Proc. SPIE. 6217, Detection and Remediation Technologies for Mines and Minelike Targets XI
KEYWORDS: Target detection, Detection and tracking algorithms, Data modeling, Sensors, Signal detection, Statistical modeling, Filtering (signal processing), Expectation maximization algorithms, Land mines, General packet radio service
In this paper, we propose an expectation maximization (EM) trained interacting multiple model (IMM) abrupt change
detector for land mine detection applications. The proposed EM algorithm learns the parameters of the different
models in real time without requiring a priori information on either the number of models or the model parameters.
Using the real ground penetrating radar (GPR) data, the learning performance of the EM-IMM technique is analyzed
and commented upon. Numerical receiver operating characteristics (ROC) analysis and detected images indicate
that the proposed EM-IMM based abrupt change detector has a better detection and imaging performance than the
conventional Kalman filter for land mine detection applications.
In today's information age, information and network security are of primary importance to any organization. Network intrusion is a serious threat to security of computers and data networks. In internet protocol (IP) based network, intrusions originate in different kinds of packets/messages contained in the open system interconnection (OSI) layer 3 or higher layers. Network intrusion detection and prevention systems observe the layer 3 packets (or layer 4 to 7 messages) to screen for intrusions and security threats. Signature based methods use a pre-existing database that document intrusion patterns as perceived in the layer 3 to 7 protocol traffics and match the incoming traffic for potential intrusion attacks. Alternately, network traffic data can be modeled and any huge anomaly from the established traffic pattern can be detected as network intrusion. The latter method, also known as anomaly based detection is gaining popularity for its versatility in learning new patterns and discovering new attacks. It is apparent that for a reliable performance, an accurate model of the network data needs to be established. In this paper, we illustrate using collected data that network traffic is seldom stationary. We propose the use of multiple models to accurately represent the traffic data. The improvement in reliability of the proposed model is verified by measuring the detection and false alarm rates on several datasets.
In this paper, we propose a novel chaos based ultra-wideband (UWB) sensor to enhance homeland security applications. The proposed chaos based modulation has a good resolution when used for wall penetrating applications. The receiver exploits the deterministic nature of chaos to cancel room reverberations avoiding complex synchronization procedure. Numerical electromagnetic (EM) simulations using finite difference time domain (FDTD) method are performed to illustrate the imaging performance of the proposed radar under real life surveillance situations with hidden and moving targets. The simulations are also employed to analyze the extent of penetrating ability of the proposed scheme for different structures. The effect of various structures and thickness on the detection performance are also commented upon.