Abstract-In a Wireless Sensor Network (WSN), intrusion detection is of significant importance in many applications in detecting malicious or unexpected intruder(s). The intruder can be an enemy in a battlefield, or a malicious moving object in the area of interest. With uniform sensor deployment, the detection probability is the same for any point in a WSN. However, some applications may require different degrees of detection probability at different locations. For example, an intrusion detection application may need improved detection probability around important entities. Gaussian-distributed WSNs can provide differentiated detection capabilities at different locations but related work is limited. This paper analyzes the problem of intrusion detection in a Gaussian-distributed WSN by characterizing the detection probability with respect to the application requirements and the network parameters under both singlesensing detection and multiple-sensing detection scenarios. Effects of different network parameters on the detection probability are examined in detail. Furthermore, performance of Gaussian-distributed WSNs is compared with uniformly distributed WSNs. This work allows us to analytically formulate detection probability in a random WSN and provides guidelines in selecting an appropriate deployment strategy and determining critical network parameters.

SUMMARY Collisions and interferences among nodes pose a challenge for data aggregation in many applications, such as target tracking by adopting dynamic convey tree‐based collaboration (DCTC). Because coordination with a time division multiple access (TDMA) medium access control (MAC) might provide an opportunity for better interference control, in this paper, we refine slot allocation to nodes in a dynamic convey tree and design an energy efficient MAC protocol called dynamic‐time division multiple access (D‐TDMA). The D‐TDMA protocol avoids collisions and interferences and allocates contiguous active slots to nodes as far as possible during data aggregation from leaf nodes to a root node. As a result, energy consumption in switching from sleep to active state is saved. In comparison with Always‐On scheme, theoretical analysis results show that the proposed protocol D‐TDMA improves energy efficiency by up to 28.3% during one data aggregation. Furthermore, simulation results show that D‐TDMA does not suffer from collisions and interferences among nodes in a dynamic convey tree and performs similar throughput to that of Always‐On scheme. Because of its advantage of parallel and continuous scheduling among node pairs in the convey tree, D‐TDMA outperforms efficient slot reservation in both energy efficiency and low delay because of to slot saving. Copyright © 2012 John Wiley & Sons, Ltd.

In this paper, machine learning is introduced to source localization in underwater ocean waveguides. Source localization is regarded as a supervised learning regression problem and is solved by generalized regression neural network (GRNN). As a feed-forward network, GRNN is built using training data with fixed structure and configuration. The normalized sample covariance matrix (SCM) formed over a number of snapshots, and the corresponding source position are used as the input and output for GRNN. The source position can be estimated directly from the normalized SCM with GRNN; the proposed approach is thus in theory data driven. In addition, there is only one parameter, the spread factor, to be learned for GRNN. The optimal spread factor is determined using cross-validation. The regression method of GRNN is compared with the classification method of feed-forward neural network (FNN), as well as the classical method of matched field processing (MFP) for vertical array data from the SWellEx-96 experiment. The results show that GRNN achieves a satisfactory localization performance that outperforms both FNN and MFP. The proposed approach provides an alternative way for underwater source localization, especially in the absence of a priori environmental information or an appropriate propagation model.

An automatic ice-fabric analyzer has been developed, which can determine individual ,-axis orientations by image-analysis techniques. The analyzer consists of four major components: a sample stage, a pair of crossed polaroids, a charge coupled device (CCD) camera and a light source. Both the sample stage and the crossed polaroids can be rotated independently of each other by the stepping motors controlled by a personal computer (PC). Measurements are conducted as follows. An ice thin section is set on the sample stage and then the crossed polaroids are rotated. Thin-section images are recorded by the PC at intervals of 2° of rotation. From the image-intensity (gray value) dataset of each crystal in the thin section the extinction angles of individual crystals can be determined. Similarly, eight other extinction angles of individual crystals are obtained from eight other CCD camera positions with respect to the thin section: Finally, the .-axis orientation of individual crystals is calculated by using these extinction angles. With this technique, all crystals within the view of the CCD camera can be analyzed at the same time. In addition, with image-processing techniques the individual crystals are recognized automatically and other parameters, such as grain-size and grain shape, can be measured simultaneously. Textural studies of Dome Fuji (Antarctica) ice cores have been conducted with this analyzer.

Mobile ad-hoc networks (MANET) is a network mode that does not depend on network infrastructure and central access. The fast and flexible networking mode of MANET renders its wide applications in specific scenarios. However, rapidly changing topology and open channels bring potential security problems. In this paper, we proposed an active-routing authentication scheme (AAS) based on the characteristics of active routing protocols. We formally demonstrated that the AAS is effective against selective forwarding attack, false routing attack, byzantine attack and route spoofing attack using the BAN logic considering the possibility of malicious nodes mingling in MANET. Experimental results show that the AAS is compatible with multiple active routing protocols and it is able to increase the packet delivery rate by 33.9%, with an average increase of 18.4% in the network containing some malicious nodes. Furthermore, the AAS is robust which remains the average network connection rate reach 1.6 times of the collusion attack prevention-OLSR(Cap-OLSR) protocol and preserves 79.2% of the network performance in simulation experiments with attacks from malicious nodes. INDEX TERMS mobile ad-hoc network; active routing; authentication scheme; secure routing;