Mobile Sensor Networks and Control: Adaptive Sampling of Spatiotemporal Processes

Author(s):  
Derek A. Paley ◽  
Artur Wolek

The control of mobile sensor networks uses sensor measurements to update a model of an unknown or estimated process, which in turn guides the collection of subsequent measurements—a feedback control framework called adaptive sampling. Applications for adaptive sampling exist in a wide range of settings, especially for unmanned or autonomous vehicles that can be deployed cheaply and in cooperative groups. The dynamics of mobile sensor platforms are often simplified to planar self-propelled particles subject to the ambient flow of the surrounding fluid. Sensor measurements are assimilated into continuous or discrete models of the process of interest, which in general can vary in space and time. The variability of the estimated process is one metric to score future candidate sampling trajectories, along with information- and uncertainty-based metrics. Sampling tasks are allocated to the network using centralized or decentralized optimization, in order to avoid redundant measurements and observational gaps.

Author(s):  
Yunfei Xu ◽  
Jongeun Choi

In this paper, a new class of Gaussian processes is proposed for resource-constrained mobile sensor networks. Such a Gaussian process builds on a GMRF with respect to a proximity graph over a surveillance region. The main advantages of using this class of Gaussian processes over standard Gaussian processes defined by mean and covariance functions are its numerical efficiency and scalability due to its built-in GMRF and its capability of representing a wide range of non-stationary physical processes. The formulas for Bayesian posterior predictive statistics such as prediction mean and variance are derived and a sequential field prediction algorithm is provided for sequentially sampled observations. For a special case using compactly supported kernels, we propose a distributed algorithm to implement field prediction by correctly fusing all observations in Bayesian statistics. Simulation results illustrate the effectiveness of our approach.


Author(s):  
J. Karl Hedrick ◽  
Brandon Basso ◽  
Joshua Love ◽  
Anouck R. Girard ◽  
Andrew T. Klesh

This paper presents a state-of-the-art survey in the broad area of Mobile Sensor Networks (MSNs). There is currently a great deal of interest in having autonomous vehicles carrying sensors and communication devices that can conduct ISR (intelligence, surveillance and reconnaissance) operations. Although this paper will discuss issues common to mobile sensor networks, the applications will generally be associated with autonomous vehicles. Areas that are addressed are: 1. Mission definition languages that allow the human to compose a mission defined in terms of tasks; 2. Communication issues including hardware, software, and network connectivity; 3. Task allocation among the assets generally by a market-based approach; 4. Path planning for individual agents; and 5. Platform motion control using autopilots with and without GPS signals and including collision avoidance.


Author(s):  
J. Karl Hedrick ◽  
Brandon Basso ◽  
Joshua Love ◽  
Benjamin M. Lavis

This paper compares some of the common tools and techniques that enable state-of-the-art systems to provide high-level control of mobile sensor networks. There is currently a great deal of interest in employing unmanned and autonomous vehicles in intelligence, surveillance, and reconnaissance operations. Although this paper addresses issues common to all mobile sensor networks, the applications presented are typically associated with autonomous vehicles. We focus specifically on three high-level areas: 1. mission definition languages that allow human users to compose missions defined in terms of tasks, 2. communication-addressing degradation and loss and relationship to underlying system architecture design, and 3. task allocation among the assets.


Sign in / Sign up

Export Citation Format

Share Document