Translational Doppler detection using direct-detect chirped, amplitude-modulated laser radar

Transcription

1 Translational Doppler detection using direct-detect chirped, amplitude-modulated laser radar William Ruff, Keith Aliberti, Mark Giza, William Potter, Brian Redman, Barry Stann US Army Research Laboratory ATTN: AMSRD-ARL-SE-EE 2800 Powder Mill Road Adelphi, MD ABSTRACT Translational Doppler detection using a direct-detect, chirped, amplitude-modulated LADAR is demonstrated. A nonoptically coherent system is utilized to lower the target-induced Doppler shift to tens of Hz, as opposed to several hundreds of megahertz for a coherent LADAR design. The Doppler shift is imposed on both the laser carrier and on the intensity modulation impressed on the laser carrier. Data indicates that the LADAR is able to measure the Doppler shift due to translational motion of a moving target. 1. INTRODUCTION The U.S. Army Research Laboratory (ARL) is developing low-cost laser radar (LADAR) systems that utilize diode-laser transmitters. Diode lasers cost significantly less than solid-state lasers and are more electrically efficient. Pulsed diode lasers have limited peak power capability due to damage at the output laser facet that, in a typical pulse operation, limits the amount of total energy that can be focused onto a target. Low-cost cw laser diodes, however, with optical power levels in the low Watts, are available commercially. With this in mind, ARL began investigating a unique, optically incoherent, LADAR architecture that uses frequency-modulated, continuous-wave (FM/cw) radar principles [1]. In this architecture, transmit and receive radar antennas are replaced by an intensity-modulated cw semiconductor laser diode and an optical receiver, respectively. This architecture uses the coherent radio frequency (RF) waveform to extract target range and velocity, while simultaneously utilizing the high frequency of the optical fields to achieve high angular resolution. Due to similarities between the LADAR and radar architectures, RF waveforms and signal processing techniques developed over the past 50 years for FM/cw radar can all be applied directly, or with only slight modification, to the LADAR architecture. More recently, ARL has been developing a near-infrared LADAR System that utilizes interdigitated-finger metalsemiconductor-metal InGaAs-based photodetectors (MSM-PDs), operating as optoelectronic mixers (OEMs) [2-4]. In this implementation, the MSM-PD bias [local oscillator (LO) voltage] and laser illumination (RF) are both amplitude modulated by a linear, frequency-modulated (chirp) signal. A portion of the laser illumination, reflected back from a target, is mixed with the LO signal in the detector. This mixing produces an intermediate frequency (IF) signal that is proportional to the target range. When utilized in the LADAR system, the MSM-PD OEMs eliminate the need for wideband transimpedance amplifiers and microwave mixers in the LADAR receiver thus allowing the use of low frequency, commercially available amplifiers. This, in turn, reduces both the cost and complexity of the system. ARL is currently designing a breadboard LADAR to measure both the range and velocity of a moving target using conventional range-doppler processing. A coherent LADAR design at an optical carrier frequency of 1.55 µm was initically considered for this application, however, target-induced Doppler shifts of several hundreds of megahertz were expected. This is not practical given other receiver constraints and, therefore, we are now considering a novel directdetection system based on our FM/cw architecture with the intensity modulation operating at a center frequency of 30 MHz on a 1.55-µm optical carrier. Lowering the modulation frequency lowers the target-induced Doppler shift to approximately 60 Hz at the expected target velocities, which is compatible with our receiver.

2 In this paper we demonstrate the ability of a non-optically coherent LADAR to measure the Doppler shift due to target motion. Data collected using a modified, existing LADAR System indicates that the Doppler shift is not only imposed on the laser carrier, but also on the intensity modulation impressed on the laser carrier. 2. FM/cw LADAR THEORY ARL s LADAR is based on frequency-modulated continuous-wave (FM/cw) radar. The data presented in the paper is obtained with a triangular, linear, frequency-chirp modulation. A detailed explanation of the signal processing used with this system can be found in most books that discuss basic radar theory [5]. Below, we discuss some of the basic operating theory for clarity and conciseness. F f stop Frequency Chirp Waveform: Triangular f IF_up τ f IF_down Transmitted Signal Received Signal f start T chirp Time f Doppler Figure 1. Transmitted and received signal for a system using a triangular chirp waveform. A frequency versus time diagram of the transmitted and received signals for the LADAR is shown in Figure 1. The dotted line in figure 1 represents the waveform received from the target. The received waveform is delayed with respect to the transmitted waveform by a time τ that represents the time it takes the signal to undergo a roundtrip to the target and back, i.e., τ = 2R c, where R is the target range: this is represented by moving the received signal left and right in the figure 1. The change in the received signal frequency is due to a Doppler shift induced by the target motion along the LADAR line of sight which is represented by moving the received signal up and down in Figure 1. The average Doppler frequency shift can be calculated using the center frequency of the modulation, f Doppler 2vf center =, (1) c where f center = f start + F 2 and v is the target velocity. The instantaneous frequency difference between the transmitted and received waveforms, or intermediate frequency f IF, can be obtained graphically from the diagram as for an up-chirp and as f F IF _ up = τ f Doppler, (2) Tchirp F f + IF _ down = τ f Doppler (3) Tchirp for a down-chirp. For the case where the target is stationary, both the up- and down-chirp intermediate frequencies will be identical and proportional to the range to the target by

3 f IF F 2R F = τ =. (4) T ct chirp chirp Taking the fast Fourier transform (FFT) of this signal over a chirp period allows the signal frequency, or equivalently, the target range to be easily measured. The frequency resolution of this measurement is equal to 1 T which corresponds to a range resolution, R, of the measurement equal to FFT is referred to as a range cell. chirp R = c 2 F [6]. Each frequency bin or cell in the For the case of a moving target, the intermediate frequencies of the up and down chirp will be different. If the Doppler frequency shift is greater than the frequency resolution of the measurement, then the Doppler frequency can be measured with one up and down chirp by f Doppler ( f f ) 1 = IF _ down IF _ up (5) 2 and the range dependant frequency can be measured using f IF ( f + f ) 1 = IF _ down IF _ up. (6) 2 If the Doppler frequency shift is less than the frequency resolution of a single measurement, the intermediate frequency of the up and down chirp will appear to be the same and the Doppler frequency cannot be measured as before. For this case, repeated measurements of the intermediate frequency of the up and down chirps, made for a length of time greater than 1 TDoppler, can be used to measure the Doppler frequency. It can be shown that the complex amplitude of the FFT used to measure f if during each up or down chirp will oscillate at the Doppler frequency over the repeated measurements. By simply taking a second FFT at each range cell across the repeated chirps, the data is mapped into a range-doppler space, and the targets range and Doppler or velocity vector along the line of site of the LADAR can be measured simultaneously. The Doppler frequency measurement resolution is limited by the amount of time the data was acquired, similar to the range dependant frequency measurement. III. EXPERIMENTAL MEASUREMENT SETUP Figure 2 is a basic electronic block diagram of the LADAR system. A trigger signal generated from a field programmable gate array (FPGA) initiates the generation of a triangular chirp signal that serves as both the laser modulation and LO signal. The chirp signal in our case is programmable by the user and may have a start frequency and a stop frequency in the tens-to-low hundreds of megahertz (an up-chirp is shown at the top of Figure 2). To modulate the laser intensity, the chirp signal is fed into a wideband RF power amplifier whose output is summed with a dc current in a bias tee to provide a modulated current drive for a semiconductor diode laser. For the experiments described in this paper, a 1.55-µm wavelength semiconductor laser operating at 1-mW average power was intensity modulated to 2-mW peak-to-peak by the RF. The divergent laser beam from the semiconductor laser is collected with a collimating lens and focused to a small spot that is aligned to intersect with the detector field of view (FOV) on the target. A portion of the laser beam is reflected from the target, collected by the receiver optics and imaged onto a single InGaAs MSM-PD OEM [8] detector located at the focal plane of the receiver optics.

4 Figure. 2. Block diagram of the scanner-less LADAR measurement setup. The photocurrent waveform from the detector is converted into a voltage signal using a transimpedance amplifier (TZA). Following the TZA are two amplifiers, one with a fixed gain and the other with a computer programmable gain, that are used to set the frequency response of the receiver and to buffer the signal for the following A/D converter. The amplifier chain has a computer-controlled equivalent impedance of 440 MΩ to 44 GΩ and a computer controlled bandwidth of 50 khz to 120 khz with a cut-on frequency beginning at 2 khz. The A/D samples the voltage from the detector and transfers the voltage data into computer memory. The detector voltage is measured periodically over each up and down chirp at a rate adequate to sample the highest expected f if. The resulting data set contains the IF waveform that includes all of the magnitude, range, and phase information derived from the light reflected from all scatterers in that pixel [1]. The entire LADAR measurement setup utilized in the paper was previously fully characterized using a GaAs MSM-PD OEM linear detector array [7]. IV. EXPERIMENTAL RESULTS A white target board, mounted on a metallic carrier, was guided on a metal track to provide a target for the Doppler detection. A string tied to the front of the carrier was used to move the target towards the LADAR and a spring was used to retract the target to its original starting position. The initial target position was located approximately 8 m from the LADAR and the range of translational motion was 109 cm, defined by two c-clamp stops. The chirp generator was programmed to output 128 up-down chirps at an 800-MHz center frequency with Fs of 75 and 37.5 MHz at rates of 49 and 98 Hz, respectively. The 800-MHz center frequency was chosen so that a number of Doppler cycles could be measured with the small length of target travel available in the laboratory. The target, when pulled the full length of travel, traverses approximately 6 Doppler cycles at 800 MHz. A 193-ns delay line was inserted between the chirp generator and the LO port on the MSM-PD OEM detector to artificially move the target out to a range far enough for the IF to be in the pass-band of the receiver electronics. The up and down chirp data was separated during the processing and are nearly identical except for the change in the sign of the target s Doppler shift. For brevity, only the down-chirp data is shown below.

5 Figure 3. Time series of Hanning-weighted FFT of IF signal from stationary target. Figure 4. Real part of Hanning-weighted FFT of IF signal from stationary target. Figure 3 is plot of the time series collected during a down-chirp for a stationary target and Figure 4 is a plot of the real part of its corresponding Hanning-weighted FFT, collected at a F of 75 MHz and a chirp rate of 49 Hz. From figures 3 and 4, the range to the target is between 10 and 11 range cells, found by counting the number of cycles in the time series data or by the signal in the FFT, which corresponds to an artificial range of between 20 to 22 meters. Figure 5 is a range- Doppler image of 64 down-chirps collected over a 2.6 second interval, formed by displaying the magnitude of the

6 complex FFT taken at each range cell for all 64 down chirps. This image simultaneously displays the target range and Doppler shift corresponding to the targets translational motion, which, for this stationary target, falls on the 0 Hz Doppler line at the center of the image. Figure 6 is a plot of the real part of the Hanning-weighted FFT of the 10 th range cell for each of the 64 down-chirps taken when the white target was pulled towards the LADAR and then allowed to retract back to its original position and, again, with the chirp generator set to a F of 75 MHz and a chirp rate of 49 Hz. It was very difficult to move the target at a constant velocity using the pull string and so the target accelerated during the image collection. This is visible in the range-doppler image, shown in figure 7, as a smearing of the target signal in the Doppler dimension. The target moved the full 109 cm during the first 35 chirps and, from figure 6, you can see that it transitioned nearly 6 Doppler cycles, as expected. The target only transitioned 3 Doppler cycles before the end of the data collection, indicating that the target had not returned to its original position by the end of the data collection. The decrease in the amplitude of the oscillation as the target moved from the back to the front of the rail indicates that the target may have moved through the 2-m range cell and into the adjacent range cell. To improve the previous experiment, we set the chirp generator to chirp at a faster 98 Hz rate to sample the target s more constant motion while it transitioned through the middle portion of the rail. We also changed the chirp F to 37.5 MHz, which corresponds to a R of 4 m, to keep the moving target in one range cell. Halving the range resolution also has the effect of moving the target signal from the 10 th range cell to the 5 th range cell. Figure 8 is a plot of the real part of the Hanning-weighted FFT of the 5 th range cell for each of the 64 down-chirps taken when the white target was pulled towards the LADAR. The target appears to be accelerating for the first 10 chirps and then moves at a more or less constant velocity for the duration of the measurement. This is indicated in the range-doppler image, shown in figure 9, where the target peak appears in the 5 th range cell and the +5 Hz Doppler cell, with some smearing in the Doppler dimension due to the initial acceleration. Figure 10 is a plot of the real part of the Hanning-weighted FFT of the 5 th range cell for each of the 64 down-chirps taken when the white target was retracted to its original position. As before, the target appears to accelerate in the beginning of the data collection and then moves at a more or less constant velocity until the end of the data collection, indicated in the range-doppler image shown in figure 11. Even though the real part of the FFT for both cases of target motion looks similar, figures 8 and 10, the complex FFT is able to distinguish differences between receding and approaching targets, figures 9 and 11, as expected. Figure 5. Range-Doppler image of stationary target.

7 Figure 6. Real part of Hanning-weighted FFT of Range cell 10 for target moving back and forth Figure 7. Range-Doppler image of target moving back and forth.

8 Figure 8. Real part of the Hanning-weighted FFT of the 5 th range cell for approaching target. Figure 9. Range-Doppler image of approaching target.

9 Figure 10. Real part of the Hanning-weighted FFT of the 5 th range cell for receding target. Figure 11. Range-Doppler image of receding target. V. CONCLUSION Translational Doppler detection using a direct-detect, chirped amplitude-modulated LADAR has been demonstrated. In this application, a non-optically coherent system is utilized to lower the target-induced Doppler shift to under a khz as opposed to several hundreds of megahertz for an optically coherent LADAR design. We have shown that the Doppler

10 shift is imposed on both the laser carrier and on the intensity modulation impressed on the laser carrier. Data indicates that the LADAR is able to measure the Doppler shift due to translational motion of a moving target. REFERENCES 1. B. L. Stann, W. C. Ruff, Z. G. Sztankay, Intensity-modulated diode laser radar using frequency modulation/continuous wave ranging techniques, Optical Engineering, vol. 35, no. 11, pp , W. Ruff, J. Bruno, S. Kennerly, K. Ritter, P. Shen, B. Stann, M. Stead, G. Sztankay, M. Tobin, Self-mixing detector candidates for an FM/cw LADAR architecture, Proc. SPIE, vol. 4035, pp , B. L. Stann, et al., Component research for the MDSS LADAR, Proc. of Advanced Sensors Consortium, pp , P. H. Shen, et al., Interdigitated finger semiconductor photodetector for optoelectronic mixing, Proc. SPIE, vol. 4028, pp , Eaves, Jerry L., Reedy, Edward K., Principles of Modern Radar, New York: Van Nostrand Reinhold Company, Brigham, E. Oran, The Fast Fourier Transform and Its Applications, New Jersey: Prentice Hall, W. Ruff, K. Aliberti, J. Damman, M. Giza, P. Shen, B. Stann, Performance of an FM/cw prototype ladar using a 32-element linear self-mixing detector array, Proc. SPIE vol. 5086, pp 58-69, K. Aliberti, H. Shen, W. Ruff, B. Stann, P. Newman, F. Semendy, M. Stead, R. Mehandru and F. Ren, Improved optoelectronic mixing of InAlAs/InGaAs interdigitated-finger metal-semiconductor-metal photodetectors, Proc. of IEEE International Conference on Sensors, pp , 2003.