Construction of an Arbitrary Waveform Radar System

Doktorsavhandling, 2024

The thesis aims to develop and deploy continuous wave noise radar systems by addressing the self-interference issue and considering real-time implementation aspects. In contrast to the traditional pulse-Doppler radar, which generally operates with short, high-powered, and deterministic pulses, noise radars transmit continuous, low-powered, random, and preferably wideband signals. Noise radar systems offer several advantages over pulse-Doppler systems. The most notable (and desired) advantage is their low probability of interception properties — detecting and localizing a noise radar system is more challenging than a pulse-Doppler radar system.

However, few (or none) commercial or military noise radar systems exist due to the challenge of achieving relevant performance. The main problem is that self-interference, such as direct signal interference or clutter echoes, severely restricts the system's detection sensitivity. A significant amount of research has been dedicated to resolving the self-interference problem, and although good results have been achieved, more is required. Noise radar signal processing also requires high-speed digital electronics, and it is only recently that the performance of digital electronics has started to be on par with the requirements.

In this thesis, bistatic noise radar is considered a solution to the self-interference problem. By constructing a bistatic noise radar system, it is shown that separating the transmitter and receiver reduces the self-interference, thereby increasing the detection sensitivity. Furthermore, bistatic operation enables adaptive beamforming, which can be applied to further suppress self-interference — this is demonstrated using a multichannel receiver.

A real-time processor operating with a time-bandwidth product of 77 dB is implemented on a state-of-the-art field programmable gate array to investigate limiting aspects of real-time noise radar systems. The processor demonstrates that wideband noise radar systems are possible, but several limiting factors exist. One limitation is that operating with high time bandwidth products leads to several effects, such as range-walk, Doppler spread, and target decoherence, which must be managed. These effects are shown using offline data, and solutions are successfully applied. However, implementing these solutions in real-time systems is still an open question.

The most significant outcome of the thesis is the construction of a real-time bistatic noise radar system capable of detecting small UAVs at an operationally relevant distance of over 3.2 km. Minor improvements can significantly increase the detection range. This achievement demonstrates the readiness of noise radar technology for commercial adoption, reinforcing the thesis's primary goal.