In focused coverage problem, sensors are required to be deployed around a given point of interest (POI) with respect to a priority requirement: an area close to POI has higher priority to be covered than a distant one. A localized sensor self-deployment algorithm, named Greedy-Rotation-Greedy (GRG) [10], has recently been proposed for constructing optimal focused coverage. This previous work assumed obstacle-free environment and focused on theoretical aspects. Here in this paper, we remove this strong assumption, and extend GRG to practical settings. We equip GRG with a novel obstacle penetration technique and give it the important obstacle avoidance capability. The new version of GRG is referred to as GRG/OP. Through simulation, we evaluate its performance in comparison with plain GRG.
We propose a localized sensor localization scheme making full use of controlled mobility of a location-aware actor and the connectivity of the sensor network. It contains two new algorithms: a unscented particle filter (UPF) based localization algorithm and an actor mobility scheduling algorithm. The former is an application of UPF. It enables sensor selflocalization using received signal strength indicator and actor position. The latter models actor mobility scheduling as traveling salesman problem and aims at fully localized network and minimized time delay. Navigated by sensors, the actor depth-first traverses a local minimum spanning tree of a connected 3-dominating set of the network.Consider a sensor S and an actor R, which have same orientation. Denote by CS the coordinate system of S and by CR that of R. CS has origin at S; CR has origin at a position known to R. Let the coordinates of R be [r, s] The objective of localization is to determine [a, b] T . From local perspective, it is equivalent to target tracking. R is aware of [r ′ , s ′ ] T while moving. When in direct contact, R informs S of [r ′ , s ′ ] T , and S tracks R's movement in CS. As soon as the result by Eqn. 1 stabilizes, S becomes localized. Below we show how S uses a unscented particle filter (UPF) [4] solution to track R based on received signal strength indicator (RSSI). UPF is different from conventional PF in that it uses as important density unscented Kalman filter (UKF) Gaussian approximation rather than transition priori. The use of UKF makes it possible to incorporate latest observation to propose new values for system state, leading to improved accuracy. The background of UPF is omitted. Dynamic state space modelWhen actor R has direct contact with sensor S, a dynamic system is formed. The system state is composed of the position of R in CS and its horizontal velocities and ver-ACM 978-1-60558-531-4/09/05. tical velocity, i.e., x k = [r k , s k ,ṙ k ,ṡ k ] T . Let ∆t be the sampling time, i.e., time interval between two consecutive states.The system dynamics is modeled as: x k = Ax k−1 + Bv k−1 , where v k−1 is a zero-mean Gaussian with covariance matrix Q k−1 = δ k−1 I2 and models unknown acceleration of R, andR transmits a beacon message containing its current position [r ′ , s ′ ] T in CR every ∆t time units. S observes its distance to R using the RSSI of communication link in between. We adopt the log-distance signal propagation model [5]. The average power at distance d from transmitter iswhere α is a slope describing signal strength change rate, and β is a constant determined by the transmitted power, wavelength, and antenna gain of the actor. The received signal strength is given by ρ(d) = µ(d) + n, where n is a Gaussian white noise N (0, δ(d)) measuring uncertainty due to factors such as shadow fading and interference. The distance d k at time k between S and R is (r 2 k + s 2 k ) 1 2 . The observation equation is z k = µ(d k ) + n k = β − 5α log(r 2 k + s 2 k ) + n k . Likelihood function p(z k |x k )Since the noise term...
No abstract
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.