Optimal transmission pulse shape for detection and identification with uncertain target aspect

Abstract

This paper investigates the optimization of the transmission radar pulse shape to maximize either target detection or identity discrimination between the T-72 and M1 tanks under conditions of aspect uncertainty, Significant performance improvements in detection and identification are obtained using optimized transmission waveforms over that of standard chirped pulses. tion discrimination using two simulated surface mobile targets: the T-72 and M1 main battle tanks. SAICChampaign [S] generated VHF-band and X-band radar signatures for a single elevation angle of 15" and the full spectrum 0" -360" of aspect angles relative to the sensor. The X-band data cover frequencies between 9.28-9.92 GHz, whereas the VHF-band data span 225-375MHz.

Waveform design with time and frequency constraints for optimal detection of elastic objects

“…Rather than making such restrictive assumptions, it is desirable to design signals that are robust to target model variations [39,40,41]. This suggests the formulation of a signal design problem whereby the optimal solution is one that maximizes the minimum probability of detecting a target within an assumed class of potential target models.…”

Waveform design for wideband radar target recognition based on eigensubspace projection

“…Aiming at the waveform design problem for target recognition with uncertain TRO, the MESMD method is proposed in [4] through maximising the expected value of the Mahalanobis distance between the received echoes. For L target recognition, the cost function of the MESMD method can be extended and written in discrete form as where …”

“…Since it may greatly improve the performance of radar system in target detection, tracking and recognition, adaptive waveform design has been regarded as a promising technique and receives intensive attention recently. Consequently, many researchers [1–7] have studied waveform design techniques for maximising the recognition rate of extended targets.…”

“…Unfortunately, in the real world, the prior knowledge of target–radar orientation (TRO) is usually imprecise and the TIR is very sensitive to TRO, namely, one cannot obtain the exact TIR in transmission waveform design. In order to address this problem, Garren et al [4] proposed MESMD method through maximising the expected values of the square of Mahalanobis distance between the received echoes of different targets in a TRO angular region. However, only under the assumption of tiny TRO angular region, maximising the expected values of the square of Mahalanobis distance (MESMD) is able to work well, since in that scenario TIR changes slightly and can be approximately modelled by a simple distribution in the angular region [4].…”

“…In order to address this problem, Garren et al [4] proposed MESMD method through maximising the expected values of the square of Mahalanobis distance between the received echoes of different targets in a TRO angular region. However, only under the assumption of tiny TRO angular region, maximising the expected values of the square of Mahalanobis distance (MESMD) is able to work well, since in that scenario TIR changes slightly and can be approximately modelled by a simple distribution in the angular region [4]. In practice, because of the TRO sensitivity, TIR varies acutely along with TRO and usually follows multimodal distribution in the angular region [10].…”