Dynamic Spectrum Access (DSA) is widely seen as a solution to the problem of limited spectrum, because of its ability to adapt the operating frequency of a radio. Mobile Ad Hoc Networks (MANETs) can extend high-capacity mobile communications over large areas where fixed and tethered-mobile systems are not available. In one use case with high potential impact, cognitive radio employs spectrum sensing to facilitate the identification of allocated frequencies not currently accessed by their primary users. Primary users own the rights to radiate at a specific frequency and geographic location, while secondary users opportunistically attempt to radiate at a specific frequency when the primary user is not using it. We populate a spatial radio environment map (REM) database with known information that can be leveraged in an ad hoc network to facilitate fair path use of the DSA-discovered links. Utilization of high-resolution geospatial data layers in RF propagation analysis is directly applicable. Random matrix theory (RMT) is useful in simulating network layer usage in nodes by a Wishart adjacency matrix. We use the Dijkstra algorithm for discovering ad hoc network node connection patterns. We present a method for analysts to dynamically allocate node-node path and link resources using fair division. User allocation of limited resources as a function of time must be dynamic and based on system fairness policies. The context of fair means that first available request for an asset is not envied as long as it is not yet allocated or tasked in order to prevent cycling of the system. This solution may also save money by offering a Pareto efficient repeatable process. We use a water fill queue algorithm to include Shapley value marginal contributions for allocation.

No abstract

No abstract

High-resolution sensors provide essential information to the oil and gas industry when planning for corridor right-of-way, land restoration, damage prevention, risk management and emergency response/ damage prevention. The world's growing need for energy has put pressure on the remote sensing community to create even higher resolution and more accurate sensors than ever before. The location of the area of interest where remote sensing is needed is often based on the location of potential petroleum reserves. Political, terrain and weather conditions often determine the choice of technology used to collect the necessary data. Continuous pipeline monitoring is becoming increasingly critical for the oil and gas industry, which faces new challenges due to higher demands for energy. This paper presents the design, development and testing of a smart, wireless sensor network for early leak detection in oil pipelines, based on sensor measurements. Today's low power sensors are capable of measuring pressure, temperature and flow changes due to leaks. We describe a system model for determining decision-making strategies based upon the ability to perform data mining and pattern discovery by utilizing sensor information to detect leaks or abnormal situations. We discuss the development of a method for determining actionable information using game theory. We focus on applying probabilistic models to leak detection and geospatial tools for assisting emergency response and pipeline repairs. Probabilistic predictions are critical in practice on many decision-making applications. However, applicability to big data is complicated by the difficulties of inference in complex probabilistic models, and by computational constraints.

No abstract