We develop an unprecedented 3D pulsed chaos lidar system for potential intelligent machinery applications. Benefited from the random nature of the chaos, conventional CW chaos lidars already possess excellent anti-jamming and anti-interference capabilities and have no range ambiguity. In our system, we further employ self-homodyning and time gating to generate a pulsed homodyned chaos to boost the energy-utilization efficiency. Compared to the original chaos, we show that the pulsed homodyned chaos improves the detection SNR by more than 20 dB. With a sampling rate of just 1.25 GS/s that has a native sampling spacing of 12 cm, we successfully achieve millimeter-level accuracy and precision in ranging. Compared with two commercial lidars tested side-by-side, namely the pulsed Spectroscan and the random-modulation continuous-wave Lidar-lite, the pulsed chaos lidar that is in compliance with the class-1 eye-safe regulation shows significantly better precision and a much longer detection range up to 100 m. Moreover, by employing a 2-axis MEMS mirror for active laser scanning, we also demonstrate real-time 3D imaging with errors of less than 4 mm in depth.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Along with the booming development of intelligent technologies, lidars play important and crucial roles in applications such as autonomous car, drone, robot, AR/VR, and surveillance [1–3]. By perceiving the surrounding environment with their superior detection range, accuracy, spatial resolution, and less constraints to weather and illumination conditions, lidars are promising for functions including ranging, 3D imaging, object tracking, recognition, and simultaneous localization and mapping.
Most lidars utilize pulsed lasers as their light sources to gain better signal-to-noise ratios (SNRs) and detection ranges, where repetitive optical pulses are emitted and the time-of-flights of each pulse reflected back from the targets are subsequently measured . With the advancements in semiconductor lasers and well-developed optoelectronic devices, a handful of pulsed lidar sensors have been commercialized in recent years for the emerging intelligent applications . However, since conventional pulsed lidars emit unspecific repetitive pulses, the periodicity and regularity of the pulses can often cause range ambiguity and make them vulnerable to interference and jamming [6,7]. With more lidar usages expected in the near future in a technology-crowded environment, reception of unspecific signals from other lidars and light sources will become inevitable and which could produce ghost images or result in failure of detections.
To mitigate the possibility of interferences, random-modulation continuous-wave (RM-CW) lidars utilizing randomly modulated lights with waveform specificities have been studied . In a RM-CW lidar, often a pseudo-random binary sequence (PRBS) is modulated on a CW laser through an external intensity modulator. By computing the cross-correlation between the received signal backscattered from the target and a transmitted replica, the range of the target can be detected while unwanted signals from other sources only contribute to noise. With the waveform specificity, the RM-CW lidars have been applied to applications including surface profiling, CO2 sensing, bio-aerosol detection, and atmospheric probing [9–11]. In a conventional RM-CW lidar, the unambiguous range is determined by the time duration of the PRBS pre-designed. The range resolution on the other hand is strongly associated with the bit width of each bit in the PRBS, which in practical is determined by the electronics and the modulation speed of the modulator used. Transmitting a recirculating PRBS that has a finite time duration and bits, nevertheless, could make a RM-CW lidar still vulnerable to jamming in which a malicious jammer can easily record the PRBS being emitted and then re-emits it to produce fake echoes.
By contrast, chaos lidars eliminate the shortcomings of both possible interference and jamming in the pulsed and RM-CW lidars by taking the advantages of the aperiodic and unpredictable nature of optical chaos [12–15]. Moreover, optical chaos can easily be generated by nonlinear dynamics of semiconductor lasers through optical injection and/or feedback without any expensive high-speed signal generator or external modulator [16–19]. In 2004, for the first time the authors analyzed the feasibilities of a CW chaos lidar and demonstrated the proof-of-concept experiments . Later with the similar concept but transmitting in optical fibers, researchers also realized optical time-domain reflectometers (OTDR) using chaos for fiber fault detection [20–22].
While CW chaos lidars are superior in their anti-interference and anti-jamming capability, however, the energy-utilization efficiencies in those previous studies were relatively low. The reasons are mainly twofold, firstly the power densities of optical chaos typically distribute unevenly in their RF spectra and tend to peak at around the relaxation oscillation frequencies of the lasers used. Only a small portion of the energy in the low frequency region close to DC can be converted for use due to the limited bandwidths of commercial off-the-shelf avalanched photodetectors (APD) and analog-to-digital converters (ADC) . Secondly, since the acquisition length is finite for an ADC, transmitting chaos light in its CW form not only wastefully delivers energy within the blind time of the ADC but also reduces the maximum power of emission permitted by the eye-safe regulation .
As the consequence, most chaos lidars reported up until now have been predominantly applied to low-loss scenarios with either short-range, high-reflectivity targets with specular reflections (such as mirrors or retroreflectors) in free-space or longer-range, but fiber-based OTDR applications [12,20–22]. Although in one study a high-power (500 mW average power) CW chaos laser was used to increase the detection range , such scaling up however is not practical for daily life consumer applications since its power is already order of magnitudes beyond the class-1 eye-safe regulation (less than 10 mW average power at 1.5 μm) .
In this study, we develop an unprecedented 3D pulsed chaos lidar system to effectively increase the energy-utilization efficiency and therefore the corresponding SNR. With its high efficiency, this system is practical for long range, diffusely reflecting object detection. We first employ homodyne interference to boost the power density of chaos at low frequency region. By self-homodyning optical chaos with a proper optical path difference longer than its coherence length, modulation depth and power density in the low frequency region can be significantly enhanced resulting from the proper constructive and destructive interferences. We then gate CW optical chaos in time into chaos modulated pulse train, namely pulsed chaos, with each pulse having a duration shorter than the acquisition length of the ADC and synchronizing with the ADC acquisition cycles. Without wasting any energy during the ADC blind time, pulsed chaos can be emitted with much higher peak power under the constraint of the class-1 eye-safe regulation. Moreover, to make it capable of acquiring and constructing 3D images of targets in real-time, we integrate a 2-axis micro-electro-mechanical-system (MEMS) scanning mirror as an active laser scanner to our 3D pulsed chaos lidar system. Further with the aid of an object recognition algorithm, which will be reported and demonstrated separately, the system is also capable of object recognition and tracking.
Compared to the chaos lidar reported previously in our proof-of-concept experiments that was only capable of ranging short-range, high-reflectivity targets with specular reflections under non-realtime post processing, the 3D pulsed chaos lidar system reported here advances its capability to be able to detect common objects with diffuse reflections at standoff distances under the class-1 eye-safe regulation, to acquire 3D data continuously in real-time, and to track and scan targets automatically. In this paper, we characterize the chaos light source of the 3D pulsed chaos lidar system, analyze its detection performance including its SNR, accuracy, and precision in ranging, and demonstrate its capabilities in long-range detection and real-time 3D imaging.
2. Setup of 3D pulsed chaos lidar system
Figure 1 shows the schematic setup of the 3D pulsed chaos lidar system. The system mainly consists of chaos light source, transmitting and receiving optics, and signal acquisition and processing modules. The chaos light source module has three sections to generate the high-efficient pulsed homodyned chaos light. In the optical feedback section, the output of a 1.55 μm single-mode semiconductor laser (Shengshi Optical SBF-D55W2-111PMS) is coupled directly to a polarization-maintaining fiber coupler. The light in one arm is reflected by a fiber mirror and fed back into the laser cavity. Optical chaos, namely the original chaos, can be generated by properly adjusting the feedback strength (defined as the ratio between the reflected and emitted optical fields) and delay time from the mirror to the laser. Passing through an isolator to avoid any unwanted feedback, the original chaos outputted to the other arm is then launched into the homodyne interference section.
In the homodyne interference section, we use a fiber-based Michelson interferometer to self-homodyne the original chaos. In order to produce homodyned chaos with maximal modulation depths, we set the power splitting ratio close to 50:50 and the optical path difference to be longer than the coherence length of the original chaos. A small portion of the homodyned chaos is detected by a 400 MHz APD (Thorlabs APD430C) to serve as the reference signal while the remaining is sent to the time gating section to further enhance the signal efficiency.
In the time gating section, we amplify the homodyned chaos by an erbium-doped fiber amplifier (EDFA)(GIP Technology CGBH1Bs128001A) and then gate it by an acousto-optic modulator to generate the pulsed homodyned chaos. By controlling the pulse repetition frequency (PRF) and the duty cycle to have a pulsewidth shorter than the acquisition length of the ADC, the energy in the pulsed homodyned chaos can be effectively utilized without wasting it during the ADC acquisition blind time. The function generator (Agilent 81150a) and the variable optical attenuator are simultaneously adjusted so that the respective duty cycle and peak power of the pulses are balanced in order to keep the average power under the 10 mW limit set by the class-1 eye-safe regulation at the wavelength of 1.5 μm .
In the transmitting and receiving optics module, the pulsed homodyned chaos is coupled out to the free-space via a collimator and then impinged on a 2-axis MEMS mirror (Mirrorcle S6249). The transmitted laser beam, with a spot size of about 1.6 mm and a divergence angle of about 1.25 mrad, can then laterally scan across a target or a scene for 3D imaging by controlling the tilt angle of the MEMS mirror in both the azimuth and the elevation directions. We use an IR camera lens (Opto Engineering SW05020) and an APD (identical to the one used in detecting the reference signal) to collect and detect the light backscattered from the target, namely the probe signal.
In the signal acquisition and processing module, we use an oscilloscope (Tektronix 7054C) as the ADC to simultaneously acquire the reference and the probe signals. A personal computer then calculates the cross-correlation between the probe and the reference signals and determines the target range from the lag time of the correlation peak. Depending on the PRF and duty cycle, we set the correlation length to be the same as the pulsewidth to max out the energy utilization. To accommodate the undergoing development of a portable, low cost, integrated 3D pulsed chaos lidar system, we limit the sampling rate of the oscilloscope to 1.25 GS/s in this study to match with the specs of those commercial cost-effective FPGA-integrated ADC boards. With a sampling rate of 1.25 GS/s, the corresponding native sampling spacing is 12 cm. To improve the accuracy and precision of the system, we further incorporate the Spline interpolation  to extract the correlation peaks in detail.
To successfully obtain 3D images in real-time, the chaos light source, transmitting and receiving optics, and signal acquisition and processing modules have to be cooperatively synchronized. Here we pre-design and upload the scanning pattern to the MEMS controller to actuate the MEMS mirror in a continuous manner. At each position during scanning, the MEMS controller sends triggers to both the function generator and the oscilloscope to start the pulse transmission and the data acquisition, respectively. After the completion of each scan, a 3D image of the target represented by its point cloud can be obtained.
3. Specification of chaos light source
To generate optical chaos, we set the feedback strength and delay time in the optical feedback section to be 0.3 and 65 ns, respectively. Not shown in the lidar setup, we use a 12 GHz PIN photoreceiver (Newport 1544-B) to analyze the microwave characteristics of the chaos generated. Figures 2(a)–2(c) show the optical spectrum, time series, and microwave spectrum of the original chaos generated. As can be seen, the original chaos has a broad 3 dB bandwidth of 13.48 GHz (a corresponding coherence length of 2.2 cm) and an aperiodic waveform in time. Despite its broad bandwidth, the power of the original chaos is distributed non-uniformly and tends to peak higher around the relaxation oscillation frequency of the laser. The fractional power detectable by the 400 MHz APD in the low frequency region is largely insufficient for practical lidar applications.
Figures 3(a)–3(c) show the respective microwave spectra, time series, and autocorrelation functions of the homodyned (blue) and original chaos (black) detected by the same APD with the same input power for comparison. We set the optical path difference of the self-homodyne at 3 m, which is much longer than the coherence length of the original chaos for optimal power enhancement. As can be seen in Fig. 3(a), even with an input power reaching the saturation power of the APD, the microwave spectrum of the original chaos is still at a level not much elevated from the noise background (green curve). By contrast, self-homodyne enhances the microwave power of the homodyned chaos by 23 dB in average compared to that of the original chaos. The enhancement in power can also be observed in the time series as shown in Fig. 3(b), where the modulation depth is increased from 2% for the original chaos to 41% for the homodyned chaos. In addition, as shown in Fig. 3(c), the autocorrelation functions of the homodyned and original chaos almost overlap with each other. It indicates that the homodyned chaos well preserves the aperiodic feature of the original chaos in which only a significant peak is present at the zero-lag. Having a delta function-like autocorrelation trace, precise detection without range ambiguity are therefore expected.
To quantitatively compare the detection capability of the homodyned and original chaos in practice, we put a piece of standardized Kodak white card (90% reflectance, diffuse reflection) at 1.5 m away from the lidar system as the target. By transmitting both the homodyned and original chaos in their pulse mode with a duty cycle of 5%, the cross-correlations between the probe and reference signals are shown in Fig. 4(a). As can be seen, although for both cases the correlation peaks appear accurately at the corresponding target distance of 1.5 m, the correlation trace of the pulsed original chaos has relatively higher amplitude variations in the detection background (off the peak) due to its lower microwave power. By taking the peak amplitude in the correlation trace as the signal and three times the standard deviation of the amplitude variations in the background as the noise, Fig. 4(b) shows the SNRs of the pulsed homodyned and original chaos with different duty cycles under a PRF of 1 kHz. As can be seen, when having the same duty cycle, the SNRs of the pulsed homodyned chaos are in average 15 dB higher than that of the pulsed original chaos. Compared to the CW original chaos (100% duty cycle) originally used in the conventional CW chaos lidars, here we show a 20 dB enhancement in the SNR with the pulsed homodyned chaos when operated at a duty cycle of 5%.
Figures 4(c) and 4(d) show the amplitudes of the signal and noise of the correlation traces obtained with the pulsed homodyned chaos (Shom and Nhom) and the pulsed original chaos (Sori and Nori) at different duty cycles, respectively. As can be seen in Fig. 4(c), having the same amount of energy delivered within each pulse repetition interval (PRI) of 1 ms (i.e., PRF of 1 kHz), the Shom for the pulsed homodyned chaos remains almost constant at different duty cycles. Note that the Shom of the CW case is 3 dB lower due to the limited ADC acquisition length of 500 μs in which only 50% of the energy (500 μs over 1 ms) is effectively received. As for the noise, the Nhom increases as the duty cycle increases. This is because that more noise is accumulated in the correlation process when the pulsewidth and the corresponding correlation length are lengthened as the duty cycle increases. As for the pulsed original chaos shown in Fig. 4(d), the Sori is relatively weak such that it soon gets submerged by the Nori when the duty cycle is larger than 40%.
4. Analyses of ranging performance
We further analyze the ranging performance of the chaos lidar at different ranges. We optimize the transmitting and receiving optics to efficiently collect the backscattered light within the 6 m range in the laboratory. Figure 5(a) shows the SNRs of the chaos lidar in ranging with the pulsed homodyned chaos, CW homodyned chaos, and pulsed original chaos at different ranges, respectively. Both the pulsed homodyned and pulsed original chaos have a duty cycle of 5%. We omit the result from the CW original chaos due to its low efficiency and poor detection capability. As can be seen, determined mainly by the configuration of the transmitting and receiving optics and thus the amount of the power collected, the SNRs from the pulsed homodyned chaos, CW homodyned chaos, and pulsed original chaos show similar variation trends as the range varies. The drop before 1 m is attributed to the large interception angle between the backscattered light and the field-of-view (FOV) of the receiver . As can be seen, for all the ranges examined, the pulsed homodyned chaos has the highest SNRs benefited by its higher energy-utilizing efficiency. Even in its CW mode, the SNRs from the CW homodyned chaos still well exceed that from the pulsed original chaos.
Figures 5(b) and 5(c) show their corresponding accuracies and precisions obtained at different ranges, respectively. As can be seen, the pulsed homodyned chaos has the best (lowest in values) accuracies and precisions at all ranges attributed to its higher SNRs, in which the shapes of the correlation traces are not altered much from the ideal delta function and the lag times of the correlation peaks are not jittered much by the excess noise. Note that, for the pulsed original chaos, we can only obtain the accuracy and precision at 1.5 m while elsewhere the SNRs are too low for a valid detection.
By replotting the data from Fig. 5, Figs. 6(a) and 6(b) show the accuracies and precisions of the chaos lidar in ranging with the pulsed homodyned chaos, CW homodyned chaos, and pulsed original chaos at different SNRs, respectively. The black dashed curves show their corresponding curve fittings. As can be seen, for all the chaos examined, both the accuracy and precision are highly correlated with the SNR where a higher SNR leads to a better (lower in values) accuracy and precision. For the pulsed homodyned chaos with an SNR of 19.8 dB, very high accuracy and precision of 1.1 mm and 0.5 mm are successfully achieved, respectively. Based on the curve fitting, with SNRs higher than 3.8 dB, the chaos lidar can achieve a 2 cm accuracy benchmark set by those high-end commercial products (e.g., Velodyne HDL-64E). For SNRs higher than 6.4 dB, millimeter-level accuracies and precisions can both be achieved at the same time.
5. Demonstrations of long range detection
To demonstrate its long range detection capability, we move the chaos lidar outside of the laboratory and test it on a 100 m hallway with a clear line of sight. To have the best performance, we configure the so-called pulsed chaos lidar adopting the pulsed homodyned chaos with a duty cycle of 5% and use this configuration in all of the following demonstrations. Two commercial lidar products, one being a pulsed lidar (Spectroscan from Spectrolab ) and one being a RM-CW lidar (Lidar-lite from Pulsedlight ), are also tested side-by-side with the pulsed chaos lidar for comparison. Here we use the same piece of the standardized Kodak white card as the target. Due to the lack of an accurate long-range distance standard, we only measure and compare their precisions and omit the accuracies in this experiment.
Figures 7(a) and 7(b) show the SNRs and precisions of the pulsed chaos lidar (blue) for ranges up to 100 m. The corresponding precisions for the Spectroscan (red) and Lidar-lite (green) up to 40 m are shown in Fig. 7(b). As can be seen, the SNR falls as the range increases, which matches well with the curve fitting (dashed curve) obtained with the inverse-square law described by the lidar equation [8, 27]. Here we only demonstrate a detection range up to 100 m limited by the length of the hallway experimented. Based on the extrapolating from the intersection of the inverse-square fitting curve and the 3.8 dB SNR baseline (dotted line) set by the 2 cm accuracy benchmark obtained from Fig. 6(a), we expect the pulsed chaos lidar under the current configuration can detect targets up to 136 m with an accuracy better than 2 cm.
In Fig. 7(b), we also show that the pulsed chaos lidar has a sub-centimeter precision up to about 100 m. Compared with the two commercial lidars tested that can only provide valid detections up to 40 m with precisions of several centimeters, the pulsed chaos lidar clearly shows its potential in commercialization for practical applications. Besides, the pulsed chaos lidar also uniquely possesses the features such as anti-jamming, anti-interference, and unambiguous detection that are not commonly available in the commercial lidars.
6. Demonstrations of 3D imaging
To demonstrate the 3D imaging capability of the pulsed chaos lidar, we first use a curvy mask as the target and acquire its 3D image by translating it laterally across the fixed laser beam. Figures 8(a)–8(c) show the photograph of the curvy mask, the 3D image, and the corresponding 3D reconstruction obtained by the pulsed chaos lidar, respectively. As can be seen, benefited from the high accuracy and precision, we obtain the 3D image of the target in detail where fine features in the eyes, nose, and lips are clearly revealed.
To further automate the target scanning, we employ a 2-axis MEMS mirror as described in the setup for active laser scanning. Figures 9(a) and 9(b) show the front- and side-view photographs of a self-made “NTHU” letter pattern and the corresponding 3D image obtained, respectively. The letters all have similar dimensions but are intentionally placed at different heights (i.e., different ranges relative to the pulsed chaos lidar). As shown in Fig. 9(b), the shapes, sizes, and depths of each letter are clearly depicted and acquired, where the maximum error in depth is less than 4 mm. Currently, the FOV of the 3D pulsed chaos lidar system is about 34 mrad. It is mainly determined and limited by the small active area of the APD used [27,30]. To expand the FOV, we will develop and incorporate further techniques to increase the equivalent active area of the detector in the future [31,32].
In this study, we develop an unprecedented 3D pulsed chaos lidar system that is capable of detecting objects with diffuse reflections at long ranges without range ambiguity, can acquire 3D images in real-time, is immune to interference and jamming, and is in compliance with the class-1 eye-safe regulation for potential practical applications. For advancement, we incorporate self-homodyning and time gating in the chaos light source that significantly enhance the energy-utilization efficiency. Compared to the CW original chaos typically used in conventional chaos lidars, we show that the pulsed homodyned chaos improves the detection SNR by more than 20 dB. Under the current configuration, even with a sampling rate of just 1.25 GS/s and a corresponding native sampling spacing of 12 cm, we successfully achieve millimeter-level accuracies and precisions with the pulsed chaos lidar for SNRs higher than 6.4 dB. Compared with two commercial lidars, namely the pulsed Spectroscan and the RM-CW Lidar-lite, the pulsed chaos lidar shows significantly better precision and a much longer detection range up to 100 m. By employing a 2-axis MEMS mirror for active laser scanning, we also demonstrate real-time 3D imaging with less than 4 mm error in depth. The lidar system is fully automated where all the modules are integrated and cooperatively synchronized. With its competitive performance and the unique anti-interference and anti-jamming capability, we show the promising potentials of this 3D pulsed chaos lidar system in future intelligent technologies.
Ministry of Science and Technology, Taiwan (MOST 103-2112-M-007-019-MY3 and 106-2112-M-007-003-MY3); National Tsing Hua University, Taiwan (106N539CE1).
The authors would like to thank Electronic and Optoelectronic System Research Lab, Industrial Technology Research Institute of Taiwan for sharing their equipment and space for the field test.
References and links
1. B. Schwarz, “LIDAR: Mapping the world in 3D,” Nat. Photonics 4, 429–430 (2010). [CrossRef]
2. Q. Li, L. Chen, M. Li, S. L. Shaw, and A. Nuchter, “A sensor-fusion drivable-region and lane-detection system for autonomous vehicle navigation in challenging road scenarios,” IEEE Trans. Veh. Technol. 63(2), 540–555 (2014). [CrossRef]
3. J. McCormack, J. Prine, and B. Trowbridge, “2D LIDAR as a distributed interaction tool for virtual and augmented reality video games,” in Proceedings of IEEE Games Entertainment Media Conference (IEEE, 2015), 15700274.
4. M. C. Amann, T. Bosch, M. Lescure, R. Myllyla, and M. Rioux, “Laser ranging: a critical review of usual techniques for distance measurement,” Opt. Eng. 40(1), 10–19 (2001). [CrossRef]
5. Velodyne LiDAR, “HDL-64E data sheet,” http://velodynelidar.com/hdl-64e.html.
6. G. Kim, J. Eom, and Y. Park, “Investigation on the occurrence of mutual interference between pulsed terrestrial LIDAR scanners,” in Proceedings of IEEE Intelligent Vehicles Symposium (IV) (IEEE, 2015), 437–442.
7. J. Petit, B. Stottelaar, M. Feiri, and F. Kargl, “Remote attacks on automated vehicles sensors: experiments on camera and lidar,” in Black Hat Europe, (2015).
9. R. Matthey and V. Mitev, “Pseudo-random noise-continuous-wave laser radar for surface and cloud measurements,” Opt. Laser Eng. 43(35), 557–571 (2005). [CrossRef]
10. Z. Yang, C. Li, M. Yu, F. Chen, and T. Wu, “Compact 405-nm random-modulation continuous wave lidar for standoff biological warfare detection,” J. Appl. Remote Sens. 9(1), 096042 (2015). [CrossRef]
11. M. Quatrevalet, X. Ai, A. Pérez-Serrano, P. Adamiec, J. Barbero, A. Fix, J. M. G. Tijero, I. Esquivias, J. G. Rarity, and G. Ehret, “Atmospheric CO2 sensing with a random modulation continuous wave integrated path differential absorption lidar,” IEEE J. Sel. Top. Quantum Electron. 23(2), 5300311 (2017). [CrossRef]
12. F. Y. Lin and J. M. Liu, “Chaotic lidar,” IEEE J. Sel. Top. Quantum Electron. 10(5), 991–997 (2004). [CrossRef]
14. C. H. Cheng, Y. C. Chen, and F. Y. Lin, “Generation of uncorrelated multi-channel chaos by electrical heterodyning for multiple-input-multiple-output chaos radar (MIMO CRADAR) application,” IEEE Photon. J. 8(1), 7800209 (2016). [CrossRef]
15. F. Y. Lin and J. M. Liu, “Chaotic radar using nonlinear laser dynamics,” IEEE J. Quantum Electron. 40(6), 815–820 (2004). [CrossRef]
16. F. Y. Lin, Y. K. Chao, and T. C. Wu, “Effective bandwidths of broadband chaotic signals,” IEEE J. Quantum Electron. 48(8), 1010–1014 (2012). [CrossRef]
17. J. Mork, B. Tromborg, and J. Mark, “Chaos in semiconductor lasers with optical feedback: theory and experiment,” IEEE J. Quantum Electron. 28(1), 93–108 (1992). [CrossRef]
18. Y. H. Liao and F. Y. Lin, “Dynamical characteristics and their applications of semiconductor lasers subject to both optical injection and optical feedback,” Opt. Express 21(20), 23568–23578 (2013). [CrossRef] [PubMed]
19. A. Uchida, T. Heil, Y. Liu, P. Davis, and T. Aida, “High-frequency broad-band signal generation using a semiconductor laser with a chaotic optical injection,” IEEE J. Quantum Electron. 39(11), 1462–1467 (2003). [CrossRef]
20. Y. Wang, B. Wang, and A. Wang, “Chaotic correlation optical time domain reflectometer utilizing laser diode,” IEEE Photon. Technol. Lett. 20(19), 1636–1638 (2008). [CrossRef]
21. A. Wang, N. Wang, Y. Yang, B. Wang, M. Zhang, and Y. Wang, “Precise fault location in WDM-PON by utilizing wavelength tunable chaotic laser,” J. Lightwave Technol. 30(21), 3420–3426 (2012). [CrossRef]
22. X. Dou, H. Yin, H. Yue, Y. Jin, J. Shen, and L. Li, “Experimental demonstration of the real-time online fault monitoring technique for chaos-based passive optical networks,” Opt. Commun. 350, 288–295 (2015). [CrossRef]
24. International Electrotechnical Commission (IEC), “Safety of laser products. Part 1: Equipment classification, requirements and user’s guide,” Standard IEC 60825-1:2001 (International Electrotechnical Commission, Geneva, 2001).
25. T. Zhao, B. Wang, Y. Wang, and X. Chang, “Free space ranging utilizing chaotic light,” Mathematical Problems in Engineering 2013, 172728 (2013). [CrossRef]
26. C. de Boor, A practical guide to spline (Springer, 1978). [CrossRef]
27. K. Stelmaszczyk, M. Dell’Aglio, S. Chudzyński, T. Stacewicz, and L. Wöste, “Analytical function for lidar geometrical compression form-factor calculations,” Appl. Opt. 44(7), 1323–1331 (2005). [CrossRef] [PubMed]
28. Spectrolab, “SpectroScan 3D MEMS lidar system,” http://www.spectrolab.com/sensors/index.html.
29. Pulsedlight, “LIDAR-Lite,” https://www.pulsedlight3d.com/.
30. G. G. Gimmestad and D. W. Roberts, “1.5 microns: the future of unattended aerosol lidar?” in Proceedings of IEEE International Geoscience and Remote Sensing Symposium (IEEE, 2004), 1944–1946.
31. X. Lee, X. Wang, T. Cui, C. Wang, Y. Li, H. Li, and Q. Wang, “Increasing the effective aperture of a detector and enlarging the receiving field of view in a 3D imaging lidar system through hexagonal prism beam splitting,” Opt. Express 24(14), 15222–15231 (2016). [CrossRef] [PubMed]
32. H. Tsuji, M. Imaki, N. Kotake, A. Hirai, M. Nakaji, and S. Kameyama, “Range imaging pulsed laser sensor with two-dimensional scanning of transmitted beam and scanless receiver using high-aspect avalanche photodiode array for eye-safe wavelength,” Opt. Eng. 56(3), 031216 (2016). [CrossRef]