Detectability models for multiple access low-probability-of-intercept networks

The concept of multiple access low-probability-ofintercept (LPI) networks has generated considerable interest in the past few years, due to increased demands for highly connected, yet covert, communications in the tactical battlefield. A number of research efforts have concentrated on network design issues, such as throughput, packet switching, and message routing, while others have focused on the detectability of the network links. However, these detectability analyses have traditionally emphasized the interception of a single network transmitter, as opposed to the network at large. This work presents a different approach to detectability analysis of the LPI network. It is assumed that the interceptor does not attempt to isolate a particular emitter, but instead decides whether or not a network is currently operational. Hence, detection of the network may be based on energy received from either a single or multiple transmitters. The candidate network considered here uses frequency hop code division multiple access (FH-CDMA), in which users are assigned orthogonal frequency hop patterns. The total network bandwidth is defined as W1, and the network “on-time” is T1 seconds. The network has M =W1=W2 contiguous channels, and uses a hop rate of 1=T2 =N=T1 hops/s. Omnidirectional communications antennas are assumed. The two intercept receivers used here are the wideband radiometer, which forms a detection decision based on a single energy measurement on the T1£W1 time-frequency space, and the channelized radiometer (Fig. 1), which makes NM energy measurements on smaller time-frequency cells T2£W2, which are then combined to form an overall decision.