The first implementation of a fully photonics-based coherent radar system shows how photonic methods for radio-frequency signal generation and measurement may facilitate the development of software-defined radar systems. See Letter p.341
For well over a decade, there has been intense pursuit of photonic techniques for the synthesis and processing of broadband radio-frequency signals in applications ranging from wireless communications to radar. Such work is motivated by one main challenge: the lack of electronic systems capable of directly generating, processing and digitizing signals that have high central frequencies and large bandwidths while maintaining signal fidelity (minimal spurious-signal content and low noise levels). That challenge is sure to remain, given the escalating interest in higher centre frequencies and greater signal agility for software-defined radio and radar architectures — those in which the desired signal can be rapidly changed under software control. Such signal agility could offer, for example, increased radar resolution. On page 341 of this issue, Ghelfi et al.1 describe the first field demonstration of a coherent radar system based on photonic techniques. Their work shows that such techniques may indeed provide capability for next-generation radar systems.
Coherent radar systems2 use both the amplitude and phase (frequency) of the radar signal that is reflected from a given target (the radar return) to provide range and velocity data for the target. In such a system, a radar waveform of low centre frequency is designed — that is, its bandwidth is chosen — to give the desired range resolution. This waveform is then shifted to a higher centre frequency by multiplication with a stable, continuous-wave radio-frequency (RF) signal (carrier), and subsequently amplified and transmitted to the target. On reaching the radar receiver, the waveform is shifted to a lower, intermediate frequency, again through multiplication with a stable RF signal, and processed to extract the target's range and velocity.
Information on range is acquired from the time of arrival of the radar return. The range resolution is determined by the waveform's bandwidth-limited duration, and improves as this duration decreases. The need for increased resolution requires the use of higher carrier frequencies, which are necessary to obtain waveforms of shorter duration. Velocity data are obtained by comparing the frequency of the return waveform against a stable, continuous-wave reference signal of the same frequency as that of the original radar waveform. For a moving target, the return waveform will be shifted in frequency relative to its original frequency owing to the Doppler effect (at acoustic frequencies, this effect is analogous to the way in which the perceived pitch of a siren varies as it passes an observer). Accurate velocity measurement demands that the frequency of both the carrier and the reference signals be known with a high degree of certainty — that is, that their phase noise is low.
In a digital coherent radar, processing of the radar return is accomplished using a computer. Such processing requires the use of a high-precision analog-to-digital converter (ADC), which measures the amplitude of the return waveform at discrete instances in time and converts these measurements (samples) to digital data suitable for computer processing. This digitization requires that the time at which each sample is taken in the ADC is known to high accuracy, which in turn necessitates that the timing jitter — the variation in the time between successive samples — is minimal. Low phase noise and minimal timing jitter are intimately related: timing jitter is directly proportional to the phase noise mathematically integrated in frequency3. Minimizing the phase noise therefore minimizes the timing jitter.
Enter the mode-locked laser — a class of laser in which the optical modes of oscillation maintain a highly stable phase relationship (hence the term mode-locked). This phase stability results in a periodic train of laser pulses exhibiting low timing jitter. When two optical modes are selected by an optical filter placed beyond the laser and subsequently directed to a light sensor known as a photodiode, the sensor's electrical output is an RF signal with low phase noise. The frequency of this signal is readily tuned on the basis of the two optical modes that are selected, and the efficiency with which the RF signal is created is determined by the electrical response of the photodiode at the chosen frequency. For coherent-radar applications, such a laser provides an appealing tool that can both generate tunable low-phase-noise RF signals and serve as the basis for a low-jitter ADC. This is precisely what Ghelfi et al. have done to build a coherent radar system as part of a project called PHODIR ('Photonics-based fully digital radar').
In this project, a single mode-locked laser is used to generate the required continuous-wave RF signals and to sample the return waveforms in an optical ADC. Such operations are at the centre of the project's photonics-based radar transmitter–receiver, or transceiver (Fig. 1). Laboratory measurements of the resulting continuous-wave signals showed that signals with a frequency near 40 gigahertz exhibit integrated phase noise and timing jitter that are approximately half those of state-of-the-art conventional RF synthesizers. More impressively, because the timing jitter of the laser-pulse train is about ten times lower than that of the best electronic ADCs, the authors achieved record-precision digitization of continuous-wave signals at 40 GHz. This optical ADC has about 100 times better digitization fidelity than its best electronic equivalent. And the device showed comparable digitization fidelity to that of similar photonic ADCs4 over a bandwidth that was ten times larger — a remarkable result.
In a first-of-its-kind demonstration, Ghelfi and colleagues next incorporated their photonics-based transceiver into a coherent radar architecture operating at a carrier frequency of about 10 GHz, and tested it in a real-world situation by measuring the take-off trajectories of several aircraft. They compared the resulting range and velocity data against publicly available trajectory data from independent sources and found excellent agreement between the two data sets. The results illustrate that a photonics-based transceiver can be successfully used in a radar system.
The individual performances of the system's transmitter and ADC are world-class with respect to those of other photonics-based devices. Furthermore, the readily tunable carrier and reference signals, as well as the photonic ADC, remove much of the complexity of frequency translation that is inherent in all-electronic radar implementations. Certainly, these elements, and their combination into a functional system, are appealing components for future frequency-agile, software-defined radar architectures.
But an important question must be addressed: can photonics-based radars achieve similar or better performance than their all-electronic counterparts? Answering this question will require direct comparison of photonic and all-electronic versions of the same radar. Compared with the system described here, current commercial digital radar systems operating at a 10-GHz carrier frequency show substantially higher dynamic range (the range of received-signal power over which the radar receiver responds linearly). Because dynamic range translates directly into radar sensitivity and false-target detection, a high value for this quantity is crucial.
Another concern is one of radar range, the distance over which the radar can successfully detect a target. Increased range requires longer processing time, placing more stringent limits on long-term timing jitter and phase noise at a low frequency offset from the carrier signal — quantities that are difficult to minimize5 in photonics-based architectures.
Ghelfi and colleagues' study raises yet other questions. In what types of radar system could photonics have the most impact? Given a particular radar application, what requirements (associated with, for example, timing jitter and phase noise) are placed on our techniques, and can they be achieved? Pursuing the answers to these questions will no doubt lead to exciting — and challenging — research in the years to come.
Ghelfi, P. et al. Nature 507, 341–345 (2014).
Skolnik, M. (ed.) Radar Handbook 2nd edn (McGraw-Hill, 1990).
von der Linde, D. Appl. Phys. B 39, 201–217 (1986).
Kim, J., Park, M. J., Perrott, M. H. & Kärtner, F. X. Opt. Express 16, 16509–16515 (2008).
Fortier, T. M. et al. Nature Photon. 5, 425–429 (2011).
About this article
Scientific Reports (2016)