Patent Grant
12038511
A portion of the disclosure of this patent document contains material which is subject to copyright protection. This patent document may show and/or describe matter which is or may become trade dress of the owner. The copyright and trade dress owner has no objection to the facsimile reproduction by anyone of the patent disclosure as it appears in the Patent and Trademark Office patent files or records, but otherwise reserves all copyright and trade dress rights whatsoever.
This disclosure relates to three-dimensional (3D) image systems and, specifically, to coherent Light Detection and Ranging (LIDAR) imaging systems.
3D imaging systems are critical for a number of applications in the automotive industry, robotics, unmanned vehicles etc. Traditionally, pulsed LIDAR systems relying on the time of flight (TOF) technique have been investigated for these applications, wherein the range to a target is determined by measuring the time taken by a single narrow pulse to be reflected from a distant target. TOF LIDAR suffers from some major challenges and limitations. Since the amount of light reflected from distant targets is very small (typically under 100 photons in the measurement time) and the pulse width needs to be very small (<1 ns) to achieve high range accuracy, these systems require sophisticated detectors, typically high-speed photon-counters. The TOF technique places very stringent requirements on the dynamic range of the detector and associated electronics (typically about 60 dB, or 20 bits). Further, these very sensitive detectors often have difficulty dealing with crosstalk from other LIDAR systems or from other sources of light (including direct sunlight) when operated in real-world situations.
Coherent LiDAR is a promising 3D imaging technology because of its potential to achieve excellent performance on a number of key metrics: high-speed (>1 million points/second), long-range (>200 m for targets with albedo of 0.1), high lateral resolution (<0.1 degrees), and fine range precision (<10 cm).
The output of the CHDL 110 is divided into two components by a tap coupler 120. A small fraction is separated from the output to be used as a Local Oscillator (LO) wave 125. The majority (typically >90%) of the CHDL output power is directed to a target 160 via a circulator 130. Typically, an optical system (not shown), such as a telescope, is used to form the CHDL output into a collimated output beam 135 that illuminates the target 160. Light reflected from the target 140 is collected by the same optical system (not shown) and returned to the circulator 130. The reflected light exits the circulator 130 and is combined with the local oscillator wave 125 in a 2×1 coupler 145. The combined LO wave and the reflected light from the target are incident on a photodetector (PD) 150. The photodetector 150 provides an output current proportional to the incident optical power. The photodetector 150 effectively multiplies the amplitudes of the reflected light and the L wave to create a coherent “beat signal” whose frequency is directly proportional to the round-trip time delay to the target, and the range to the target is thus determined.
While
ω=ω0+ξt, (1)
The optical frequency of the reflected wave from a target follows a similar function, but is offset in time from the output wave by a period given by
τ=2R/c (2)
i≈(PLOPRef)1/2 cos(Δωt+ω0τ) (4)
where i=current output from the photodetector; and
A coherent LIDAR system can determine both a range to a target and a rate of change of the range (range rate) to the target.
ωD=2π/λ(dR/dt) (7)
When a target is illuminated with a laser beam with an up-chirp (i.e., a beam that increases in frequency with time), the Doppler shift and the frequency shift due to the delay of the reflected beam are additive, such that
Δω+=ωD+ξτ; (8)
Throughout this description, elements appearing in figures are assigned three-digit or four-digit reference designators, where the two least significant digits are specific to the element and the most significant digit or digits provide the number of the figure where the element is introduced. An element that is not described in conjunction with a figure may be presumed to have the same characteristics and function as a previously-described element having the same reference designator.
Description of Apparatus
The key requirement for a coherent chirped LIDAR system is a laser whose optical frequency varies with time in a precisely controlled fashion. LIDAR systems commonly incorporate semiconductor diode lasers and attempt to control the laser to produce a precisely linear chirped wave. However, the principles described in this patent can be more generally applied to any type of laser whose output frequency can be varied by changing one or more input parameters. These principles can also be applied to generating nonlinear chirps.
Feedback-controlled chirped diode lasers measure the frequency output characteristic of the laser and use the measurement to provide closed-loop feedback to control the laser output frequency. However, measuring and controlling the rate of change of the laser output frequency typically requires a finite time interval. For example, a fraction of the laser output power may be transmitted through an unbalanced (or asymmetric) Mach-Zehnder interferometer (MZI) and onto a photodetector. The output frequency of the beat signal produced by the MZI—photodetector is directly proportional to the slope of the frequency chirp (i.e., the rate of change of frequency with time). By comparing this beat frequency to a desired beat frequency (corresponding to a desired frequency chirp rate), an error signal can be generated and fed back to the laser (typically after additional filtering). This closed-loop system can generate a precisely controlled linear chirp, but only works well if the chirp duration and the LIDAR measurement time are substantially longer that the time interval needed to measure and control the slope of the frequency chirp.
The chirp rate of high-speed coherent LIDAR systems is dictated by the required resolution and image update rate. Some systems may use chirp durations or measurement times less than 1 microsecond. Closed-loop laser control, as described in the preceding paragraph, does not work for these high speed LIDAR systems, because the propagation delays in the feedback system (such as optical delays in the unbalanced MZI and the response time of the laser and any loop filters) are comparable to or larger than the chirp duration itself.
The transmitter output in a coherent LIDAR system is a sequence of identical (at least in theory) frequency chirps that repeat periodically, as shown in
A portion of the output beam from the laser 410 is extracted by an optical tap 420 and applied to an optical frequency discriminator 425 that provides a measurement of the rate of change of the output frequency of the laser output. The optical tap 420 may be a tap coupler as shown, an asymmetric beam splitter, or some other device that extracts a small portion of the laser output. The optical frequency discriminator may be, for example, an asymmetric MZI and photodetector as shown in
An error determination module 430 receives the output from the optical frequency discriminator 425, and determines the deviation of the laser frequency from its intended value as a function of time during the chirp period. The error determination may be performed by hardware and/or by a processor executing a method, such as a Hilbert transform, implemented in software. The error determination for the present chirp (“chirp k”) is provided to a correction determination module 435 that determines a correction to be applied to the drive signal for a subsequent chirp k+1 or, more generally, a future chirp period, where the delay between the measurements and the future time period is less than the time constant of any environmental fluctuations. The correction module 435 may determine the correction to be applied to the future chirp using hardware and/or a processor executing a method implemented in software. The correction determination module 435 may determine the correction to be applied to the future chirp based upon the error determination and/or the correction for one or more prior chirps. For example, the correction determination module 435 may determine the correction to be applied to the future chirp based, at least in part, on a weighted sum or weighted average of the determined errors for two or more prior periods of the periodic frequency versus time function. The correction determination module 435 may determine the correction to be applied to the future chirp based, at least in part, on digital or analog filtering of the determined errors for one or more prior periods of the periodic frequency versus time function.
The laser 400 will be subsequently refer to as a “state machine CHDL” because where the information about the current “state” of the chirp is fed back to influence the future state of the chirp.
Although not shown in
As illustrated in
It is desirable that the 3D-IR be as high as possible. For the standard system of
3D-IR<½τmax=c/4Rmax. (8)
This is a limitation imposed by the finite speed of light c. For example, for a maximum range of 300 m, the 3D imaging rate is limited to 0.25 million pixels per second. This limitation may not be acceptable in some applications.
The “speed of light limit” on the 3D imaging rate can be overcome with two improvements to the basic coherent LIDAR system of
In the LIDAR system 500, elements with reference designators from 510 to 560 have the same function as the counterpart elements in the LIDAR system 100 and will not be further described. The LIDAR system 500 includes second CHDL 570, which is chirped in the opposite direction as the first CHDL 510. The first and second CHDLs 510, 570 may be multiplexed by a beam combiner 575. For example, the first and second CHDLs 510, 570 may have orthogonal polarization, and the beam combiner 575 may be a polarization beam splitter. The first and second CHDLs 510, 570 may have different wavelengths, and the beam combiner 575 may be a dichroic beam splitter. In any case, the beams from the first and second CHDLs 510, 570 are combined and directed to the target 560. The reflected beams from the target are separated by beam divider (using polarization or wavelength as appropriate) are directed to separate detectors 550, 585. With this approach, the time to obtain R and dR/dt for a pixel is T, and the 3D imaging rate is (1/T) pixels per second. Again, by minimizing the measurement time TM, the limitation on 3D-IR is 0.5 million pixels per second for a maximum range of 300 meters. A 3D-IR of 0.5 million pixels per second, while twice the rate of conventional coherent LIDAR systems, may still not be sufficient for some application.
The next improvement in the conventional coherent LIDAR system results from recognition that the speed of light fundamentally imposes a delay or latency in the measurement, rather than a restriction on the imaging rate. The improvement to overcome the speed of light limit is to not reset the chirp after every pixel, but instead to use a chirp with longer extent T+ that spans a number of (scanned) pixels as shown in
As shown in
LIDAR systems impose a stringent requirement on the number of photons that need to be collected by the coherent receiver in order to achieve accurate measurements of range and range-rate. The number of photons collected is determined by the transmitted laser power, the reflectivity of the target, and the size of the receiver collection optics. Long-range (i.e., longer than 100 meters) LIDAR systems benefit from the use of high output powers (e.g., 100 mW to 10 W) to minimize the size and complexity of the collection optics used in the coherent receiver. However, long-range coherent LIDAR systems also require very narrow laser line width, which is generally incompatible with high laser output power. Semiconductor lasers with narrow line widths typically have output powers less than 100 mW.
A semiconductor laser may be used in conjunction with a semiconductor optical amplifier (SOA) or flared tapered amplifier in order to achieve the desired higher output powers while maintaining the required narrow line width. The output of a narrow line width master oscillator, which may be a state machine CHDL, may be fed into an optical amplifier, typically in a semiconductor medium. However, the optical and spectral properties of the oscillator can be affected by optical feedback effects and coupling of amplified spontaneous emission (ASE) from the amplifier section to the oscillator, which can dramatically increase the line width. Thus, a feedback barrier may be disposed between the CHDL and the amplifier to ensure that the line width and other properties of the CHDL are not affected by feedback from the amplifier section. The feedback barrier may be, for example, an optical isolator or an attenuator. A master-oscillator power-amplifier (MOPA) laser with a broad-area or flared/tapered amplifier can provide single-mode operation at high (i.e., greater than 10 W) output power on a single integrated semiconductor chip.
With few exceptions, LIDAR systems can be segregated into staring systems and scanning systems. In a staring LIDAR, the transmitted laser beam illuminates the entire scene to be imaged. Reflections from the entire scene are imaged onto a detector array, with each detector element in the array corresponding to respective pixel in the scene. A staring coherent LIDAR must spread the reflected and LO beam over the entire detector array, which leads to insufficient signal-to-noise ratio unless the available laser power is very high. Thus, coherent LIDAR systems typically use an optical system to scan the transmitted bean sequentially over the scene.
A coherent receiver for a coherent LIDAR system employing a scanning transmitted beam has a fundamental architectural challenge. A transmitted beam having a beam diameter D0 is scanned across the scene within a wide field of view Θ. The scanning is typically performed in two dimensions. However, the subsequent figures only show scanning in one direction for ease of representation. The same design can be easily extended to two-dimensional scanning. The size of the transmitted beam, D0, is determined by the size of the scanning optic and the required angular resolution (typical values of D0 are 1-3 mm). The coherent receiver has to modally overlap the received photons from the target with the LO wave on a photodetector. One solution for the coherent receiver is to use an imaging lens that images the entire field of view Θ on a fixed detector or detector array, and illuminate the entire detector area, whether it is a single large detector or a detector array, with the LO at all times. However, since only a fraction of the field of view is imaged at any given time (this fraction can be 1/105 or smaller), this leads to an inefficient use of LO power and detector area, and can result in very poor signal to noise ratio due to LO shot noise. Thus the receiver in a typical coherent LIDAR is typically scanned along with the transmitted beam. In other words, the LO beam needs to be mode-matched with the return beam from the scene as different parts of the field of view are illuminated.
The effective collection aperture of the coherent receiver, D1, is dictated by the requirement to collect enough photons from the target to make a high-SNR measurement. With a sufficiently high-power laser, D1=D0, which is to say the transmitted beam diameter and the receiver collection aperture may be the same. In this case, the LIDAR optical system can be a simple “cat's-eye” configuration iso-called because the transmitted and reflected light propagate in opposing directions along the same optical path, as is the case with light reflected from the eyes of a cat) where the return beam from the target retraces the optical path of the transmitted beam, as in the coherent LIDAR system 800 shown in
In the LIDAR system 800, elements with reference designators from 810 to 855 have the same function as the counterpart elements in the LIDAR system 100 and will not be further described. The laser beam (other than the fraction extracted for the LO) is output from the circulator 830 and expanded/collimated by lens 870 to form an output beam having diameter D0. The lens 870, and all lenses in subsequent drawings, are depicted as single refractive lens elements but may be any combination of refractive, reflective, and/or diffractive elements to perform the required function. The output beam impinges upon scanning mirror 875, which can be rotated about an axis 880. The scanning mirror 875 may be, for example, a MEMS (micro-electro-mechanical system) mirror capable of high speed scanning. Rotation of the scanning mirror causes the output beam to scan across the scene through a field of view Θ.
Light reflected from the target impinges upon the scanning mirror 875 and is directed to the optic 870. Optic 870 captures (i.e., focuses) the reflected light, which is directed to the circulator 830. The captured reflected light exits the circulator 830 and is combined with the LO beam and detected as previously described. The diameter Do of the transmitted beam and the collection aperture of the receiver are defined by the optic 870. Increasing the diameter of the transmitted beam correspondingly increases the diameter of the receiver aperture (and thus the number of received photons) at the expense of increasing the size of the optic 870 and the scanning mirror 875. The requirement for high speed scanning over the field of view limits the allowable size of the mirror. The transmitted beam diameter of scanning LIDAR systems is typically limited to 1 mm to 3 mm.
In the absence of a very high-power laser, the necessary diameter of the receiver aperture D1 is preferably larger than the transmitted beam diameter D0, and the challenge is to ensure mode overlap between the received beam and the LO beam at the detector with sufficient signal-to-noise ratio.
To achieve coherent detection of the reflected light, it is necessary that the LO beam and the received light be superimposed at the detector. It is possible to achieve this with an LO beam that illuminates the entire detector 945. Since the received light forms a spot that scans across the detector 945 as the transmitted beams is scanned across the scene, it is advantageous for the LO beam to instead scan across the detector 945 in a corresponding manner. To this end, a portion of the transmitted beam is extracted by a tap beam splitter 920 to form the LO beam 925. The LO beam is then combined with the received light 935 by a second beam splitter 950. Since the LO beam 925 is extracted from the transmitted beam 930 as it is scanned, the angle of the LO beam changes in conjunction with the scanning of the transmitted beam. The tap beam splitter 920 and beam splitter 950 effectively form a corner reflector such that the LO beam is parallel to the received light when the LO and received beams are combined. Thus, the second lens 940 focuses the LO beam to a spot that superimposed on the spot of received light at the detector 945.
A third lens 1060 and a fourth lens 1065 form a 1:1 telescope that relays the transmitted beam 1030 from the scanning mirror 1015 to the scene. A portion of the transmitted beam is extracted by a tap beam splitter 1020 between the third and fourth lenses 1060, 1065 to form the LO beam 1025. The LO beam and the received light 1035 are combined by a second beam splitter 1050. A relay lens 1070 in the path of the LO beam focuses the LO beam to a spot at the plane of the detector 1045. When the focal lengths of the third lens 1060 and the further lens 1065 are equal to the focal length of the second lens 1040, the focused spot of the LO beam will have the same size as the focused spot of received light.
While
The number of optical modes, or unique angular positions, a beam can assume within a field of view is determined by the angular resolution of the beam. A beam with diameter D0 has an angular resolution ˜λ/D0 (ignoring constant scaling factors of order unity throughout this discussion). Therefore a beam with diameter D0 can be scanned to fill out a certain number of optical modes N0˜Θ*D0/λ within the field of view Θ. When a scanned beam of diameter D0 is optically transformed into a new beam with a diameter D1, the angular resolution of the new beam will be ˜λ/MD0, where M=D1/D0 is the magnification factor. However, the total number of available optical modes remains constant when the diameter of the beam is magnified. A simple telescope that magnifies the beam diameter from D0 to D1 will reduce the total field of view from Θ to Θ/M to conserve the number of optical modes.
After the beam diameter is magnified to D1, the scanner will provide N0 modes, with an angular resolution of λ/MD0, distributed over a field of view of Θ/M. To recover the original field of view, the N0 modes must be spaced apart with an angular distance between adjacent modes of λ/D0, such that the N0 modes are sparsely distributed across the field of view Θ. This means that only a portion of each scene resolution element is illuminated as the transmitted beam is scanned over the scene. Practically this means that the LIDAR system images the same total field of view in the same measurement time by trading off scene pixel fill factor for an increase in the received optical power.
A mode transformer may be used to transform a set of N0 closely spaced modes spanning a field of view Θ/M into a set of N0 modes that sparsely sample a field of view Θ. For example, a mode transformer can be implemented by coupling each available mode into a respective single mode optical fiber, and then moving the fibers apart from each other to “sparsify” the modes. A collimating lens can then be used to convert the light from each optical fiber into a D1-sized beam.
Another practical embodiment of a mode converter uses a microlens array.
When the focal lengths of the second lens 1120 and the third lens 1140 are the equal, the field of view Θ will be equal to the beam scan angle. When the focal lengths of the second and third lenses 1120, 1140 are unequal the field of view can be expanded or compressed compared to the beam scan angle.
The optical system 1100 of
Coherent LIDAR measures the amplitude and phase of the electromagnetic field reflected from the target. This reflected field is strongly influenced by surface irregularities on the target within the illuminated spot. These irregularities result in random variations in the amplitude and phase of the reflected field, commonly known as speckle. In most practical cases, the intensity of the reflected field has an exponential probability distribution and the phase has a uniform probability distribution. This means that even a bright target can occasionally have a low intensity and may be below the detection threshold of the LIDAR system. The spatial scale of the speckle variations is given by the resolution of the receiver optics. For example, in a conventional coherent LIDAR system, if the angular resolution of the LIDAR system is 0.1 degrees, the amplitude and phase of the speckle pattern change every 0.1 degrees.
The probabilistic nature of target reflections also occurs in RADAR systems, and techniques for RADAR detection of so-called “fluctuating targets” have been developed. These techniques rely on multiple RADAR measurements of the target to overcome the target strength (and phase) fluctuations. These measurements are then coherently or incoherently “integrated” to overcome the negative effects of target fluctuations. Integration refers to the process of combining multiple measurements to extract range and/or Doppler measurements, and can provide a higher probability of detection of the target for a given signal to noise ratio of the measurements. The process of incoherent integration ignores the optical phase of the reflected field, whereas coherent integration utilizes the phase information. The process of integration works best if the multiple measurements being integrated are in some way uncorrelated from each other, so that the target fluctuations can be “averaged” out. Speckle in a LIDAR system is comparable to a fluctuating target in a RADAR system, and similar mathematical techniques can be applied to mitigate the effects of speckle. However, the goal of a high-speed coherent LIDAR system is to obtain range and Doppler information from every scene pixel in a single scan of the field of view. Thus integration over multiple scans, as used in RADAR systems, cannot be directly applied to a LIDAR system.
The key to speckle mitigation in LIDAR systems is to obtain multiple measurements over each scene pixel during each scan of the field of view, and then coherently or incoherently integrate (combine) these measurements to mitigate target fluctuations (speckle).
In the first approach shown in
Alternatively, the LIDAR measurements on N subpixels within a pixel can be performed simultaneously and in parallel, as illustrated in
A hybrid of the techniques shown in
The output of a CHDL 1410 is divided into N parallel beams. A tap coupler 1420 is used to extract a small fraction of each beam as a respective LO beam. The majority (typically >90%) of each CHDL beam is directed toward a target via an N-channel circulator 1430. The N-channel circulator may be, for example, implemented using discrete optical elements (e.g. known combinations of polarizing beam splitters and a Faraday rotator) such that N beams can pass through the circulator in parallel. A lens 1470 converts the N output beams 1435 from the circulator 1430 into N collimated output beams at slightly different angles such that the N beams illuminate N subpixels within a target pixel, as shown in
The lens 1470 collects light reflected from the target (not shown) and forms the reflected light into N received light beams 1440. The received light beams separated from the output beams by the circulator 1430 and are combined with respective LO beams by N couplers or beamsplitters 1445. The combined LO waves and the received light beams from the target are respectively incident on N photodetectors 1450. The N photodetectors 1450 provide respective measurements indicative of the range and range rate of respective subpixels, as previously described. A processor (not shown) coherently or incoherently combines the N subpixel measurements to provide a composite range and/or range rate measurement for the pixel.
A first lens 1510 receives N beams from a circulator, such as the circulator 1430 in
The reflected light from these N subpixels is received by the third lens and propagates through the scanning optical system 1500 in the reverse direction, which results in N separate beams returned to the circulator, as was the case in the LIDAR system 1400 of
A high-resolution coherent LIDAR or imaging system has two important components: i) a swept-frequency laser or “chirped laser” with a large frequency sweep range B to provide high axial resolution at the range of interest; and ii) a technique to translate the one-pixel (fixed (x,y)) measurement laterally in two dimensions to obtain a full 3-D image.
Coherent LIDAR systems for 3-D imaging typically rely on the scanning of a one-pixel measurement across the scene to be image. Examples of scanning LIDAR systems were previously shown in
An alternative to scanning LIDAR systems is a staring system consisting of a swept frequency laser, and a detection technique that is capable of capturing or measuring the 3-D image of a scene, or a multi-pixel portion of a scene concurrently. Such a system has the potential to be inexpensive, robust, and contain no moving parts.
This single pixel detector described above is extended to an N-element array (one-dimensional or two-dimensional) to implement the full-field 3-D imaging system as shown in
The output of the detector arrays (whether K single detectors or 2K detectors to form K balanced pairs) is amplified with an array of TIAs 1665, digitized with an array of ADCs 1670, and one or more signal processors 1675 performs the Fourier transforms (typically using the FFT algorithm), the detection algorithms, and the input/output communication. The output of receiver array is typically two “images” (i.e., two values per pixel) corresponding to the depth map and the intensity of the reflections. Measuring the scene with alternate up- and down-chirps will also allow measuring the range rate of each scene pixel. Further data processing algorithms may also be implemented by the smart detector array or external processors.
The various electronic components (detectors, TIAs, ADCs and signal processors) may be fully integrated on the same wafer substrate, e.g., silicon electronics with silicon or silicon-germanium detectors, or III-V semiconductor detectors and electronics. Alternatively, hybrid integration may be used where dies with different functionalities (detectors, amplifiers, mixed-mode circuits etc.) are incorporated on the substrate using pick-and-place techniques. Different functional elements may be spatially separated as shown schematically in this figure, or may be implemented as “composite pixels” that incorporate these different functional elements in close proximity to each other as shown in
Throughout this description, the embodiments and examples shown should be considered as exemplars, rather than limitations on the apparatus and procedures disclosed or claimed. Although many of the examples presented herein involve specific combinations of method acts or system elements, it should be understood that those acts and those elements may be combined in other ways to accomplish the same objectives. With regard to flowcharts, additional and fewer steps may be taken, and the steps as shown may be combined or further refined to achieve the methods described herein. Acts, elements and features discussed only in connection with one embodiment are not intended to be excluded from a similar role in other embodiments.
As used herein, “plurality” means two or more. As used herein, a “set” of items may include one or more of such items. As used herein, whether in the written description or the claims, the terms “comprising”, “including”, “carrying”, “having”, “containing”, “involving”, and the like are to be understood to be open-ended, i.e., to mean including but not limited to. Only the transitional phrases “consisting of” and “consisting essentially of”, respectively, are closed or semi-closed transitional phrases with respect to claims. Use of ordinal terms such as “first”, “second”, “third”, etc., in the claims to modify a claim element does not by itself connote any priority, precedence, or order of one claim element over another or the temporal order in which acts of a method are performed, but are used merely as labels to distinguish one claim element having a certain name from another element having a same name (but for use of the ordinal term) to distinguish the claim elements. As used herein, “and/or” means that the listed items are alternatives, but the alternatives also include any combination of the listed items.
This application is a continuation of U.S. patent application Ser. No. 16/032,535, which was filed on Jul. 11, 2018, which claims priority to U.S. Provisional Application 62/536,425, filed on Jul. 24, 2017 and to U.S. Provisional Application 62/675,567, filed on May 23, 2018, the contents of each of which are fully incorporated herein by reference.
| Number | Name | Date | Kind |
|---|---|---|---|
| 5373460 | Marks, II | Dec 1994 | A |
| 10649088 | Sebastian | May 2020 | B1 |
| 11187807 | Satyan | Nov 2021 | B2 |
| 20160313445 | Bailey | Oct 2016 | A1 |
| 20180284240 | LaChapelle | Oct 2018 | A1 |
| Number | Date | Country |
|---|---|---|
| 101876705 | Nov 2009 | CN |
| 103630888 | Mar 2014 | CN |
| Number | Date | Country | |
|---|---|---|---|
| 20220043155 A1 | Feb 2022 | US |
| Number | Date | Country | |
|---|---|---|---|
| 62675567 | May 2018 | US | |
| 62536425 | Jul 2017 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 16032535 | Jul 2018 | US |
| Child | 17505640 | US |
