Geometric aspects of long-term noncoherent integration

A time-shift model for OFDM radar

“…Range walk is particularly problematic when using long integration times for detection. Without compensation, it limits the coherent integration gain, lowering the probability of detecting weak targets [7], [20], and is also a problem with incoherent integration [2]. This paper derives a time-shift model that addresses the range-walk problem and that can be efficiently implemented for OFDM signals, among others.…”

“…2 However, not all CP-OFDM codes are equally suited for the type of low-complexity radar processing considered here. As was the case in [5], each OFDM symbol is treated as a "pulse" and it is assumed that the maximum delay of a return from the target compared to the direct-path return from the nearest receiver is less than the duration of the cyclic prefix, which occupies the interval between pulses.…”

“…Section IV discusses the discrete-time implementation of the matched filters. In Section V, the delay-Doppler plots created using each of the methods in scenarios involving synthetically gen- 2 The DTMB standard puts pseudonoise in the guard interval between symbols rather than using a cyclic prefix. It has been claimed that the use of pseudonoise facilitates synchronization between the transmitter and the receivers.…”