Patent Application
20120236379
Current scanning LADAR (Laser Detection and Ranging) devices commonly use dual or nodding/rotating galvanometric polygonal mirrors. These traditional mirror systems are relatively expensive, large, and heavy and correspondingly require more power than would a miniature system. Systems and methods for reducing the cost, size, and/or energy consumption of LADAR devices are currently being sought.
In accordance with an aspect of the invention, LADAR devices employ microelectromechanical system (MEMS) mirrors that enable use of LADAR in situations where cost, weight, power and form-factor are constrained. Suitable angular ranges and scanning frequencies for a MEMS structure can be achieved using low resistance hinges or flexures with or without angle multiplying optics. Feedback loops that receive input from sensors and generate drive signals for actuators can drive the mirror for triangular motion for both fast and slow axis oscillations. Weighting of the mirror structure and selection of the spring constants of flexures can also provide the mirror with natural or resonance frequencies corresponding to odd multiples of the desired oscillating frequencies to assist in achieving the desired triangular motion. In particular, the mirror can be driven at or near resonant frequencies for efficient energy transfer and large amplitude oscillations while still providing triangular motion.
One embodiment of the invention is a MEMS mirror system. Another embodiment of the invention is a scanning LADAR containing a MEMS mirror system. The MEMS scanning LADAR can provide a real-time 3D image sensor that is compact and lightweight. Image processing and control systems can interface with a LADAR front-end device to post-process output data from the LADAR system for purposes such as image display, analysis, and/or autonomous system control.
In one embodiment, a MEMS mirror can be fabricated without clean-room semiconductor type processes and does not require the development costs of systems manufactured using conventional semiconductor processing. While MEMS mirror configurations can be converted to semiconductor processes for high-volume production, the non-semiconductor mirrors are producible in moderate volumes without costly clean-room processes. The non-semiconductor MEMS mirror can employ piezoelectric actuators such as lead zirconate titanate (PZT) actuators. For the piezoelectric actuator to produce a large angular rotation of a mirror, the PZT or other piezoelectric material can be made thin because part of the resistance to rotation arises from the bulk modulus “stiffness” of the actuator. The substrate supporting the PZT and mirror may also be thin and lightweight to reduce resistance to rotation. Further, hinges can attach the PZT actuators to the mirror to further reduce resistance to rotation. A hinge or flexure can similarly be thin to offer very little resistance to flexing. A hinge can be fabricated in the plane of the supporting substrate (e.g., by patterning the supporting substrate or overlying layers) or provided by a structure that is attached to and extending from the substrate. In one configuration, the actuators attach to the hinges or flexures, and the hinges or flexures transmit force to the mirror from the actuators, causing the mirror to rotate when the actuators move. In another configuration, the actuators attach to the mirror area, so that movement of the actuators rotates the mirror about the hinges or flexures attached to the mirror.
Some further inventive aspects of systems and methods disclosed herein include but are not limited to the following. 1.) The use of closed loop control systems with a MEMS mirror and the proportional-integral-derivative (PID) controllers. Embodiments of closed loop control systems can implement mirror control processes including continuous mirror scan control, as some embodiments of the closed loops control a position waveform with another control signal waveform rather than point by point. 2.) The MEMS mirror structures and fabrication processes can be completed with or without semiconductor processing requiring clean room environments. 3.) Sensors and associated electronics can be integrated into MEMS mirror configurations to measure angles for feedback loop control of the mirror. 4.) MEMS mirror mechanical parameters can include mass distribution and spring constants that provide resonant frequencies that are odd multiples of a scan frequency to facilitate constant angular velocity or triangular motion of the MEMS mirror. 5.) LADAR or MEMS mirror systems can generally contain any combination of the features disclosed herein.
Use of the same reference symbols in different figures indicates similar or identical items.
In accordance with an aspect of the invention, a Laser Detection and Ranging (LADAR) system can include microelectromechanical system (MEMS) scanning mirrors, low resistance flexures or hinges, and drive circuits for continuous scanning in a relatively low cost, energy efficient, compact device.
Laser system 120 operates in a pulsed mode with pulse timing under the control of timing control circuit 130 or signal processing unit 180. The characteristics of laser 120 and laser drive circuits 125 used in LADAR module 100 may be chosen according to the desired performance parameters of LADAR module 100. In general, applications of LADAR module 100 may specify or require a specific range or maximum depth of scanned space 116 and resolution (e.g., number of pulses per scan line) that laser 120 must achieve. Laser 120 and drive electronics 125 may thus be chosen to provide a sufficiently high output power to provide an adequate reflected signal at a desired maximum range, a sufficiently short pulse duration to enable a desired range resolution, a sufficient pulse repetition frequency to match the motion of a scanning mirror 114 in MSM 110 and a desired number of measurement points per scan line, and a laser output wavelength in the infrared (IR) or other part of the electromagnetic spectrum to reduce background light interference and provide eye-safe operation. In an exemplary embodiment, laser 120 produces 1550 nm light or light with a range from 869 to 1550 nm. Other characteristics of laser 120 (e.g., environmental and package characteristics) may also be selected based on the above and further user-requirements, e.g., to provide a compact, lightweight, portable, and low input-power LADAR module 100.
The beam from laser 120 is directed through transceiver optics 150 before reflecting off a scanning mirror element 114 in MSM 110. Scanning mirror 114 preferably has a flat and continuous mirror surface rather than including an array of separate mirrors. In one embodiment, scanning mirror 114 has a continuous reflective surface that is about 4 mm by 4 mm square, but the reflective surface of scanning mirror 114 could have other shapes and sizes, e.g., circular, provided that the reflective surface is large enough to accommodate the transmitted beam and or collection of returning light. The reflectivity of scanning mirror 114 generally should be high, e.g., above about 90%, at the frequency of light produced by laser 120 for energy efficiency and to avoid overheating at the specific power and duty cycle of laser.
The reflective portion of scanning mirror 114 is rotatably mounted for scanning an incident beam through a raster type motion when a mirror drive system 112 drives scanning mirror 114 as described further below. In an exemplary embodiment, scanning mirror 114 has independently controlled oscillations about X and Y axes and provides up to about ±8° of rotation about each axis with triangular mirror motion. Other embodiments may have smaller ranges of motion or different angular ranges in X and Y directions, e.g., about 4° in the X direction and about 3° in the Y direction. Triangular motion ideally forces the mirror motion into a linear angular motion except at turning points where the motion reverses direction. With triangular motion, light pulse generated at a constant frequency can provide distance measurements that are evenly distributed through space 116. Providing triangular motion generally requires overcoming a natural tendency for sinusoidal oscillations. Techniques such as additional mirror drive inputs, additional resonators, and/or closed loop control as described further below may be used to provide triangular motion. Relatively large mirror rotations (e.g., ±8°) in a MEMS structure require either sufficient static force or dynamic multiplication effects provided by driving at resonance of the mechanical mirror system. Driving near resonance may be particularly useful for higher frequency motion, sometimes referred to herein as “fast axis” oscillations. For example, static driving below the rotational resonance about a “slow axis” and resonant drive at the natural frequency of a “fast axis” can be used for the slow and fast axes respectively. Magnetic, electrostatic, and/or piezoelectric drive mechanisms can be employed. However, piezoelectric drive may have the advantage of providing a self-contained drive mechanism and freedom from inherent constraints such as pull-in.
An optical angle multiplier 140 such as a lens 142 or other optical system can increase the angle of the transmitted scan beam to increase the angular range of LADAR module 100 and provide a larger scanned volume 116. Alternatively or additionally, a static (fixed) flat mirror 144 that reflects the already scanned beam back to scanning mirror 114 for a second reflection can increase the scan angle. However, for a second reflection, scanning mirror 114 may need a significantly larger area because reflection from mirror 144 will generally move the beam laterally. Angle multiplier 140 can thus increase the angle of the scan to produce a larger scanned volume 116 and correspondingly enhance the “field-of-view” of LADAR module 100.
Some of the laser light transmitted from angle multiplier 140 reflects off the surfaces of objects in scanned volume 116 and returns to LADAR module 100.
Detector 160, which can be a photodiode or similar light detector, receives the returning light and produces an analog signal representing the measured intensity of light having the transmitted frequency. Optical filters or frequency separation optics can be employed in detector 160 or receiver optics (e.g., optics 150) if detector 160 is otherwise insufficiently selective of the desired frequency. Data acquisition circuit 170 converts the analog signal from detector 160 into a digital signal and provides the digital signal to signal processing unit 180. Signal processing unit 180, which can be implemented using a general purpose microcontroller or microprocessor with suitable software or firmware, receives the signal from data acquisition circuit 170, identifies each pulse corresponding to reflected light in the digital signal, determines the time between transmission and return of the light, and determines a distance to a reflecting surface from the time of flight of the light pulse.
Timing control circuit 130 can produce a master timing signal for control and synchronization of laser driver 125, mirror scanning module 110, and signal processing unit 180 so that signal processing unit 180 can identify each transmitted pulse and the corresponding returning pulse of light. As a result, signal processing unit 180 produces a series of measurements of the distances to reflecting surfaces in scanned space 116, and each distance corresponds to a specific direction in the scan pattern of MSM 100. Accordingly, each scan through space 116 produces a two-dimensional array or frame of distance measurements that indicate the locations of points on reflecting surfaces in space 116. Signal processing unit 180 can include an interface for transmission of a series of distance measurements or a series of distance measurement frames to a computer or memory system that uses or records temporal changes in the three-dimensional volume of scanned space 116.
In one embodiment that may be used for automotive systems, LADAR module 100 provides a frame rate of about 15 fps (frames per second) with an image resolution of 320 pixels or distance measurements per scan line (e.g., per horizontal line), 240 scan lines per frame, and a maximum distance measurement of about 100 m. Other embodiments (e.g., for Natural User Interface applications) may have different requirements, e.g., a higher frame rate of 30 fps or 60 fps, resolution of 640×480 or higher, and a reduced maximum range (to about 12 m or less).
MSM 110 can use a variety of MEMS based mirror configurations to achieve desired frame rates and field of view for LADAR system 100. In one scanning mirror configuration, torsional flexures 210 as shown in the partial cutaway view of
MEMS mirror structure 300, particularly base 310, can be mounted or affixed on a printed circuit board (PCB) 340 as shown in the cross-sectional view of
Actuators 322, 324, 326, and 328 can be formed using a piezoelectric structure that bends or arches in response to an applied voltage. Piezoelectric material such as zinc oxide (ZnO) can be applied to MEMS mirror 300 using a thin film, spin on, or SolGel process. In operation, actuators 322 and 324 bend in opposite directions to raise one end of frame 320 and lower the other end of frame 320. Flexures 312, which connect frame 320 to base 310 and define a slow axis of rotation of mirror 330, are thin enough (10 to 50 μm) to twist when actuators 322 and 324 apply force to frame 320 through hinges or connectors 314, thereby lifting one end of frame and pushing down the opposite end of frame 320. Mirror 330 and underlying support area 318 are connected to frame 320 through flexures 316 that are perpendicular to flexures 312. Actuators 326 and 328 directly connect to flexures 316 and act to twist flexures 316 and rotate mirror 330 about a fast axis when actuator drive voltages are applied.
Frame 430 is connected to a base 450 by flexures 440 that permit rotation of frame 430 about a slow axis of mirror rotation. Flexures 440 are connected to piezoelectric actuators 455 that actuate rotation of frame 430 as described above with reference to
An alternate embodiment for fast axis electrodes 665 can be driven with inclined electrodes 665 as shown in
Each of four actuators used to drive oscillations of frame 820 about the slow axis includes two parts 850A and 850B that have a non-linear arrangement. The greater combined length of the parts 850A and 850B forming an actuator gives each of these actuators a greater range of movement, which may permit a greater angular range for slow axis oscillations.
In the embodiment of
The MEMS mirror systems described above can generally be fabricated using know semiconductor fabrication techniques for forming MEMS structures. For example, a fabrication process can start by depositing a first metal layer, a thin layer of piezoelectric material, and a second metal layer on a semiconductor or semiconductor-on-insulator substrate. Multiple photolithographic and etching processes can then be employed to define the areas of piezoelectric actuators including piezoelectric material sandwiched between upper and lower electrodes, create conductive traces from the metal layers, and thin or etch through the substrate to separate the mirror area, inner frame, and outer frame or base and to create flexures or flat hinges. Areas of the substrate can generally be thinned where desired through a top or back etch process. For embodiments of the invention employing HAR hinges, photolithograph and etching process can create alignment features such as grooves on the substrate, and the hinges can be created in a separate process and attached, e.g., glued using an epoxy, to the substrate using semiconductor processing methods.
An alternative process can fabricate MEMS mirror systems without requiring the clean room environment normally used for semiconductor processing. In particular, instead of using a semiconductor or SOI substrate, a metal substrate (e.g., stainless steel substrate) can be coated with piezoelectric material, and a multi-step etching process can create gaps and regions of different thicknesses needed for different structures. For example, the main material may be 500-μm thick stainless steel. A first etch step removes 200 μm to thin areas such as mirror area 1010 to decrease weight or create grooves for attachment of HAR hinges. A second etch step removes a total of 400 to 450 μm to leave regions thin enough to act as flexures. A final etch process etches through where separations between areas are required. Actuator areas, particularly areas for actuators that drive fast axis scanning of a mirror area, may further include a pattern of holes that may be etched through the actuator areas to further lighten those areas for fast actuation. Layers of piezoelectric material (e.g., PZT) can be coated on one or both sides of the actuators areas. Thin layers of piezoelectric material (e.g., PZT 2 to 125 μm thick) may be preferred again for faster actuation. It may be noted that although these etching processes can be implemented using conventional integrated circuit or semiconductor processing techniques, the dimensions of area are typically large enough that stringent clean room techniques may be unnecessary, thereby allowing lower manufacturing costs.
In accordance with another aspect of the invention, scanning mirrors can be driven to provide a trianglular motion in which the angular velocity is constant except at the turning points. However, mechanical systems generally have a tendency for sinusoidal oscillation at a characteristic or resonant frequency of the mechanical system. To aid in providing triangular motion, the mass distribution of the system and the spring constants of flexures or hinges can be selected so that resonant oscillations are at odd multiples of a desired scan frequency. For example,
A LADAR system, having a mirror structure with natural mechanical modes that are odd multiples of the desired oscillation frequency or not, can employ feedback loops using sensors to measure the movement of the scanning mirror and controllers that generate actuator drive signals that produce the desired triangular motions.
Sensing circuits 1230 sense the angular displacement of the mirror, e.g., as a change in capacitance, and a charge amp can convert the capacitance change into the voltage of a sensed signal θ. A gain stage (not shown) can amplify sensed signal θ for comparison to the target reference signal ref by a subtractor 1320.
LADAR systems such as described above have many uses including, for example, collision avoidance and parking systems in automobiles and vision systems for robots. Another use of LADAR systems is for a Natural User Interface (NUI) for a computer system, such as a personal computer or a game console. Some NUI systems currently use “Triangulation” Technology (TT) to sense 3D spatial information. In contrast, LADAR uses time-of-flight (TOF) technology to sense 3D spatial information. There are areas where TOF may have advantages over TT as follows.
LADAR module 100 can measure the time-of-flight for a light pulse that travels from LADAR module 100 to user 1510 or a solid object in the surroundings of user 1510 and then returns to LADAR module. LADAR module provides the measured times or derived distance measurements to computer 1520. A distance measurement for LADAR module 100 is simply the product of a time-of-flight measurement and the speed of light. In general, LADAR module 100 can construct a frame of distance measurements by scanning the laser beam through the desired field of view as described above. Computer 1520 may contain a program module (not shown) that uses the frames of spatial measurements from LADAR module 100 to identify the location of user 1510, movement of user 1510, and possible specific movements of body parts such as the legs, arms, hands, or fingers of one or more users 1510. Changes such as changes in facial details caused by moving one's mouth or jaw or blinking one's eyes may be detected using LADAR or separate imaging and image analysis systems (not shown). The measured movements can then be interpreted as control instructions for a program being run on computer 1520, e.g., to control the action of a curser, select program objects, enter information, or operate a game program.
The operating parameters of LADAR module 100 for an indoor application that is expected to be most common for NUI systems may be different from parameters that may be optimal for outdoor applications such as uses of LADAR systems in automobiles. For NUI use, the field of view may be on the order or 45° or more, and the desired range of measurement may be roughly from a minimum distance of about 0.3 to 1 m and a maximum of about 5 to 10 m. The resolution of each frame may be sufficient to detect leg, arm, hand, and possibly finger movements of user 1510 when the user is within the target range of LADAR module 100. A frame containing 640×480 distance measurements may be suitable for many NUI systems. The frame rate needed for an NUI system will generally depend on the reaction rate required for the programs being controlled. For example, frame rates similar to those used for video (e.g., 15 fps to 60 fps) may be desired in order to simulate smooth control of a program or a fast action game. The LADAR modules described may have physical parameters that are adjusted to optimize performance for a particular NUI system.
NUI systems using LADAR time-of-flight (TOF) technology may provide the following advantages over NUI systems using triangulation technology (TT).
a) Range Resolution-1: TT has good resolution at a specific distance (e.g., 1 cm at 3 m) but the resolution quickly degrades with distance. TOF LADAR systems can be designed for long ranges, and there is no inherent limit to improving the resolution and TOF range. In particular, the resolution of a TOF system is fundamentally independent of distance. NUI systems using TOF may be able to provide required sensing resolution at a much greater range than can be provided by TT systems.
b) Range Resolution-2: TOF range resolution is dictated by the time duration of four basic steps in the creation of range values: pulse generation time, detector speed, signal processing time, and backend processing time. The backend processing times are not a problem factor for achieving higher resolutions, but all the other steps can be. For TOF to achieve the desired level of range resolution for NUI, use of high-performance or custom devices for pulse generation, detection, backend and signal processing may be needed.
c) Range Resolution-3: TT determines distance by computing the difference between the 2D dot or grid patterns projected onto surfaces separated by some distance from the TT sensor. TT range resolution depends on analyzing the 2D pattern projected on all surfaces. Hence, measurements depend on the spatial resolution of a 2D pattern and frame rates depend on the speed of available video or image sensors. In TOF, there is no such 2D resolution requirement; there simply is round-trip pulse detection from individual points on the surfaces.
d) Lag-Time: TT is computation intensive, which may cause time delays between the user's action and the corresponding display of that action on a computer display. Such delays are a big problem in high speed action games and other potential applications. TOF is much less computation intensive. Again, for example, TT determines distance by computing the difference between the 2D dot patterns projected on all surfaces some distance from the TT sensor. In TOF, there is no such calculation; there is simply pulse detection from individual surfaces at each point (pixel). Hence TOF could reduce lag times and increase the speed and time resolution of a NUI system.
e) Frame Rate (fps): Although a frame rate of 15 fps (for a 100 m range LADAR) are primarily described above, higher frame rates are more easily achieved when the maximum range of a LADAR system is shortened to the few meters that NUI systems need. The relationship between the fps and range depends on several other performance specifics, and there is plenty of room to increase a LADAR system to the NUI's 30 fps or even 60 fps. Further, high frame rates may be difficult and expensive to achieve with TT systems because TT systems use 2-D image sensors and very high frame rate video sensors may be expensive and difficult to obtain.
Although embodiments of the invention have been described to illustrate specific examples, the description of such examples and should not be taken as a limitation. Various adaptations and combinations of features of the embodiments disclosed are within the scope of the invention as defined by the following claims.
This patent document claims benefit of the earlier filing dates of U.S. provisional Pat. App. Ser. No. 61/376,223, filed Aug. 23, 2010 and U.S. provisional Pat. App. Ser. No. 61/435,729, filed Jan. 24, 2011, which are hereby incorporated by reference in their entirety.
| Number | Date | Country | |
|---|---|---|---|
| 61376223 | Aug 2010 | US | |
| 61435729 | Jan 2011 | US |
