Agile Multiple Pulse Coherent Lidar for Range and Micro-Doppler Measurement

Transcription

1 Agile Multiple Pulse Coherent Lidar for Range and Micro-Doppler Measurement Stephen M. Hannon, J. Alex Thomson, Sammy W. Henderson, Philip Gatt, Robert Stoneman, Dale Bruns Coherent Technologies, Inc. Lafayette, Colorado USA ABSTRACT A novel high time-bandwidth product waveform lidar has been developed. The lidar operates at the eyesafe 2 tm wavelength and produces a sequence of two or more cavity-dumped pulselets with a controllable intra-pulse spacing. The number of and spacing for the individual pulselets is adjusted to match the target and atmospheric characteristics. This waveform agility enables the sensor to operate at very long stand-off ranges. Performance predictions and results from recent field demonstrations are described. Keywords: coherent lidar, solid-state lidar, ranging, Doppler, vibrometry, imaging 1. INTRODUCTION We have developed a novel, agile waveform coherent lidar concept to provide a robust multifunction sensor capability for target acquisition, tracking and identification purposes. The base waveform, its generation and efficient processing, represents a breakthrough in coherent lidar hard target detection and discrimination. The coherent lidar operates at an eyesafe solid-state wavelength (2 im initially, and 1.5 im as well as MWIR wavelengths as transmitter technologies progress) in order to achieve a compact, rugged design capable of stand-off ranges of 2 km or more. Simultaneous precision range and micro-doppler target signatures are extracted with an extremely efficient processing architecture. Range resolution of one foot with simultaneous velocity precision better than 1 mm/sec are expected. Target ID can be performed with such imagery to a relatively high degree of fidelity because of the inherent spatial scale information that it provides. Additional target pixels can be accessed through the vibration signature that can be produced with the base waveform and/or its higher order variants. This paper presents a summary of recent development and demonstration activities associated with the agile waveform coherent lidar. Section 2 gives a brief background discussion of the base doublet pulse waveform and Section 3 details the rationale for and benefit of the agile waveform concept. Section 4 discusses recent and ongoing agile pulse laser transmitter development efforts. Section 5 presents sample results for both hard targets and aerosol (distributed) targets. Performance predictions for long-range hard target applications are then presented in Section 6. Section 7 summarizes the paper and outlines areas of focus for continued development. 2. DOUBLET PULSE BACKGROUND Pulse pair waveforms, limiting cases of the general pulse train, are routinely used in Doppler radar for both hard target and aerosol target applications. The pulse pair and poly-pulse-pair transmit formats have been applied to microwave radar for radial velocity and spectral width probing of weather echoes.1 We have extended the pulse 'Zrnic, D.S., "Estimation of spectral moments for weather echoes," IEEE Trans. Geosci. Electron. GE-17, 113 (1979). Part of the SPIE Conference on Laser Radar Technology andapplications Ill. Orlando, Florida April 1998 SPIE Vol X/98/$1. 259

2 pair (doublet pulse) waveform to eyesafe, near infrared optical wavelengths. The doublet pulse provides coherent Doppler lidar systems a substantial time bandwidth product (TB) of 1, or more with a very modest processing requirement. The waveform format comprises a pair of pulselets, each of duration t, separated by T, seconds. The range resolution is governed by the pulselet duration i, while the velocity precision goes as one over the pulselet separation, T. Figure I illustrates the base pulse format. Ambiguities in the velocity measurement arise as a result of the periodic structure of the waveform and occur every X/2T, mlsec, where? is the operating wavelength. These ambiguities can he readily dealt with in software in many applications. Higher order variants of the general doublet pulse (e.g., the triplet pulse) would also help eliminate such ambiguities. In the past, extension of the doublet pulse to optical wavelengths was not done for several reasons. First, it was difficult to efficiently produce the doublet pulse waveform optically. Second, unlike radar, the frequency and phase of the transmitted waveform was difficult to keep constant from one pulselet to the next. Our innovation solves these problems via two key advances. First, we generate the doublet pulse by intensity-triggered cavity dumping of an injection-seeded Q-switched 2 micron laser. Second, because the phase and frequency of the individual pulselets vary randomly, a monitor sample is taken of each transmitted pulselet. This monitor pulselet is used to correct for the random phase and frequency with the help of the stable CW master oscillator laser used in the heterodyne detection process. The general doublet pulse coherent lidar concept provides a truly novel, robust sensor architecture that can be applied across a broad range of applications and operating wavelengths. It enables extremely efficient, low cost processing architectures for very high TB waveforms. T s:fl 1/PRF T Figure 1 General Doublet Pulse format showing pulselet duration T, pulselet separation T, and doublet repetition frequency PRF. Lidar L t = to Lidar L+vT H V =4+ 4itvT/2 vt t = t + T Figure 2 The doublet pulse measures the target velocity by essentially measuring the phase change between the first pulselet and the second pulselet caused as the range to target is increased or decreased by an amount vt, where v is the velocity and T is the pulselet separation. 26

3 Both range and Doppler (velocity) processing precision performance will generally improve as 1/sqrt(CNR), where CNR is the carrier to noise ratio. Range resolution will depend on the pulselet duration and will go as LR = c-r12. The Cramer-Rao lower bound for the velocity estimate precision (Gaussian pulse) is given by2 8v = rms 2CNR[vg 2=2p,n O.26pin T JCNR JNavg The Cramer-Rao lower bound for the range estimate precision is given by SR = rms c/v J8ln 2 JCNR JNavg Figure 3 presents some early experimental results in which hard target was located a short distance from the transmitter and a range-doppler data sequence was compiled. The estimates were sorted by CNR on an estimate-to-estimate basis so that a glint-target statistical model was utilized in a companion simulation. Figure 3 plots the velocity estimate performance as a function of CNR. The experimental data is seen to be in excellent agreement with both the simulation and the CRLB. Departure from the CRLB at high CNR is due to finite monitor pulse CNR present in both the data and the simulation. For CNRs below 1 db, the velocity estimates become poor and are uniformly distributed over the 1 cm/sec velocity uncertainty interval for the nominal 8 tsec pulselet separation. U) E E C U) a) >. a) > Waveform CNRn (db) 4 5 Experiment results compare well with Monte Carlo simulation when a 33 db monitor waveform CNRn is incorporated in the Monte Carlo simulation Figure 3 Experimental data (symbols) and simulation prediction (solid line) of velocity measurement accuracy as a function of CNR assuming glint target amplitude statistics. Stationary hard target. CRLB is also shown (dashed line). Departure from CRLB at high CNR due to finite monitor pulse CNR; departure from CRLB at low CNR due to velocity estimate error saturation over velocity ambiguity interval set by 8 psec pulselet separation. 3. AGILE WAVEFORM COHERENT LIDAR CONCEPT Ongoing development efforts are extending the original doublet pulse concept and early demonstrations in several ways. The first is to extend to an agile waveform lidar transmitter in order to enable a more robust waveform that is matched to the target. One drawback of a fixed doublet pulse waveform is the velocity 2 H.L. Van Trees, Detection, Estimation, and Modulation Theory, Part I, JohnWiley & Sons,

4 ambiguities that arise at velocities proportional to one over the pulselet separation. For example, a 1 psec pulselet separation yields a maximum velocity ambiguity interval of 1 cm/sec. Continuity arguments can be applied to eliminate these ambiguities in many circumstances, a common practice in radar meteorology. Continuity-based de-aliasing techniques can be readily applied provided the target accelerations are below VaPRF rn/sec2. where Va 5 the velocity ambiguity interval and PRF is the waveform repetition frequency. However, for many scenarios of potential interest, the vibration signature may ride atop a very large amplitude, low frequency oscillation. For example, a rotating helicopter blade will have peak velocities of order 2 mis, with small-scale (a few nmilsec to a few cmlsec) vibrational velocities on top of the large, slowly varying velocities. Similarly, the sensor platform velocity may induce such a large scale velocity variability. With a selectable pulselet separation, or the potential for adding a third pulselet (a triplet pulse), the velocity ambiguity interval can be increased to permit more robust estimation of the small-scale and large-scale velocities. Another reason to consider an agile multiple pulse format is that micro-doppler signatures may only be of interest in certain measurement scenarios. More generally, range-doppler imaging may be performed with subfoot range resolution while maintaining velocity precisions of a few cm/sec. For such purposes, the pulselet separation need only be 1 sec or so. Pulselet separations shorter than 1 psec are also appropriate for atmospheric wind sensing applications. An agile waveform transmitter provides the flexibility to optimize the use of the available average power across multiple target engagement scenarios. At long ranges, when the target is not spatially resolved, the sensor must operate in a diffraction-limited mode and the transmitter must operate at high pulse energy and low PRF. Once the target is initially acquired, the pulse energy can be reduced to increase the PRF and provide estimates of higher frequency motions. At shorter ranges, the target becomes spatially resolved, at least partially. The target brightness increases, enabling the photons to be distributed across multiple array elements in a floodlight illumination mode. The PRF and waveform format can also be modified to extract target-specific range and Doppler features in a 3D imaging mode. As the range to target continues to decrease, the number of pixels available increases and the transmitter adapts in terms of energy, pulselet separation and pulselet format. Figure 4 illustrates these concepts. Spatially Unresolved - > 1 km- Partially Resolved - diffraction-limited mode km - - low PRF - - flood illumination mode - Fully Resolved < - high pulse energy - - moderate to high PRF moderate pulse energy - flood illumination mode -, A ' - - -highprf-,' L: \ 1 'A" -lowpulseenergy- I ' :3:: I t higher order waveforms I 1D RangeEcho Add Doppler (2D) Types of Motion: Transport Body Rotation Body Flexing/Flutter Rotor Movernent/BPF Engine Vibration,/.,uu Partially Resolved Quasi 3D Add Doppler (4D) 1-5 rn/sec.1-1 rn/sec (.1-1Hz).1-1 rn/sec (.1-2 Hz) 1-3 rn/sec (1-2 Hz) up to lcrn/sec (1-1kHz) Fully Resolved Fully 3D Add Doppler (4D) Doppler => large velocities-> vibration Figure 4 Generalized long range engagement scenario. An agile waveform transmitter coupled with a coherent focal plane adapts to optimize NCID over a continuum of spatial resolution regimes. The Doppler metric spans a wide range of target motion scales, from large-body motion to engine vibration. 262

5 Now consider the application classes for the coherent lidar sensor. The maximum coherent dwell period is often matched to the target decorrelation time. For hard targets, the range of pulselet separations is given below 2-2 jisec: very precise Doppler resolution, compatible with micro-doppler or vibration measurements (ambiguity interval is to.5 cm/sec for doublet waveform). 1-3 psec: moderately precise Doppler resolution, compatible with range-doppler imaging or moving target indication (MTI) (ambiguity interval is to.3 rn/sec for doublet waveform) For wind measurement applications, the range of allowable pulselet separations is significantly less than that for typical hard target measurements. Here, the spread of velocities in the probe volume (a 5-2 cm diameter by 1-1 m long cylinder of air) generally controls the decorrelation time:.1-.5 isec: good Doppler resolution, compatible with high resolution wind sensing (ambiguity interval is to 2 rn/sec for doublet waveform) 4. LASER TECHNOLOGY DEVELOPMENTS Initial laser developments focused on generation of a pair of cavity-dumped, injection-seeded 2 jtm laser pulses. A fast LiNbO3 Pockels cell was used for the cavity dumping. The cavity-dumped pulses up to several hundred microjoules have been generated with a nominal 7 ns FWHM in duration, corresponding to the round-trip cavity propagation time. A single-frequency 2 tm CW laser was used to injection seed the pulsed laser and an example of a seeded 7 nsec pulselet heterodyne signal and its return echo from a nearby target are shown in Figure 5. A demonstration of the doublet pulse generation is shown in Figure 6. The waveform is generated by cavity dumping each of two Q-switched pulses. Here, the cavity-dumped pulselets are spaced approximately 8 tsec apart with a waveform PRF of 1 khz. 1 Monitor and signal coherent ladar pulses L.onv Time (nsec) Figure 5 Sample heterodyne signal showing single- Figure 6 Sample pair of doublet pulses, each with frequency cavity-dumped transmitted pulselet and return 8 sec pulselet separation and inter-pulse-pair time of 1 signiet from short-range target. msec (1 khz PRF). The two-micron Tm:YAG laser transmitter has the capability for generating a wide range of flexible waveforms. As a result, waveforms with more complex temporal patterns can be generated. Specifically, waveforms with multiple pulselets, variable pulselet spacings, and variable pulselet durations are now available. This flexibility significantly expands the utility of the transmitter for lidar applications. Multiple pulselets can be used to eliminate velocity ambiguity concerns, variable pulselet separations allow operation over adjustable ranges of velocity and velocity precision, and variable pulselet durations allow adjustable detection bandwidths. The improved capability for flexible waveform generation results from the development of fast and highly controlled high voltage switching techniques applied to the intracavity pulselet-generating electro-optic modulator. Two general techniques are used to produce pulselets with a wide range of temporal separation. 263

6 For pulselet separations in the range of 1 ts to several hundred is the laser produces a set of Q-switched pulses with the required separation times, each of which is cavity dumped to produce a single short-duration pulselet. For pulselet separations in the range of 5 ns to 1 ts the laser produces a single Q-switched pulse, which is cavity-dumped repetitively at the required separation times. Figure 7 shows a series of demonstrated example output formats available from a single agile waveform transmitter. The waveform is generated by repetitively cavity dumping a single Q-switched pulse. Each pulselet has a duration of 7 ns. The fast and highly controllable high voltage switching techniques can also be used to generate pulselets with continuously adjustable pulselet durations, thereby allowing continuous adjustment of the signal detection bandwidth. The pulselet duration can be lengthened to accommodate the use of narrower bandwidth data acquisition electronics, and shortened to provide finer range resolution. [Higher order waveforms riplet with 1 ns and 23 ns separatlo a- U) U) -J A Time (ns) Doublet with 1 nsec separation Time (ns) 4 [Variable pulselet separation Time (ns) Figure 7 Demonstrated lidar waveform agility from a single laser transmitter. Repetitive cavity dumping permits variable pulselet separations as well as multiple pulselet generation. 5. SAMPLE MEASUREMENT RESULTS In this section, we present sample results for both hard and aerosol (diffuse) targets. Hard target data has been collected using a.5-lw average power Tm:YAG 2.12 tm doublet pulse transceiver developed by CTI for the Navy Command, Control and Ocean Surveillance Center (NCCOSC) in San Diego, CA. The lidar produces a pair of 25-5.tJ pulselets at a PRF of.5-1 khz. It is generally coupled to a 2" refractive telescope. The transceiver typically produces 7 nsec duration pulselets; however, with a simple adjustment of the cavitydumping electronics, longer-duration pulselets can be generated. The heterodyne frequency offset is 15 MHz. For the validation measurements described here, the sensor produced lower-energy 1.tJ pulselets at a PRF of 5 Hz and a pulselet separation of 15 psec. The measurements explore individual vibration spectra for a two-frequency, single range target. The general measurement setup is shown below. The measurements utilized a pair of speakers to mimic a complex target. Measurements of the individual speaker vibration spectra were conducted in order to determine the degree to which the vibration SNR is reduced for the doublet pulse sensor. 264

7 Figure 8 shows sample velocity time series (left panel) and vibration spectra (right panel) measured with the doublet pulse lidar. Figure 8 corresponds to a case in which only a single velocity is present at any instant. The velocity time history, however, comprises two distinct frequencies. The lidar easily detects these frequencies. The next set of measurements explore individual vibration spectra for a two-frequency, single range target. The general measurement setup is shown in Figure 9. The measurements utilized a pair of speakers to mimic a complex target. Measurements of the individual speaker vibration spectra were conducted in order to determine the degree to which the vibration SNR is reduced for the doublet pulse sensor. Figure 1 shows that the output SNR is reduced by a 5-2 db when two velocities are simultaneously present within a pixel. However, both frequencies are still clearly detectable in the vibration spectrum. This type of complex target is relevant to conditions where the beam footprint overlaps a sizeable portion of a targets structural vibration mode. 3 2 One Sided PowerSi rum / Gates E C) > Time (ms) Frequency (Hz) TSAV9 = us File: TsAvg = us Vamb/2 = 3.32 mm/s Tr = 6. ns Vamb/2 = 3.32 mm/s Vrms = 1.27 mm/s Avg MnCNRw = db Vrms = 1.27 mm/s Vavg =.15 mm/s Avg Si9CNRw = db Vavg =.15 mm/s Mon Width = 16 samples 5 of 5 Good Waveforms Mon Width = 16 samples Sig Width = 12 samples Sig Width = 12 samples File: Tr = 6. ns Avg MonCNRw = db Avg Si9CNRw = db 5 of 5 Good Waveforms Figure 8 Sample complex vibration spectrum for a simple target. Here, two frequencies are present simultaneously; however, there is only one target velocity ai any particular moment. The two frequencies are clearly evident in the spectrum (right panel) at 1 and 17 Hz. Target #2 17 Hz Speaker Beamsplitter Target #1 1Hz Speaker Figure 9 Configuration for two-frequency, two-velocity, single-range doublet pulse lidar vibration measurements. 265

8 Velocity vs. Time / Gates Time (ms) File: b TsAvg = us Tr = 6. ns Vamb/2 = 3.31 mm/s Avg MOnCNRw = db Vrms =.583 mm/s Avg SigCNRw = db Vavg =.22 mm/s 5 of 5 Good Waveforms Mon Width = 16 samples Sig Width = 12 samples Frequency (Hz) File: b TsAvg = us Tr = 6. ns Vamb/2 = 3.31 mm/s Avg MonCNRw = db Vrms =.583 mm/s Avg SigCNRw = db Vavg =.22 mm/s 5 of 5 Good Waveforms Mon Width = 1 6 samples Sig Width = 12 samples Figure 1 Sample complex vibration spectrum for a complex target. Here, two frequencies and two velocities are present simultaneously. The two frequencies are clearly evident in the spectrum (right panel) at 1 and 17 Hz; however, the SNR in the vibration spectrum is reduced over that in Figure 8. In order to better understand how the doublet pulse can still resolve the two vibration frequencies, albeit with reduced amplitude, consider the figure below. The doublet pulse lidar makes a phase difference estimate for the target return over the waveform duration. The resultant velocity time series for a two-velocity target will still yield a vibration spectrum with the appropriate (correct) frequency components (at 1 and 17 Hz in Figure 1 above), but modified vibrational amplitude estimates. For diffuse speckle targets, the resultant velocity time series will be further degraded by the amplitude fading for the component phasors. Single Velocity Target Two Velocity Target 2(t) i(t)=2ictvi(t)i? Figure 11 Phasor diagrams illustrating single-velocity and two-velocity target echoes sensed by a doublet pulse lidar. The estimated velocity results from the vector sum of the two phase delay vectors at any instant in time. We next consider sample wind measurement results collected with a shorter-pulselet-separation doublet pulse lidar. These measurements were collected in support of the development of a high resolution atmospheric velocity probe for the U.S. Navy. When completed, this doublet-pulse-based lidar will provide high spatial resolution (<3 ft), high velocity precision (1-2 kt) volumetric measurements of a localized region (5-1 ft cube) characterized by high wind shear. One such measurement application is the diagnosis of shear environments in and around helicopters and associated structures. Figure 12 shows sample wind measurement results for the doublet pulse lidar operating with a pulselet separation of 23 nsec. The pulselet duration was 7 nsec. A simple wind tunnel was constructed for the 266

9 measurements. The right hand panel of Figure 12 shows the doublet-pulse-lidar-estimated velocity time series for a nominal 3 rn/sec wind speed. The velocity spread within the wind tunnel was the likely cause for the larger velocity fluctuations relative to a zero-wind speed case (left panel) E 4 Predicted Value=Om/sec 4 '.4 Mean = -.1 rn/sec.4 Mean = 2.6 rn/sec Vrrns =.19 rn/sec Vrrns =.57 rn/sec Time(ms) Time (ms) C.)..; >-2 >-2 Figure 12 Sample wind measurements made with a 23 nsec pulselet separation. Left panel shows the velocity time series data for a zero-wind-velocity case. Right panel shows the velocity time series data for a rn/sec wind speed generated with a simplified wind tunnel. The larger fluctuations (larger RMS value) for the larger wind speed is believed to be due to the high turbulence generated by the wind tunnel. 6. SENSOR PERFORMANCE PREDICTIONS This section presents sample performance predictions for an airborne sensor. CTI's coherent lidar simulation software is utilized to generate these predictions. Model atmospheres are available via FASCOD3, and we have assumed a tropical model atmosphere with either 5 km or 25 km near-surface visibility. A broad range of atmospheric refractive turbulence profiles are available for predicting CNR losses for each slant path. Figure 13 and Figure 14 plot sample results in contour format. The lidar is airborne and at an altitude of 4 km with a 2 cm aperture. Much smaller apertures (5 cm) will be appropriate for shorter range applications. The carrier to noise ratio (CNR) is a measure of the average number of detected photons in a matched filter bandwidth. CNR contours are plotted at the 1 db level. These CNR levels are required to achieve acceptable pixel-level dropout and velocity precision performance. A diffuse hard target with a 2 m diameter and a diffuse reflectivity of. 1/it sr1 is assumed. The contour plots indicate for which target ranges and heights, the threshold CNR level is met. All target ranges and heights closer to the lidar will exceed this threshold. Figure 13 plots comparative 1 db CNR contours for 5 km and 25 km ground-level visibility and four operating wavelengths: 2. 1 m, I.55.tm, tm, and 1.59.tm. The lidar has an aperture diameter of 2 cm and an altitude of 4 km. We see that the near infrared wavelengths of 2. 1 im and 1.55 mm provide the best performance, a result of the excellent atmospheric transmission at that wavelength. We see that the best performance is achieved with the 2.1 pm wavelength due to the combination of better atmospheric transmission and smaller beam diameter at a given range relative to the target diameter. CO2 absorption significantly limits the 1.6.tm sensor performance. Water vapor absorption adversely effects both 1.6 m and 9.11.tm. 267

10 E 8 6, IU 2-2 LdUOd (rim) I I m Eft: D:16cm,-" ta:4m'.. Foc: mt. / t,431am: 2 coi - / 7Ref:.1/it Horizontal Range (km) E a, C) I 8 LdFTlUUd tilili) I E: 1 mj , / Eff: ! D:16cm - Foc: int/ 6 / tl3iam: 2 CI)1 / ZRet:.1/it / NM1-:oo /'A:4m Horizontal Range (km) Figure 13 Predicted 1 db CNR range-height contour plots tr four operating wavelengths and a mj/pulse coherent lidar operating at an altitude of 4 km. Contours correspond to operation at 2. It) pm, 1.55 pm, 9.11 pm, and 1.59 pm. Curvature of earth is indicated by shaded region. Left panel corresponds to 5 km ground-level atmospheric visibility and the right panel corresponds to 25 km groundlevel atmospheric visibility. Other sensor parameters are listed. Figure 14 plots the predicted 1 mm/sec velocity precision contours for an airborne coherent lidar. Four operating wavelengths are again considered: 2.1.tm, 1.55.im, 9.11 tm, and 1.59 im. Signal decorrelation during the measurement time can significantly limit the range performance of a micro-doppler lidar. Sources of signal decorrelation include relative lidar-target translation, target rotation, relative lidar-target rotation, and atmospheric refractive turbulence. Target rotation and/or relative lidar-target rotation dominate the loss of signal coherence in many scenarios. Relative target rotation results in different parts of the target having different velocity components along the lidar line of sight. The result is RMS velocity measurement noise that is equivalent to the RMS velocity spread in the illuminated portion of the target. At high CNR, the velocity precision is dominated by this velocity spread and no longer improves with increases in CNR. Figure 14 shows that the target rotation can significantly reduce the ability to precisely measure the point velocity data on the target. It should be noted that the range measurement performance is not significantly impacted by the presence of target rotation. In low refractive turbulence (such as for an airborne platform), the beam divergence of the shorter wavelengths is less, resulting in a smaller illuminated spot size on the target, which in turn results in less velocity measurement noise at shorter wavelengths. The target rotation limitation can be partially overcome by signal averaging as long as the PRF at which the averaging takes place results in complete signal decorrelation between pulses. It is important to note that any vibration sensing lidar must contend with these type of decorrelation mechanisms. Fortunately, for many realistic scenarios, effective rotation levels are relatively small and should not significantly reduce the maximum vibration stand-off range. For example, the rotation of the lidar-target line of sight is.1 rad/sec for a target located at 1 km range with a target velocity perpendicular to the lidar-target line of sight of 1 mlsec. 268

11 /inji: 2 Radial Velocity Precision (mmls) Radial Velocity Precision (minis) Radial Velocity Precision (minis) 1 ElnJ ad/inc Rol: 1 rad/. V4 io :_.. 1 Eli: 1 ElI:.1 ElI..1 TDI F i Tin F / i TDI F if rnioio/ TRIoiO/ 41 TRIO / / NM1.O A.' NM i NM1.O I(fl:_ Ein,J T Eln,J 6 / 6 I Horizontai Range (km) Horizonisi Range (kin) Horzoniai Range (kin) radls.1 rad/s.1 radls Figure 14 Predicted 1 mmlsec velocity precision contours for liur operating wavelengths and a I mj/pulse doublet pulse coherent lidar operating at an altitude of 4 km. The pulselet separation is 1 psec. Three values for target rotation are considered: rad/sec (left);.1 nod/sec (middle);.1 radfsec (right). 7. SUMMARY AND FUTURE DIRECTIONS We have developed and demonstrated a novel coherent lidar concept for precision range and micro-doppler measurement at large stand-off ranges. The agile waveform comprises one or more short-duration pulselets with separation times matching the controlling decorrelation time. This agile waveform provides a very high time-bandwidth product but with a very efficient signal processing architecture. The waveform agility expands the concepts application base to include very high resolution measurements of localized wind velocity fields. Laser demonstrations show that short-duration coherent lidar pulses can be produced at the eyesafe 2 trn wavelength and that selectable pulselet separations can be generated. For precision hard target velocity measurements, the waveform duration is 1-1 l.tsec, and for high-resolution wind measurements, the waveform duration (pulselet separation) is.1-i sec. The flexible pulse transmitter has been used to collect a variety of hard target and wind measurements that match theoretical predictions of sensor performance. Hard target velocity precisions less than.5 mm/sec and range precisions better than.1 m have been demonstrated. Wind velocity measurements with precisions better than I kt have also been demonstrated. Our development of the agile pulse lidar concept continues. Ongoing Air-Force-sponsored efforts will develop a more robust, higher-power 2.im transmitter laser. In addition, we are developing a scaleable and affordable coherent array receiver architecture by exploiting unique aspects of the high time-bandwidth product waveform. The Navy-sponsored high resolution wind sensor will be completed in late 1998 and demonstrated in ACKNOWLEDGMENTS The prototype doublet pulse lidar used in support of this work was developed under sponsorship from the Naval Command, Control and Ocean Surveillance Center (F. Hanson, technical monitor). Hard target demonstration measurements were funded under Phase I SBIR sponsorship from the U.S. Air Force Research Laboratory (M. Dierking, technical monitor). Agile waveform and wind measurement demonstrations were funded under Phase II SBIR sponsorship from the Naval Air Warfare Center, Aircraft Division (D. Carico, technical monitor). 269

12 本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 ( 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具