Radar systems serve as foundational components in both civil and military aerospace infrastructures. Modern radar must not only distinguish between detection and non-detection but must also classify detected objects. Signal processing increasingly integrates machine learning models into complex systems, such as radar. Additionally, developments have fused signal processing with quantum computing, creating an emerging field of research. This paper examines the applicability of quantum machine learning models for radar signal classification, focusing on the impact of Ansatz depth on expressibility. Multiple challenges arise due to the immature state of noisy intermediate-scale quantum hardware and the computational complexity of quantum circuit simulation. Nonetheless, results indicate that shallow Ansätze with fewer than 70 gates are sufficient to achieve the maximum available performance per data-encoding operation.
Martinez et al. (Wed,) studied this question.