Improved computational performance for distributed passive radar processing through channelised data

Passive radars provide excellent Doppler spectral estimates for deep, fluctuating targets. Because the radar illumination is unpredictable, a full lag-profile or cross-ambiguity analysis must be performed to extract the scatterer power spectrum as a function of range. Two techniques (coherent integration and channelisation), that may be used to achieve significant speed improvements, are demonstrated. Coherent integration reduces the data bandwidth. While this step reduces the computational burden to an effort readily achievable by modern desktop computers for observing FM broadcasts from a single receiver, the load increases when observing more transmitters, using more antennas in interferometric analysis, and using a larger bandwidth source such as DTV. A further improvement is developed; it works by 'channelising' the illuminating and scattered signals, and splitting the spectral bandwidth by an integer factor, resulting in a computational reduction. The initial coherent integration method is presented, then a variety of implementations of the channelised improvements are shown. Properly implemented, these methods produce significant speed increases on ordinary desktop workstations. Channelised methods work nicely with transmitters whose spectral content is not white, permitting the elimination of noise power which lies outside the transmitter bandwidth, as well as beyond the decimation filter bandwidth of digital receivers.