IN041-14
MULTI-AGENT MULTI-SCALE OBSERVATIONS OF SOIL MOISTURE VIA SPCTOR: SENSING POLICY CONTROLLER AND OPTIMIZER

Wednesday, 16 December 2020: 04:39
Virtual
Mahta Moghaddam, University of Southern California, Ming Hsieh Department of Electrical and Computer Engineering, Los Angeles, CA, United States, Ruzbeh Akbar, University of Southern California, The Ming Hsieh Dept. of Electr. Eng., Los Angeles, CA, United States, Agnelo Silva, METER Group, Pullman, WA, United States, Samuel Prager, University of Southern California, Los Angeles, United States and Dara Entekhabi, Massachusetts Institute of Technology, Department of Civil and Environmental Engineering, Cambridge, MA, United States
Abstract:
In support of the NASA ESTO/AIST New Observing System vision, we are developing a combined ground- and drone-based Wireless Sensor Network (WSN) technology for observations of soil moisture at multiple spatial and temporal scales. This technology is developed based on two objectives: (a) developing a Sensing-Policy Controller (SPC) for multi-Agent – WSN and Unmanned Aerial Vehicle (UAV) – observation strategy coordination and optimization, and (b) demonstrating in-field integrated operations of WSNs and UAV-based software defined radars (SDRadars).

Existing ground networks have limited capabilities in spatiotemporal sampling flexibility and agility, multi-user coordination, and multi-sensor data integration. These limitations are further compounded when multiple users with different application requirements seek access to the network. These application requirements encompass calibration/validation activities, such as those for SMAP, SMOS, and NISAR, and stand-alone watershed-scale observations. The observations scales could range from meters to kilometers (spatially) and days to weeks (temporally).

Heritage WSN technology from the in-situ Soil moisture Sensing Controller and oPtimal estimator (SoilSCAPE) networks will be leveraged to define the architecture of the new multi-agent network. The agents will include in-situ sensors as well as UAV-based SDRadars that provide soil moisture observations on a larger synoptic scale, yet are lower-cost than traditional airborne sensors and can be deployed on-demand and in near-real-time. Within a given observation domain that could span an entire watershed, the in-situ network will be utilized to determine regions of high soil moisture uncertainty using an upscaling method. The UAV-based SDRadar will then be deployed to collect data over such locations to reduce the uncertainty. A network of UAVs will ultimately be used to extend the region of coverage and the accuracy of observations. To achieve optimal observational scenarios, we introduce a new physics-aware decision-making framework, which enables coordination of UAV- and in-situ WSN soil moisture sampling strategies. This presentation will describe each of these project elements and will include the latest status of developing the SPC, the SDRadar, and the path planning algorithm.

<< Previous Abstract | Next Abstract