Model-based waveform design for optimal detection: A multi-objective approach to dealing with incomplete a priori...

Area of Science:

• Signal processing
• Optimization theory
• Sensing technologies

Background:

• Designing optimal transmit signals for active sensing is challenging with incomplete target/environment knowledge.
• Existing methods may not perform well under model uncertainty.
• Multi-objective optimization offers a framework to handle diverse potential models.

Purpose of the Study:

• To develop energy-constrained optimal transmit signals for active sensing.
• To address uncertainties in target and environmental models.
• To enhance detection probability in challenging sensing scenarios.

Main Methods:

• Formulated the problem as a multi-objective optimization task.
• Incorporated multiple potential target, interference, and clutter models.
• Utilized a maxmin (maximize the minimum) objective function for detection probability.

Main Results:

• Maxmin waveforms judiciously allocate energy for robust detection across models.
• Demonstrated improved detection performance compared to LFM signals and wrongly assumed target spectra.
• Achieved performance comparable to designs matched to the correct model.

Conclusions:

• The maxmin formulation provides advantageous optimal waveforms for energy-constrained active sensing.
• This approach robustly handles uncertainties in target and environmental characteristics.
• The maxmin problem formulation is proven to be convex, ensuring efficient solution.