Skip to main content
eScholarship
Open Access Publications from the University of California

Genetic Algorithm-Based Embedded Networked Sensing Design Coupled to an Environmental Simulator

Abstract

Embedded networked sensing (ENS) technology is rapidly expanding into environmental application domains, where network coverage issues are tightly coupled to the environmental media and observational objectives. The goal of this work is to develop and test an automated, real-time ENS coverage design algorithm in the context of an environmental simulation model. The algorithm combines the application of a genetic algorithm (GA) with a deterministic inverse modeling approach, and is demonstrated in the context of a bench-scale groundwater test bed in which the ENS objective is to identify the location of a heat source. More specifically, optimal sensor locations are determined in real-time using a GA-based evolution algorithm whose objective function is the trace minimization of the model-prediction covariance with respect to potential sensor locations. Next, measured temperature sensor data and a descent-based inverse technique are used to update the source location estimate. The procedure is repeated (2 monitoring sensors per design cycle) until a pre-determined sensor supply is exhausted. Two transient heat transport experiments are undertaken in which sources placed upstream of a manually configurable ENS comprising thermocouples for mapping spatiotemporal temperature distributions. The ENS approach successfully corrected an erroneous initial source location estimate and incrementally improved upon this estimate with the addition of new sensors. A point of diminishing improvement was eventually achieved at an imperfect source location estimate. This result was most likely the result of discrepancies between the mathematical model and the experimental system. For the dual-source experiment, the real-time source locator converged on a single source between the two sources, indicating the need for more sophisticated logic for increasingly complex cases.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View