CON 0: CENS Contaminant Transport Observation and Management (Contam) Research Overview

Center for Embedded Networked Sensing CENS Contaminant Transport Observation and Management Research: Overview Thomas C. Harmon 3 , Deborah Estrin 2 , Jennifer Jay 1 , William Kaiser 5 , Steve Margulis 1 , Miodrag Potkonjak 2 , Jose Saez 4, J. Eric Haux 3 , Jason Fisher 3 , Juyoul Kim 1 , Mohammad Rahimi 2 Naim Busek 2 , John Ewart 3 , Yeonjeong Park 1 , Nithya Ramanathan 2 , Tom Schoellhammer 2 1 UCLA Civil & Environmental Engineering, 2 UCLA Computer Science, 3 UC Merced School of Engineering, 4 Loyola Marymount University Civil & Environmental Engineering, 5 UCLA Electrical Engineering Introduction: Multiscale observation of distributed contaminants in environmental systems system s Integrating multiscale observations using remote and embedded networked sensing The embedded networked sensing (ENS) problem being addressed by the contaminant transport application domain projects is the design of sensor networks supporting physical and chemical observations within and between environmental media, including land, air and water. An interdisciplinary approach is needed because it is difficult to conceive of an effective spatiotemporal sampling plan without domain- specific knowledge and network programming tools are not yet user-friendly enough to see widespread use amongst application domain experts. The contaminant transport observation projects are evolving from controlled test beds to specific real-world deployments to a proposed large, multiscale water quality observation and management network. The current emphasis is on the soil domain which is by nature a rich context for ENS development because of the natural heterogeneity of such media and the inherent cost and technical challenges of deploying sensor networks in these environmental media. Problem Description: Finding optimal network configurations to safely manage contaminants contamin ants 1. Contaminant Source Assessment Use laboratory test beds with simplistic sensing demands (temperature) in order to develop spatiotemporal design strategies 4. Future Directions To maximize the impact of the contaminant observation systems being developed here, we will expand our approach to: (1) New modality in soil, aquatic sediment, riparian and forest modular ENS systems (2) Multiscale environmental observatory designs fusing of ENS- level data streams with larger scale embedded and remote (aerial/satellite) sensing 2. Soil Pylon Sensor Array Design and Validation Create a robust, modular soil sensor array design; use sensing to automatically hone simulators; then us feedback-control to balance irrigation and fertilization with groundwater protection 3. Multiscale Soil Sensor Network Deployment Attack test cases at multiple scales to challenge the hardware- software designs: field scale agricultural and single plant rhizosphere test beds Proposed Solution: Integrating network design with process simulation and management managemen t Contaminant Source Assessment in a laboratory test bed SCXI chassis Heat source injection Peristaltic pump Flow in Sensor network Sandy porous medium Constant temperature water Potential source region 3 point heat 0 source Temperature (C) Time (min) Tailoring the ENS design to multiple scales Drawing by Bill Swenson (UCR) DAQ card Laptop Computer Flow out This completed project proved the concept of real-time ENS design using a combined genetic algorithm-inverse process modeling strategy Initial guessing of source location, the number of active monitoring points, and time interval for monitoring network update Run GA Determine sampling points 0 m depth 1 m depth Constant head downstream sampling region Update estimates and increase active monitoring points Deploy (or reposition) sensors and collect temperature data Inverse modeling 2 m depth 3 m depth Estimate source location Soil Pylon Design and Validation • Local rain gauge sensors monitor spatially distributed irrigation rate • Soil moisture sensors monitor local water content to support the observations of water infiltration and redistribution • Thermistors monitorlocal a ir temperature and below-ground gradient to support energy balance in evapotranspiration calculations • Off-the-shelf vs. CENS-fabricated nitrate sensors (in situ test bed system for CENS Sensor Group) Geopatial statistically based deployment of soil pylon-based observation networks at an agricultural scale (30 acres) and at the single plant scale. Long term vision: multiscale observatories Feedback Control for Optimal Irrigation UCLA – UCR – Caltech – USC – CSU – JPL – UC Merced