Multi-Sensor-Based Estimation Algorithms For Wireless Networked Systems

In recent years, with the development of wireless communication technology and the enhancement of sensor node performance, wireless sensor networks (WSNs) have been involved in a wide range of applications. As a hub for information transmitting from sensor to data processing center or from sensor to sensor, wireless communication network is introduced into multi-sensor-based estimation systems, which may be termed as wireless networked multi-sensor-based estimation systems (WNMESs). Nowadays, WNMESs have been one of the hot research topics.The utilization of WSNs in multi-sensor-based estimation systems offers lots of ad-vantages, such as easy deployment and maintenance, low cost, high reliability, good fault-tolerance and scalability. However, it also brings in many new problems, for example, en-ergy limitation, bandwidth constraint and sensor data asynchronization. Based on the set-membership estimation theory, projection theory, Lyapunov theory, least-squares estima-tion method and covariance intersection fusion algorithm, this paper is concerned with the multi-sensor-based estimation problems with quantized communication, sensor scheduling, clustering sensor data and asynchronous transmission. The main work and contributions are summarized as follows:1. A distributed set-valued estimation problem with local quantized communication is investigated for a class of WNMESs subjected to power and bandwidth constraints. A static logarithmic quantizer is used to make the manipulation of quantization on the infor-mation to be transmitted. By rewriting the logarithmic quantization error as an uncertainty, a distributed set-valued estimation strategy is proposed based on the sensors'own measure-ments and the received local information. A sufficient condition for the existence of the distributed set-valued estimators is derived, and the estimator parameters are determined by solving a convex optimization problem. Two examples of tracking manoeuvring targets described by a two-dimensional model and a three-dimensional model, respectively, are provided to demonstrate the effectiveness of the proposed theoretical results.2. A distributed consensus-based Kalman estimation problem is investigated for sen-sor networks with quantized communications and random sensor failures. A probabilistic quantization strategy is adopted for local encoding communication. A recursive equation is presented to calculate the estimation error covariance matrix and an upper bound is derived for the estimation performance index. Moreover, a sufficient condition for the convergence of the upper bound of the estimation performance index is also presented. According to the relationship between the estimation performance and the quantization parameter, two types of optimization problems are constructed for cases of energy is constrained or not, respec-tively. Illustrative examples are provided to demonstrate the effectiveness of the proposed theoretical results.3. A stochastic competitive transmission strategy is proposed to deal with the trans-mission constraint problem in a single-channel-based sensor network. A periodic mixed storage strategy combing the zero-input and the hold-input mechanisms is presented. By introducing a random variable to describe the transmission situation in each sensor, the WNMESs are rewritten as a class of periodic discrete-time stochastic systems. By using the orthogonal projection theory, a multi-sensor-based Kalman estimation algorithm with SCT and PMS is derived. Lastly, a continuous stirred tank reactor and a moving target are provided to demonstrate the effectiveness of the proposed theoretical results.4. A multi-sensor-based H? estimation problem is investigated for heterogeneous sen-sor networks with a common communication channel. The SCT strategy is used for sensor scheduling, such that the sensors communicate with the estimator in a strict asynchronous manner. By using the Lyapunov theory, a multi-sensor-based H? filter is designed, where the filter parameters are determined by solving a linear matrix inequality. Based on the asynchronous sampled information from the sensors, the estimator aims to periodically generate estimates. Lastly, an illustrative example is provided to demonstrate the effective-ness of the proposed theoretical results.5. A multi-sensor-based aperiodic least-squares estimation problem is investigated for heterogeneous sensor networks with a common communication channel. The SCT strategy and the PMS scheme are used for sensor scheduling and measurements storage, respec-tively. A set of Bernoulli random variables are introduced to describe the sensor failure phenomenon. By introducing a cost function regarding as the estimation performance in-dex of WNMESs, the aperiodic estimation problem is translated into solving a least-squares optimization problem. By using the least-squares approach, a multi-sensor-based aperiodic estimation algorithm is given. Moreover, performance analysis is presented for the estima-tion system and an upper bound is derived for the estimation error. Additionally, a sufficient condition is presented to ensure the convergence of the obtained upper bound. Lastly, an ar-tificial weir system is provided to demonstrate the effectiveness of the proposed theoretical results.6. A hierarchical fusion estimation problem is investigated for clustered sensor net-works with clustering sensor data and asynchronous local estimates. According to the hi-erarchical estimation strategy, local estimates are firstly generated based on the clustered measurements and then be fused by using the centralized fusion strategy. A minimum vari-ance estimation algorithm is presented for each cluster, and a covariance intersection fusion strategy is presented for the fusion center by using asynchronous local estimates. Moreover, the accuracy analysis of the fusion estimation algorithm is also given. Lastly, an illustrative example is provided to demonstrate the effectiveness of the proposed theoretical results.