the use of robots, where data is not available or accessible, local water resources can more easily be managed for the greater Bangalore region. Three case study sites, including two irrigation tanks and one urban lake, within the Arkavathy basin were selected where unmanned aerial vehicles and unmanned surface vehicles collected data in the form of aerial imagery and bathymetric measurements. The data were further processed into 3D textured surface models and exported as digital elevations models (DEMs) for post-processing in GIS. From the DEMs, topographic and bathymetric maps were created and storage volumes and surface areas are calculated by relating water surface levels to tank bathymetry. The results are stage-storage and stage-surface area relationships for each case study site. These relationships provide valuable information relating to groundwater recharge and streamflow generation. Sensitivity analysis showed that the topographic surface data used in the stage-storage and stage-surface area curves was validated within ± 0.35 meters. By providing these relationships and curves, researchers can further understand hydrologic processes in the Arkavathy River Basin and inform local water management policies. From these case studies, three formative observations were made, relating to i) interpretation of the data fusion process using information collected from both UAV and USV systems; ii) observations for the human-robot interactions for USV and; iii) field observations for deployment and retrieval in water environments with low accessibility. This work is of interest to hydrologists and geoscientists who can use this methodology to assist in data collection and enhance their understanding of environmental processes