Coordination mechanism based on mobile actuator design for wireless sensor and actuator networks

Abstract

Wireless sensor and actuator networks combine a large number of sensors and a lower number of actuators that are connected with wireless medium, providing distributed sensing and executing appropriate tasks in a special region of interest. To accomplish effective sensing and acting tasks, efficient coordination mechanism among the nodes is required. As an attempt in this direction, this paper develops a collaborative control and estimation mechanism, which addresses the nodes coordination in a distributed manner. First, we propose a regional controllability-based virtual force algorithm as an actuator deployment strategy to enhance area coverage after an initial random placement of actuators. During this process, a dynamic coordination mechanism is adopted to control nodes. This mechanism incorporates two components, namely, proportional-integral-derivative neural network and recursive least squares-based Kalman filter algorithms. Taking advantage of feedback control and online learning technology, the proposed coordination mechanism schedules the corresponding nodes on the basis of the characteristics of current events, utilizes proportional-integral-derivative neural network controller inside each actuator to improve system transient and steady-state responses, and deals with system state/parameter estimation problems according to the recursive least squares-based Kalman filter algorithm, so as to achieve better control accuracy. Simulations demonstrate the effectiveness of our proposed methods.

Distributed Node Coordination for Real-Time Energy-Constrained Control in Wireless Sensor and Actuator Networks

“…A fault tolerant algorithm is provided in [29] to jointly optimize the scheduling, routing and control. A neural network based control algorithm is designed in [30] to achieve a precise light control and maintain a robustness against the inaccurate system parameters.…”

“…During the A-A coordination, we compare the proposed Lagrange-based Distributed Algorithm (LDA) with three ex- isting methods: 1) PID Neural Network (PIDNN) [13], [30], 2) Gradient Descent method (GD) [7], [31], and 3) Sequential Unconstrained Minimization Technique (SUMT) [1]. The results are shown in Fig.…”

E-eco: Performance-aware energy-efficient cloud data center orchestration

“…He et al (2012) considers the quality of sensing as a utility function, proposing a greedy algorithm to perform placement and scheduling through the activation or deactivation of the sensors. In the same direction, Mo and Xu (2015) uses a Kalman filter algorithm for scheduling sensors, based on current events, in order to develop a deployment strategy that enhances the coverage of the sensors. E-eco also manages the state of the participating hosts, but these hosts have considerable delay time between transitions.…”

Learning‐based adaptive feedback control for tracking optimisation in wireless sensor actuator networking systems

“…The authors of ref. [30] proposed an actuator deployment strategy to enhance area coverage after an initial random placement of actuators. During this process, a dynamic coordination mechanism was adopted to control nodes, which incorporates proportional–integral–derivative neural network and recursive least squares‐based Kalman filter algorithms.…”