Autonomic Cyber Physical System for Biosphere 2 Landscape Observatory ResearchGraduate Students: Venkata Krishna Nimmagadda, and Shafiul Islam PIs: Drs. Salim Hariri, Steve DeLong, and Ali AkogluMotivation: Interaction of hydrological cycle with other global changes is affecting ecosystems on Earth. Landscape Evolution Observatory (LEO) is an interdisciplinary project at Biosphere 2 aimed to quantify the interactions between soils, biological, hydrological, geological and atmospheric processes. To develop a good u 
nderstanding of these interactions of processes, there is a pressing need for development of experimental infrastructure which can couple complex physical systems with hydrological models. An ability to control the physical systems based on the information from monitoring systems equipped with a variety of field instrumentation, will considerably accelerate the discovery cycle by iteratively testing models against experimental system. To this date, there is no complete solution to control and manage compu
tational models and experimental test bed in a closed loop form with the autonomic sensing, data‐model fusion, autonomic workflows and online analysis capabilities. Such a solution will accelerate research and discovery for grand Earth science challenges, and as well as many other cyber physical systems such as intelligent transportation systems, greenhouses, etc. ;) Background and Current State: LEO experiments are characterized by large amounts of the data streaming and processing to effectively study hydrological cycles and their interactions with environment. Ring Buffer Network Bus (RBNB) Data Turbine is an open source initiative to develop an infrastructure for data management, processing and synchronization. Kepler is another open so..ware tool for scientific workflow management. In parallel we also set up a physical experimental testbed in the B2. The test bed consists of 30 sensors to monitor features such as the weight, water flow, and conductivity with an operational infrastructure to simulate rain on the test bed. National Instrument’s Compact RIO is programmed with the LabView to collect and store data in real time. However, the experimental test bed and the workflow management components have not been integrated yet. Project Goals: To enable seamless coupling of computational models with the unique experimental infrastructure of Biosphere 2 (B2). We propose a service-oriented cyber- infrastructure shown in figure 1 for establishing a distributed system which integrates physical infrastructure at B2 with data assimilation models executing on a virtualized cloud platform with hundreds of cores available at Autonomic Computing Lab at University of Arizona. Our objective is to develop strategies for robust data acquisition, autonomic decision making in sensing, data-model fusion, analysis and visualization to support accelerate research and discovery.
Research Tasks:- B2.LEO Sensor Network Design: The current implementation of our sensor networks can be accessed online to keep track of real time data generated from the soil test.bed which is available at http://ponderosa.b2science.org:8080/SensorPlots/RTSDPs_exp3.jsp.
- Sensor Data Streaming and Architecture
- Eco-hydrological Modeling
- Model Calibration and Improved Estimates through Assimilation of Data and Model
- Autonomic Control, Analysis, and Visualization Dashboard
| 
