Physical Modeling and Software Synthesis for Self-Reconfigurable Sensors in River Environments
Abstract
Our work examines the role of software synthesis for monitoring and planning of autonomous sensors evolving on tidally forced rivers. The goal of the sensors is the coordinated sampling of currents and salinity to reconstruct the distributed state of the river and detect salinity intrusion. Solutions to problems such as this require an approach that considers reliable sensing, fast computation and safety verification. The recent results include Voronoi partitioning of regions of the medium based on energy models [1], and algorithms for covering those regions that utilize various flow-field approximations [2].
We also have recent results from a demonstration of 100 floating self-contained sensors for river environments. This demonstration was conducted in the Sacramento River near Walnut Grove, CA on May 9, 2012. The mixed fleet of active (self-propelled) and passive (purely floating) sensors was deployed from a dock and retrieved by boat after four hours. The active sensors used their propulsion to avoid obstacles and navigate the junction between the Sacramento River and the Georgiana Slough. The propulsion feature allows the floating sensors to stay safe without human supervision as well as to distribute themselves across the network of channels in order to fulfill sensing objectives. Our future work on this front will be dedicated to the development of new metrics and strategies that can help us deal with the uncertainty in the river state (which we would like to reconstruct) and (moderately) dynamic scenarios [3].
Results in model-based software synthesis allow new techniques for synthesizing the drifter controllers from high-level models. The modeling languages are used to determine failure modes for a drifter (e.g., flipping over, remaining at a single GPS position for too long) and what course of action should be taken. This work has also led to success in utilizing mobile phone programming as a motivating course for undergraduate and graduate students, to understand how CPS problems can be attacked using this ubiquitous device [4].
1. Y. Ru and S. Martinez, “Energy-based Voronoi partition in constant flow environments”, submitted to IEEE Transactions on Automation Science and Engineering, 2011. http://arxiv.org/abs/1111.0071
!2. Y. Ru and S. Martinez, “Coverage control in constant flow environments based on a mixed energy-time metric”, to appear in IEEE Conf. on Decision and Control, 2012.
!3. A. Tinka, M. Rafiee and A.M. Bayen. “Floating sensor networks for river studies” IEEE Systems Journal. To appear, available at http://dx.doi.org/10.1109/JSYST.2012.2204914 !4. J. Sprinkle. “Teaching students to learn to learn mobile phone programming.” In Proceedings of the compilation of the co-located workshops on DSM'11, TMC'11, AGERE!'11, AOOPES'11, NEAT'11, & VMIL'11, pages 261-266, 2011.!
Award ID: 0930919