Semisupervised Location Awareness in Wireless Sensor Networks Using Laplacian Support Vector Regression

Abstract

Supervised machine learning has been widely used in context-aware wireless sensor networks (WSNs) to discover context descriptions from sensor data. However, collecting a lot of labeled training data in order to guarantee good performance requires much cost and time. For this reason, the semisupervised learning has been recently developed due to its superior performance despite using only a small amount of the labeled data. In this paper, we extend the standard support vector regression (SVR) to the semisupervised SVR by employing manifold regularization, which we call Laplacian SVR (LapSVR). The LapSVR is compared with the standard SVR and the semisupervised least square algorithm that is another recently developed semisupervised regression algorithm. The algorithms are evaluated for location awareness of multiple mobile robots in a WSN. The experimental results show that the proposed algorithm yields more accurate location estimates than the other algorithms.

Distributed estimation using online semi‐supervised particle filter for mobile sensor networks

“…The proposed particle filter incorporates the learning algorithm into the likelihood in (6). First, the likelihood is designed as Gaussian distribution given by…”

“…As an extension from the conventional static WSNs, mobile sensor networks (MSNs) can be used for a wider range of applications by incorporating some mobile nodes. Depending on the role of mobile nodes, MSNs are utilised for navigation [2], cooperative system [3][4][5], target tracking [6] and surveillance [7,8]. For many MSN applications, limited computation and communication power render distributed estimation more suitable than the centralised version.…”

“…To address these problems, semi-supervised learning algorithms have been recently developed. A main idea of semisupervised learning is to exploit the prior knowledge from the unlabelled data to enhance classification performance [30][31][32] and regression performance [6,17,33]. In this paper, unlabelled data is utilised to build manifold regularisation (or graph-based semi-supervised learning) that captures intrinsic geometric structure of data and eventually makes the estimator smooth along the manifold [34].…”

“…Experimental setup with (a) 1 mobile node and 13 stationary nodes (called beacons)[6] (b) The line of sight between two nodes can be blocked by an appearance of an obstacle, which causes the change in the RSSI pattern. (c) The relationships between RSSI measurement and distance, before (Sample 1) and after (Sample 2) the RSSI characteristic is disturbed by an obstacle a Setup b Obstacle between two nodes c RSSI models…”