Adaptive Waveform Design for Target Localization and Tracking

Adaptive Waveform Design for Target Localization and Tracking for Cognitive MIMO Sonar Joseph Tabrikian Underwater Acoustics Symposium Tel-Aviv University June 17, 2013 Research students: W. Huleihel and N. Shraga

Outline q q q Introduction Cognitive MIMO radar/sonar configuration Adaptive waveform for target localization Adaptive waveform for MIMO sonar – static scenario Adaptive waveform design for MIMO sonar – dynamic scenario Conclusions and future work

Introduction - Cognitive Radar/Sonar Adaptive Waveform Design Transmit Signal Environment Receive Signal Adaptive Receiver Detection/ Localization/ Tracking/ Classification Key point: Transmit waveform is designed at very low SNR’s before the target is detected.

Cognitive MIMO Radar/Sonar Configuration

Cognitive MIMO Radar/Sonar Configuration

Cognitive MIMO Radar/Sonar Configuration Target dynamic model Optimal Adaptive Waveform Design Optimal Receiver noise Detection/ Estimation/ Tracking

Waveform Design for Optimal Target Localization Considered criteria: q Bayesian Cramér-Rao bound (BCRB) q Simple, analytic expressions q Ignores large-errors/threshold phenomenon q Reuven-Messer bound (RMB) q Higher complexity q Takes into account large-errors/threshold phenomenon and therefore is able to control the sidelobes

Simulations – Cognitive MIMO Radar

Simulations – Cognitive MIMO Radar BCRB-based waveform design Posterior pdf’s and transmit beampatterns . Auto-focusing effect: Automatic beamforming before detection/estimation.

Simulations – Cognitive MIMO Radar RMB-based waveform design Posterior pdf’s and transmit beampatterns . Auto-focusing effect: Automatic beamforming before detection/estimation.

Simulations – Cognitive MIMO Radar Single target – direction estimation accuracy: ASNR=-6 d. B k=6

Cognitive MIMO Sonar

Simulations – Cognitive MIMO Sonar

Simulations – Cognitive MIMO Sonar Single target – posterior pdf:

Simulations – Cognitive MIMO Sonar Single target – beampattern:

Waveform Design for Optimal Target Tracking o Dynamic model: o What is the optimal transmit (spatial) waveform for tracking?

Simulations – Cognitive MIMO Sonar Target Tracking

Simulations – Cognitive MIMO Sonar Single target – posterior pdf (via Monte-Carlo):

Simulations – Cognitive MIMO Sonar Single target – posterior pdf (via Monte-Carlo):

Conclusions and Future Work o o o A new optimal waveform design approach for cognitive MIMO radar/sonar is proposed based on minimizing the BCRB and RMB at each step using the measurements from previous steps. The RMB-based algorithm was shown to provide better results, since it is able to control the sidelobes. This approach provides an automatic focusing array: beamforming before detection or estimation. The method was adapted to consider dynamic targets, which can be interpreted as track-before-detect in transmission. Further research will cover the following issues: o Taking into account environmental uncertainties, o Wideband signal model, o Realistic shallow water channel simulations, o Considering other optimization criteria, such as probability of detection.

Thank you!

Simulations – Cognitive MIMO Radar