Authors: Minglei You; Ting Jiang
Addresses: Key Laboratory of Universal Wireless Communication, Ministry of Education, Beijing University of Posts and Telecommunications, P.O. Box 96, No. 10, XiTuCheng Road, Beijing 100876, China ' Key Laboratory of Universal Wireless Communication, Ministry of Education, Beijing University of Posts and Telecommunications, P.O. Box 96, No. 10, XiTuCheng Road, Beijing 100876, China
Abstract: Obstacle identification is a difficult task, which is more challenged in foliage environment. In this paper, a method of target detection is proposed, which attempts to extract information from the data being transmitted around the wireless sensor network (WSN) to identify targets that might be within the local, foliage obscured scene. The selected bispectra algorithm is applied to extract the feature vector, as well as radial-basis function (RBF) neural network is used to realise the obstacle classification. According to the experiment results, this method is able to identify the existence and the different distances of the obstacles measured in outdoor foliage scene with a good recognition rate.
Keywords: target detection; obstacle identification; UWB; ultra wideband; WSNs; wireless sensor networks; selected bispectra; RBF neural networks; radial-basis function; foliage environment; vegetation effect; trees.
International Journal of Sensor Networks, 2015 Vol.17 No.1, pp.63 - 72
Received: 19 Jan 2013
Accepted: 26 Dec 2013
Published online: 19 Feb 2015 *