在毫米波段ISAR成像仪器CWLFM高距离分辨率雷达

418 IEEE SENSORS JOURNAL, VOL. 11, NO. 2, FEBRUARY 2011 Instrumental CWLFM High-Range Resolution Radar in Millimeter Waveband for ISAR Imaging Álvaro Blanco-del-Campo, Alberto Asensio-López, Javier Gismero-Menoyo, Blas Pablo Dorta-Naranjo, and Javier Carretero-Moya, Student Member, IEEE Abstract—This paper shows a High-Range Resolution (HRR) Radar prototype development, transmitting in millimeter-wave band. The waveform selected to be used is the continuous-wave linear frequency modulation (CWLFM), which allows, at the same time, good range resolution, moderate unambiguous doppler range and reasonable operational range taking into account that high-frequency band is used. The radar prototype looks for High-Range Resolution (HRR) characteristics as well as flexibility, in order to obtain the best Inverse Synthetic Aperture Radar (ISAR) images possible depending on the applications and targets involved. That is, depending on the speed and target’s trajectories, some radar features, such as transmitted bandwidth, modulator frequency or sampling frequency, play a key role in obtaining the best possible results while using lowest electrical requirements. In Section IV, several different ISAR images are shown, pointing out not only the radar features achieved and their variety, but also their quality thanks to the hardware and electronics improvements developed. Index Terms—Continuous-wave linear frequency modulation (CWLFM), doppler processing, high-range resolution (HRR), inverse synthetic aperture radar (ISAR) images, millimeter wave- band, RF zoom scheme. I. INTRODUCTION I N RECENT YEARS, the advances made in solid-state tech-nology and real-time processing have given rise to a con- ceptual change in radar systems and, therefore their possible uses. Conventional radars, labeled as low range-resolution be- cause the targets are smaller than the size of one range-cell, are only able to detect and measure the target’s location during dwell-time. So, the measurement of their speed and calculation of their trajectories lack of precision, needing several turns of the antenna to improve it. Manuscript received January 21, 2010; revised May 07, 2010; accepted May 11, 2010. Date of publication October 07, 2010; date of current version November 19, 2010. This work was supported in part by the Project TEC2008 of the Spanish National Board of Scientific and Technology Research. The associate editor coordinating the review of this paper and approving it for publication was Prof. Evgeny Katz. Á. Blanco-del-Campo, A. Asensio-López, J. Gismero-Menoyo, and J. Carretero-Moya are with the Grupo de Microondas y Radar Departamento de Señales, Sistemas y Radiocomunicaciones, E.T.S.I. Telecomunicación (C-407), Ciudad Universitaria (UPM Madrid), 28040 Madrid, Spain (e-mail: alvaro@gmr.ssr.upm.es; vera@gmr.ssr.upm.es; javier@gmr.ssr.upm.es; jcar- retero@gmr.ssr.upm.es). B. P. Dorta-Naranjo is with the Grupo de Ingeniería de Comunicaciones, Departamento de Señales y Comunicaciones, Pabellón B de Telecomunicación, Campus Universitario de Tafira (ULPGC), 35017 Las Palmas, Spain (e-mail: pdorta@dsc.ulpgc.es). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/JSEN.2010.2053703 The current state of the art technology has brought about a greater thinking on high-range resolution radars, which im- plies transmitting much wider radio-frequency bandwidths. In this way, targets become extensive detections, that is, occupying many range-cells from the Radar Matrix [1]. This fact means that a big increase in the operations has to be made, feasible due to the advance in electronics, but allowing better opera- tional features, such as obtaining the target’s radar images [2], improved Automatic Target Recognition (ATR) algorithms or higher precision in their measurements. That is, completely new radar applications. Radar range resolution ( ), can be explained as the radar’s ability to distinguish between different targets, which are being illuminated simultaneously by the antenna’s system, but located at different ranges. According to Rayleigh criterion [2], range resolution and transmitted bandwidth are related by the equation (1) where range resolution; electromagnetic wave media propagation velocity; transmitted radio-frequency bandwidth. Using this formula it can be concluded that, the lower range resolution is, the bigger transmitted bandwidth is needed. But unfortunately, for the same operational range, more data has to be processed involving more complicated and expensive equip- ment. Conventional radar systems have range-cells of between 15 and 150 meters [3], [4], but reducing them, a priori, provides many advantages to the system and, potentially, new features. A. Clutter Power Reduction Clutter is defined as all unwanted reflections, or echoes, re- ceived by the system. It can be produced by several and dif- ferent objects such as the sea, the surface of the earth or rain, but what really matters is that these echoes can reflect more power than real targets which should be detected by the radar. So, it becomes necessary to develop quite sophisticated signal processing techniques to detect targets embedded in clutter. The unwanted power of the echoes is proportional to the size of the range-cell [5], so by improving the resolution, surface surveillance shipboard sensors [5], or coastal ones, can be devel- oped achieving better detection probability under bad weather conditions or spiky sea states, feature not easily possible using common resolutions. 1530-437X/$26.00 © 2010 IEEE BLANCO-DEL-CAMPO et al.: INSTRUMENTAL CWLFM HIGH-RANGE RESOLUTION RADAR IN MILLIMETER WAVEBAND FOR ISAR IMAGING 419 Another possible uses, such as airport-surface surveillance [7], can take advantage of the HRR characteristics. Showing or plotting raw video data directly onto the operator’s display, re- sulting a 2-D target image, providing useful additional informa- tion like Foreign Object Debris (FOD) detection. On the other hand, clutter radar cross-section reduction allows very complex difficulties such as periscope detection, small boats with immigrants, or launches for illegal trading to be overcome. All of them are uses characterized by very low radar cross section targets embedded in strong clutter and not easily differentiable using Doppler processing. Conventional Track Before Detection (TBD) techniques [8], [9] are no more well suited, and new processing schemes have to be included in order to detect targets rivaling the sea clutter [10], [11]. Its behavior, or statistical characterization, is no more “classic,” due to HRR features [12], and requires new studies in order to improve the detection algorithms. Unfortunately, there are no published articles using, or obtaining, experimental data with range resolution at about 15 cm. or 7.5 cm. The most, being at 1.5 m. and 3 m. [13], [14], or more recently with 0.3 m [15] but, in any case, using coherent data, that is, being able to process the sea echo phase. So, this lack of data became another purpose of the prototype [16], [17], requiring or imposing some electrical features, like modulation (Section II), which will be explained later. B. Minimization of Multipath Influence As is well established, the multipath effect cause, both in communications and radar systems, fading signal events which decrease the received signal power when the unwanted, or sec- ondary, path arrives at 180 out of phase compared with the straightforward way. In a High-Range Resolution (HRR) radar, this effect is minimized because of its intrinsic operation. If the difference in distance between the straight and reflected path is greater than the resolution’s cell, both signals are resolved and not added in the receiver. So, no fading or interference effect oc- curs. Then, the bigger range or better target detection features can be finally achieved. This specific characteristic is, or could be, especially useful for sea-skimming missile detection. C. Target Classification and Identification HRR radars are unquestionably well qualified for this task, not only for detecting and tracking targets over the surface of the earth [18], [19] but also, and taking a step forward, developing automatic target recognition algorithms, a hot topic in which several research groups are now focusing on [20] and [21]. Changing high-resolution radar into a lower one is equivalent to processing the data with a low-pass filter, losing or erasing fundamental information for establishing the target’s classifica- tion, or even identification tasks, in a proper way. As an example, on the upper side of Fig. 1, plots a simple range radar profile, belonging to the vessel shown in Fig. 7 and obtained using the prototype developed in this project. 1 GHz of bandwidth was transmitted, that is using (1), 15 cm of range resolution. It reveals, in a very clear way, several high-power reflectors which are differentiated and resolved in distance. These resolved high reflective areas of the whole vessel, or rather, their relative locations, are essential values for efficient ATR algorithms. Meanwhile, on the lower side of Fig. 1, the Fig. 1. Ship’s range dimension radar profiles with 0.15 (upper figure) and 3 meter (lower figure) range resolutions. same range profile using 3 meter range-resolution is shown. This second image is equivalent to having passed through a low-pass filter previous data. The location of the prominent scatters has been filtered out, and no ATR algorithm is able to achieve rea- sonable operational features under these conditions. Moreover, due to the worldwide current social and political situation, there is an increasing interest in surveillance, classifi- cation and identification tasks, especially using unmanned vehi- cles, where millimeter-wave band and continuous-wave linear frequency modulation (CWLFM) signals have several advan- tages, such as their light weight, small size and good power requirements [22]. D. Target ISAR Images Inverse Synthetic Aperture Radar (ISAR) processing tech- niques use the target’s motion for increasing the cross-range radar resolution. They work in a completely opposite way to the SAR ones but with the same aim, to obtain good bidimensional target images. While SAR techniques use the radar’s motion to induce Doppler frequencies onto targets, and use it for im- proving cross-range resolution, ISAR means that the radar has to remain static and exploit the target’s own dynamics in order to detect the Doppler frequencies. This fact, added with a suitable range resolution, makes the system able to resolve or distinguish between various parts of the whole target, even if they lay at the same range cell (see Section IV). Assuming that the target can be modeled as a solid rigid body, there is a direct relationship between the cross-range angle and Doppler frequency [23]. Unfortunately, the relationship depends on both, the radar’s line-of-sight (LOS) and the target’s dy- namics so, a priori in a general way, the usefulness and quality of the resulting images can be only estimated under certain con- ditions (Section IV), that is, to assure a good resemblance be- tween optical and ISAR resulting image. This paper is organized as follows. Section II explains why the CWLFM signal is transmitted, listing the most significant advantages for achieving the project’s aims. Section III ex- plains the prototype block scheme, emphasizing both the main 420 IEEE SENSORS JOURNAL, VOL. 11, NO. 2, FEBRUARY 2011 innovations and their feature. Later, Section IV summarizes the main experiments carried out, the performance of the prototype through studying the results obtained, and finally, throughout Section V, main conclusions obtained under the development of the project are brought together. II. TRANSMITTED SIGNAL: CWLFM The requirements of the project listed below, without a doubt, determine the best signal to be transmitted, since the continuous- wave sawtooth linear frequency modulated signal was, in all cases, the most suitable option available. A. Solid-State Technology: Power Transmitted 1 Watt If a pulsed signal is transmitted, much more peak power should be needed for the same maximum operational range as compared to transmitting a continuous one. Then, managing such high peak voltages, another manufacturing circuitry technology should have been selected, losing the inherent advantages of solid-state technology such as size, weight, reliability, and power consumption [24]. B. 0.075 m of Range Resolution Using Low Sampling Frequencies A CWLFM radar, for obtaining the target’s distances, com- pares a sample of the transmitted signal with the received echoes, after being amplified by a low-noise amplifier (LNA). This fact, made with a mixer by multiplying both signals, produces the “beating signal,” whose spectral components (labeled as “ ”) define target distances (“ ”). There is a linear relationship between them, combining transmitted signal features: modulator frequency and transmitted bandwidth (“ ” and “ ,” respectively) (2) In this way, beating frequencies (“ ”) can be, and usually are, much smaller than the transmitted bandwidth (“ ”) so, in con- trast to radiating other waveforms, a very high compression bandwidth is achieved at the receiver, reducing the binary rate originated. In other words, very good range resolutions can be achieved, needing very low sampling frequencies at the receiver, at least, for reduced distances. Moreover, (2) states that the ex- isting linear relationship between distances and frequencies, de- pends on transmitted signal slope ( ), and so the radar operator only can decrease the sampling frequency needed at the receiver by reducing the modulator frequency if the range resolution has to be maintained. Moreover, the CWLFM signal allows us to do this without changing the receiver chain or the processing scheme implemented, enabling a real-time modu- lator frequency selectable radar to be developed in a very simple way (Section III). In a mathematical way, (2) can be rewritten as follows: (3) desired maximum range to be controlled; sampling frequency at the receiver. So, the maximum operational range (“ ”) can be increased by maintaining the transmitted bandwidth (“ ”) and sampling frequency (“ ”), with just reducing the modulator frequency (“ ”). Obviously, this change brings about other limitations as explained next. C. Maximum Unambiguous Doppler Range First premise, using solid-state technology, ruled out pulsed waveforms from the very beginning. However stepped fre- quency waveforms, pulsed trains with frequency hopping, are very common in published systems whose goal is to achieve a very good range resolution without needing high sampling frequencies [25], [26]. A priori, it is possible to achieve all objectives of the project using intrapulse modulation [25], but unfortunately, for HRR, a large number of pulses would be needed, resulting in a low unambiguous Doppler range, not big enough for good ISAR images [27]. The unambiguous Doppler margin using a CWLFM is (4) while for a pulsed one with frequency hopping (5) pulse repetition frequency; number of pulses integrated for the required bandwidth. Range resolution can be improved by increasing “ ” but, as is stated in (5), the unambiguous Doppler margin will also be re- duced. Then, if the target’s Doppler bandwidth is greater than the system’s unambiguous one, the images would appear “bent” and its usefulness enormously reduced. D. Waveform Parameter Flexibility and Ease of Generation As will be explained in the following section, using a DDS in the generation scheme allows the characteristics of the trans- mitted waveform, “ ” and “ ,” to change very easily and in real-time, an essential feature for optimizing the ISAR im- agery results. Moreover, the CWLFM waveform enables both the receiver scheme and signal processing to become unaltered in spite of changing, that is, varying the radar’s range resolution and Doppler’s unambiguous margin [(2) and (4)]. In fact, there is only one real problem. Due to its continuous wave characteristic, two antennas instead of one are required, needing more space and weight. But, at least, thanks to the use of the millimeter waveband, they become smaller and lighter. III. SYSTEM BLOCK SCHEME The signal generation block can be considered as the most striking subsystem of the prototype (Fig. 2). Linear frequency modulation is obtained by using a commercial VCO (Voltage Controlled Oscillator HMC398 from Hittite) which is able to sweep 1 GHz of the bandwidth, between 14 and 15 GHz, used within a PLO (Phase Locked Oscillator) scheme whose refer- ence signal is generated by a DDS [28]–[30]. Inherent frequency BLANCO-DEL-CAMPO et al.: INSTRUMENTAL CWLFM HIGH-RANGE RESOLUTION RADAR IN MILLIMETER WAVEBAND FOR ISAR IMAGING 421 Fig. 2. System block scheme. linearity, an essential feature for the overall project success [31], [32], lies in the DDS itself, while induced nonlinearities, such as thermal or power variations as well as time degradation, are compensated continuously by the locked loop. The proposed scheme (Fig. 2), divides the RF VCO output by 128, due to an embedded divisor in the VCO itself and two other commercial ones ( ), thus, for 1 GHz of RF bandwidth sweep, between 14 and 15 GHz, the DDS has to generate an LFM of 7.8125 MHz bandwidth, from 109.375 to 117.1875 MHz. The selected DDS was the AD9954 model, from Analog Devices, which uses an external clock of 400 MHz generated from a 100 MHz Crystal oscillator, placed as overall prototype reference signal. The AD9954 model, in fact, has two different DDS working with the same clock, and both can be programmed in real time and independently through a parallel bus connected to radar’s operator workstation. One of these two DDS generates the LFM reference signal for the loop, while the other one is used as a tuned local oscillator in the heterodyne scheme receiver implemented [27]. As a control measure, a dig- ital oscilloscope has also been placed for sampling the “Tune” signal, which excites the VCO, and plots it on the radar’s oper- ator display through the oscilloscope’s USB port. Once the PLO’s LFM waveform is generated, it goes through a commercial passive doubler from Hittite, HMC331, reaching the whole RF transmitted bandwidth goal, 2 GHz between 28 and 30 GHz. This doubling generation scheme is very easily of developing, but has its limitations. Maybe most important one is the maximum slope available “modulator period-bandwidth swept,” imposed by the cutoff frequency of the low-pass filter used in the locked loop. In this case, the maximum “time-fre- quency” slope available is 6 THz/s, that is, a maximum modu- lator frequency of 3 KHz sweeping whole 2 GHz RF bandwidth of the prototype. Fig. 3 shows the spectrum obtained for a continuous wave of 28 GHz (dark line) without modulation. At first sight, doing a few easy calculations, it can be assumed that the phase noise could be a great problem for the resolution required but, fortu- nately, it is strongly decreased by the equalization process made by the receiver mixer taking advantage of the correlation be- tween the local oscillator, in fact, a sample of the transmitted signal, and the echoes received [28]. The other feature noticeable in Fig. 3 is the power obtained, made at the output of a high gain amplifier, model HMC283 from Hittite, placed after the aforementioned passive frequency doubler. Moreover, transmitter has another last block, made up of a power amplifier, Triquint’s TGA1172, and a directional coupler, whose purpose is to get the sample of the transmitted signal to use it as a local oscillator in the receiver. So, at the power amplifier’s output, 29 dBm are finally achieved. As regards the data acquisition system, one