CROSS-REFERENCE TO RELATED APPLICATION 10001} This application claims priority to U.S. Patent Application No. 15/054,570, filed

February 26, 2016, which is hereb incorporated by reference in its entirety,

BACKGROUND

|0002} Unless otherwise indicated herein, the materials described in this section are not prior art t the claims in this application and are not admitted to be prior art by inclusion in this section. 0003} Radio detection and ranging (RADAR) systems can be used to actively estimate distances to environmental e ures by emitting radio signals and detecting returning reflected signals. Distances to radio-reflective features can. be detennined according to the time dela between ransmissio and reception. The radar system can emit a signal that varies in frequency ove time, such as a signal with a time-varying frequency ramp, and then relate the difference in frequency between the emitted signal and the reflected signal to a range estimate. Some systems may also estimate relative motion of reflective objects based on Doppler frequency shifts in the received reflected .signals.. 0004} Directional. antennas can be used for the transmission and or reception. of signals to associate each range estimate with a bearing. More generally, directional -antennas can also be used to locus radiated energ on a given field, of view of interest. Combining the measured distances and the directional information allows for the summnding environment features to be mapped. The radar sensor can thus be used, for insianee, by an autonomous vehicle control system to avoid obstacles indicated by the sensor information.

I [9Θ05] Some example automotive .radar systems may be configured to operate at. an electromagnetic wave frequency of 77 Gigahertz (GHz), whic corresponds to a millimeter (mm) wave electromagnetic wavelength (e.g., 3,9 mm for 77 GHz). These radar systems may use antennas that can focus the radiated energy into tight beams i order to enable the radar system to measure an environment with hig 'accuracy, such as a environment around an autonomous vehicle. Such antennas may be compact (typically with- rectangular form factors), efficient (i.e., with little of the 77 GHz energy lost to heat in the antenna or reflected back into the transmitter electronics), and low cost arid easy to manufacture (i.e., radar systems with these antennas can he made in high volume).

SUMMARY

0006J Disclosed, herein are embodiments that relate to determining an offset, for automotive radar based on unstructured data. In one. aspect the present application describes a method including transmitting a plurality of radar signals from a pluralit of different locations by a radar unit of a vehicle. The method also includes receiving a plurality of reflection signals, wher each reflection signal i associated with one of the transmitted radar signals. Additionally, the .method includes determining, b a processor, at least one stationary object that caused reflection in the plurality of reflection signals. Further, the method includes based oft the determined stationary object _* determining, by the processor, an offset for the radar unit. The method yet further includes operating the radar unit based on the determined offset Furthermore, the method includes controlling an autonomous vehicle based on the radar unit being operated with the determined offset.

(00073 In another aspect, the present application describes a vehicle. The vehicle includes a radar unit The radar unit is configured to transmit a plurality of radar signals, f om a plurality of different locations of a vehicle _* and receive a plurality of reflection signals, wherein each reflection signal is associated with one of the teansmitted radar signals. Additionally, _. the apparatus includes mounting plate configured to couple the radar unit to a mounting ^' structure- on a vehicle. The apparatus also includes a computational unit configured to perform operations. The computational unit is configured to determine at least one stationary object that caused a reflection in the plurality of radar signals. The computational unit is further configured to based on the determined stationary object, determine an offset for the radar unit Additionally, the computation unit is configured to operate the rad r unit based ^' on the determined offset. Further _* the computation unit is configured to coatiOl the vehicle based on the radar unit being operated with the determined offset,

[0008 In yet another example, a computing device is provided. The computing device may include a processor and a computer readable medium having stored thereon program instructions that when executed by the processor cause the computing device t perform functions. The functions include causing the transmission of a plurality of radar signals from a plurality of different locations by a radar unit of a vehicle. The functions also include, causing the reception of a plurality of reflection signals, where each reflection signal is associated with one of the transmitted radar signals. The -functions additionally include determining, at least one stationary object that caused reflection in the plurality ^' of reflection signals. Further, the functions include based on the determined ^' stationary object, determining an offset for the radar unit Furthermore, the functions include operating the radar unit based on the determined ^' offse Additionally, the functions include controlling an autonomous vehicle based on the rada unit being operated with the determined offset. In another aspect, the present application describes an apparatus. The apparatus m be configured i¾r determining an offset for automotive radar based on unstructured data. The apparatus may further include means for transniitthig a plurality of radar signals from a plurality of different locations by a radar unit of a vehicle, The apparatus also includes means for receiving a pluralit of reflection signals, where each reflection signal is associated with one of the transmitted radar signals. Additionally, the apparatus meludes means for determining at least one stationary object that caused reflection i the plurality of reflection signals. Farther _* the apparatus includes based on the determined stationary object, means for determining an offset, for the radar unit The apparatus yet further includes means for operating the radar mean based on the determined offset. Furthermore, the apparatus includes means for controlling an autonomous vehicle based on the radar means being operated with the determined offset.

|0019J The foregoing summary is illustrative only and is not intended to be in any way limiting. In addition to the illustrativ aspects, embodiments, and features described above, former aspects, embodiments, and features will become apparent by reference to the figures and the following detailed description.

BRIEE DESCRIPTION OF Till FIGURES

mill Figure 1 illustrates an. example layout of radar sectors.

100121 Figure 2 illustrates example beam steering for a sector for a radar unit.

;00131 Figure 3 illustrates an example radar unit mounting.

00141 Figure 4 illustrates an example computing device for performing some

methods disclosed herein. [991.5) Figure 5 is an example method for radar mounting estimation with tinstraetnred data.

DETAILED PESCMPTOK

|9916) I» the following detailed description, reference is made to the accompanying figiffcs, which form a part hereof. In the figures, similar symbols typically identify similar components, unless context dictates otherwise. The illustrative embodiments described in the detailed description, figures, and claims ^' are not meant to be ..limiting. Other embodiments may be utilized, and other changes may be made, without departing from the scope of the -subject matter presented herein, it will be readily understood that the aspects of the present disclosure, as generally described herein, and illustra ed in the figures, can be arranged, substituted, combined, separated, and designed in a wide variety of different configurations, all of which are explicitly .contemplated, herein.

[9917J The following detailed description relates to an apparatus and methods for automotive sensor offset determination based on the capture of unstructured data. In practice, vehicular radar systems may feature multiple rada units with each radar unit ^' having an associated field of view. Typically, the alignment of the various radar units was a time- consuming- labor-intensive _, procedure requiring precise measurements and many expensive tools. The apparatus ami methods presented herein allow the ^' offset determination of radar unite in a more time efficient and -equipment-efficient manner.

0018! Traditionally, vehicular radar systems have featured single radar unit looking in a forward direction. These traditional radar systems would also typically only direct the radar beam in a single direction. When a vehicle has only a single radar unit with a beam that only was directed in a single direction, the precision required in tire placement of the radar unit may be not as great as systems with multiple radar units with sieerab!e beams. The non~steerable beam only interrogating one direction may cause the tower precision requirement For example, a traditional vehicular radar system may only be configured to deteci the presence of another vehicle directly in front of the vehicle having the radar system, Thus* as long as a radar beam generally points in the forward direction the traditional radar system would be able to detect the vehicle.

£W19J More advanced radar systems may be used with a vehicle in order to obtain a wider field of view than just that directly in front of the vehicle. For example, it may be desirable either for radar to fee able to steer a radar beam or for a vehicle to feature multiple radar units poi ting in different directions. Additionally, the radar units may be configure to detect more than just other vehicles located directly in front, of the vehicle with the radar unit. Thus, the radar system ma be able to ^■■ interrogate different regions than just the ^■ region in front of the vehicle, in some examples, multiple radar units ma fee combined with steerabie radar beams to further increase the interrogation region and the ^' 'ima inin resolution of the vehicular radar system,

10020$ An example radar system for use- with the present disclosure may include mufti- sector 90~degree field of view radar antenna architecture that may enable an antenna to both scan across approximately 90-degrees of the azimuth plane (e.g. the horizontal plane) while also being monntabie on various surfaces of a vehicle. Having a radar antenna with a 90-degree field of view may enable a radar system to scan a full 360 azimuth plane by having four radar units each configured to scan one 90-degree non-overlapping sector. Therefore, the example disclosed radar system may be able to steer a radar beam to interrogate the entire region in the azimuth plane of the vehicle. So that for example, four such radars located on four comers of a vehicle would provide a full 360 coverage around the vehicle. For example, -a system such as this may aid in autonomous driving of a vehicle,

£6021 J When each radar unit can scan or span a 0-degree region, placing 4 radar units on a vehicle may enable the vehicle to scan a beam over the full 360 azimuth plane. Each of the four radar units may be configured to scan beam over one sector (i.e. one quarter of the azimuth plane) and thus the entire plane may he scanned by the combination of the four radar units. In various examples, the placement of the radar units may be adjusted depending on the specific vehicle, the re uirements of the radar system, or other design criteria. In some additional examples, the radar units raay be configured to scan a region of an angular width that is not 90 degrees. For example, some radar unite ma scan 30 degrees, 120 degrees, or another angle. Further, in some ex m les, the radar unite on the vehicle may scan less than the full 360 azimuth plane. Additionally, in some examples, radar units may have overlapping scanning segments. -of the azimuth plane.

In some examples, the radar sectors may be defined based on where the rada units may be mounted on the vehicle, in one example, one radar unit ma be mounted in each of the side mirrors of the v ehicle. The other two radar units may be mounted behind the tai H igh is of the vehicle, in this example, the quadrants may be defined based on axes where one axis aligns with the direction of vehicular moti on and the other axis aligns with the middle of the vehicle from front to back. In another example, the radar units ma be mounted in orde to have one pointing forward, one pointing backward, and one pointing to each side. In this second exam le, the axes of the quadrants may be at a 45-degree angle to the direction of motion of the vehicle. Additionally, the radar unit ma be mounted on top of the vehicle.

[00231 The modular multi-sector 90-degree field of view radar antenna architecture may be able to steer the radar beams emitted from each radar unit. The radar beams may be steered by the radar units in var ous ways, For example, in some embod ments the radar units may be able to steer the beam in an approximately continuous maimer across the 90-degree field of view for. the respective antenna or the radar unite ma ^' be configured with sectoral sub beams- spanning the 90 degrees. In other embodiments, the radar units may be able to steer the radar beam to predetermined directions within the 90-degree field of view for the respective antenna _*

[00241 One aspect of the present disclosure provides an apparatus for the calibration of the placement of radar sensor units- on- vehicle, for vehicular radar, in some instances, it may be desirable to have a 360-degree Field of View (FOV). By having data front 360 degrees, the vehicle ma more accurately sense objects near the vehicle to aid in safe driving and navigation. The present disclosure ma use multiple separate radar sensor units, each configured, to scan ove a portion of the azimuth plane to achieve a full 360 FOV,

[00251 I» some approaches, the vehicle ma be configured with multiple radar units. For ease of explanation of the present disclosure, it will be assumed that the vehicle has four radar ^• unite; however, more or fewer radar units, in examples -with four radar units, each radar unit may be configured to each scan over one particular quadrant (i.e. 90 degrees) of the azimuth plane of the vehicle. In traditional approaches, it may be desirable to locate the radar sensor units on the vehicle with high accuracy. Fo example, it ma be desirable for each radar unit to have an- angular tolerance of ± 1 degree ra both the azimuth and elevation planes. Unlike traditional systems, the present disclosure pro vides a radar apparatus and method that allows the radar units to have a larger angular tolerance, but enables radar calibration through a calculation of a radar offset,

|0026J The radar calibration of the present disclosure is based on making radar measurements. Radar nits ma be mounted to vehicle without using the traditional tow- toleranee alignment, in order to operate the radar system, an offset (or lack of offset) may be calculated for each of the radar" unite coupled the to the vehicle. Once the offset for each radar unit is calculated, the radar unit may be operated with the veliicle in a manner similar to that if the radar units had been aligned using the traditional low- tolerance alignment,

0O27J The method for calibrating the radar system may include operating the vehicle on which the radar units are mounted while making radar measurements at several locations. In some examples, the several locations may be a series of locations as the vehicle is traveling. The several locations ma be relatively closely spaced to each other (e.g. within, a few feet) s that the radar system may be able to see at least some of the same objects causing radar reflections across at. least a subset of the several locations. A processor of the system may attempt to determine which objects that cause radar reflection are stationary objects.

Based on the measurements taken with the radar calibration apparatus at each location, the radar reflections of the -stationary objeci(s) may be analyzed. When a static object is seen by a .radar unit when the radar unit is at a plurality of different positions* the system may be able to determine the alignment for the radar unit. The alignment ma include both and azimuth angle offset and an elevation angle offset. The azimuth angle offset and elevation angle offset for each radar unit may be stored in a memory. Based on the azimuth angle offset and elevation angle offset, the .radar unit may be operated as if it had been mounted with more precise tolerances. la sotne embodiments, the offsets may only he calculated with respect to either azimuth or elevation angle.

10029} If a radar (coupled to a vehicle) is moving and if the scene around tire radar is stationary, then it is possible to solve for the unknown radar position and orientation with respect the navigation frame of the radar's platform. Scenes are rarely completely stationary, especially if unstructured, so the present method and apparatus solve the harder problem of also distinguishing between, moving and stationary elements of the scene. Taken jointly, this represents a very difficult estimation problem. Disclosed herein is a solution that employ an expectation maximization approach to iteratively converge to a correct estimate of the unknown parameters.

|003OJ Figure 1 illustrates an example layout of radar sectors for an autonomous vehicle

102. As sho wn In Figure 1, each of the radar sectors may have an angular width approximately equal to the scanning range of the radar units (as will be described. -with respect to Figure 2). For example, the sectors of Figure I divide the azimuth plane around the autonomous vehicle 102 into 90 degree sectors. However, in examples where the radar units are configured to sca a radar beam over a different angle than 90 degree (not shown), the width and number of sectors may change. Although f igure 1 shows a car, the methods and apparatuses presented herein may b used with other vehicular systems as well, such as aircraft, boats, etc.

0031 J As shown in Figure 1, the radar sectors ma align with the axes (112a and 112b) of the vehicle 102. For example, there may be a front left, front right, rea left, and rea right sector defined by the midpoints of the vehicle 102. Because each sector corresponds to one radar unit,, each radar unit may be configured to scan across one sector. Further, becawse each example radar unit of Figure 1 has a scanning angle of approximately 0 degrees, each radar unit scans a region that approximately does not overlap with the scanning angle of any other radar unit. The layout: of radar sectors shown in Figure 1 is one example. Other possible layouts of radar sectors are possible as well

10032} In order to achieve radar sectors defined by the midpoints of the vehicle 102, each radar unit may he mounted at a 45-degree angle with respect to the two axes of the vehicle 102. By mounting each radar unit : -a 45 degree angle with respect to the two axes of the vehicle 102, a 90 degree scan of the radar unit would scan from one vehicle axis to the other vehicle axis. For example, radar unit mounted at a 45-degree angle to the axes in side mirror unit 104. may he able to scan the front left sector (i.e. from the vertical axis 3 12a through the front of the vehicle 102 to the axis 1 12b that runs through fee side of the vehicle). An additional radar unit may he mounted at a 45-degree angle to the axes in side mirror unit 106 ma he able to scan the front right sector. In order to scan the back right sector, a radar unit may be mounted in iaillight unit. 110. Additionally, in order to scan the back left: sector, a radar unit may be mounted in Iaillight ^• unit 108. The radar unit placements shown Figure 1. are merely one example. In various other examples, the radar units may be placed in other locations, such as on top of the vehicle, or within or behind other vehicle components. Further , the sectors may also be defined ^' differently in various embodiments. For example, the sectors may be at a 45~degree angle with respect to the vehicle. In this example, one radar unit may face forward, another backward, and the other tw to the sides of the vehicle. [99331 I» some examples, all the radar units of vehicle 102 ma be configured with the same scanning angle. The aztBiuih plane around the vehicle is equal to a Ml 360 degrees. Thus, if each radar unit is configured with the same scanning angle, then the scanning angle for the radar units would be equal to approximately 360 divided by the number of radar units on the vehicle. Thus, for foil azimuth plane scanning, a vehicle 102 wit one radar unit would need that radar unit to be able to scan over the foil 360 degrees.

[0034j If the vehicle 102 had two radar units, each would scan approximately 180 degrees. For three radar units, each would foe configured, to scan 120 degrees. For four radar units, as shown in Figure * each may scan approximated 90 degrees. Five radar units ma be configured on the vehicle 102 and eac may be able to scan 72 degrees. Further, si radar units may be configured- on the vehicle 102 and. each ma be able to scan approximately 60 degrees.

[0035J The number of radar units may be chosen based on a number of criteria, such as ease of manufacture of the radar units, yehiete placement, : or other criteria. For example,: some radar units may foe configured with a planar structure that is sufficiently small. The planar rada ^• unit may be mouritable at various positions on a vehicle. For example, a vehicle may have a dedicated radar housing mounted on the top of the vehicle. The radar housing may contain various radar units. However, in other embodiments, radar units may be placed within the vehicle structure.

(Θ036| When radar units are located within the vehicle structure, each may not be visible f om outside the vehicle without removing parts of the vehicle. Thus, the vehicle may not he altered aesthetically, cosmetically, or aerodynamically from adding radar units. For example, radar units may be placed under vehicle trim work, tinder bumpers _* under grills, within housings for lights, within side mirrors, or other .locations as well in some embodiments, it ma be desirable to place radar waits in positions where the object covering the mdar unit is at least partiall transparent to radar. For example, various plastics, polymers, and other materials may form part of the vehicle structure sad cover the radar- waits, while allowing the radar signal to pass through.

10037} Additionally, in some embodiments, the radar units ma he configured with different scanning ranges for different radar units, For example _* in some embodiments a specific radar' unit with a wide scanning angle may not he able to be placed on the vehicle in the proper location, thus, a smaller radar unit, with a smaller -scanning angle may be placed in that location. However, other radar units may be able to have larger scanning angles. Therefore, the total scanning angle of the radar units may add u to 360 degrees (or more) and provide Ml 360 degree azimuthal scanning. For example, a vehicle may have 3 radar units- that each scan over 100 degrees and a fourth radar unit that scans over 60 degrees. Thus, the radar units may be able to scan the full azimuth plane, but the scanning sectors may not be equal in angular size,

|ββ3ί?| Figure 2 illustrates example beam steering for a sector for a radar unit 200. The radar unit 200 may be configured with a steerahle beam, i.e.. the radar unit 200 may be able to control a direction in which the beam is radiated. By controlling the direction in which the beam is radiated, the ^' radar unit 200 ma be able to direct radiation in specific direction in order to determine mdar reflections (and thus objects) in. that direction, in some embodiments, the radar unit 200 may be able to scan a radar beam in a continuous manner across the various angles of the azimuth plane. In other embodiments, the radar unit 200 may be able to scan the radar beam in discrete steps across the various angles of the azimuth plane. Ϊ0039]; The example radar unit 200 in Figure 2 has a radar beam 206 thai can be steered across a plurality of different angles. As shown in Figure 2, the radar beam 206 may have a halt- power bea width of approximately 22.5 degrees. The half-power bearawidth describes the width, measured in degrees, of a main lobe of the radar beam 206 between two points that correspond to half the amplitude of the maximum of the rada beam 206. In various embodiments, the half-power beamwidth of the radar beam 206 may be different than 22,3 ^• degrees. Additionally, in some embodiments, the half-power bearawidth of the radar beam 206 may change depending on the angle at which the radar beam 206 is pointed. For example, the half-power beamwidth of the radar beam 20 may be narrower when the radar beam 206 is pointed more closely to the orthogonal 204a (i.e. broadside) direction to th radiating surface and widen and the radar beam 206 is steered away from the orthogonal direction 204a.

|0649J In the example ^' Shown in Figure 2, the radar beam may be able to be steered to four different angles. The steering ^' angle may be measured with respect to the orthogonal 20% (I.e. broadside) direction t the radiating surface. The beam may be steered to +36 degrees at 204e and -3 degrees at 204e. Also, the beam may be steered to +12 degrees at 204b and -12 degrees at 204d. The four different angles may represent the discrete angles to which the radar beam 206 may be steered. In some additional examples, the radar beam may be able to be steered to two angles simultaneously. For example, the radar beam may be steered to both +12 and -12 degrees at the same time. This may result in a beam that is overall steered in the direction of the sura of the angles (e.g. -12+12=0, thus the beam in this example would be in the broadside direction 204a). However, when a radar beam is steered at two directions at once, the •tialf-power beamwidth of the radar beam may be widened. Thus, a radar resolution may decrease.

[0941]; By steering the radar beam 206 to each of angles 204b-2O4e _s the Ml 90-degree field of view can be scanned. For example, when the radar beam 206 is steered to +36 degrees 204e, the half-power beamwidth of the radar beam 206 will cover from +47,25 degrees to +24,75 degrees (as measured from the broadside direction 204a), Additionally, when the radar beam 206 is sieered to -3 degrees 20 0, the half-power beamwidth ^' of the radar beam 206 will cover from -47.25 degrees to -24,75 degrees. Further, when the rada beam 206 is sieered to +12 degrees 204b, the half-power beamwidth of the radar beam 206 will cover from +23.25 degrees, to +0,75 degrees. And finally, when the radar beam 206 is steered to -12 degrees 2044, the half- power beamwidth of the radar beam 206 will cover from -23.25 degrees to -0.75 degrees. Thus, the radar beam 206 will effectively be able to scan (i.e. selectively enable or disable the four beams spanning the angular width) from -47,25 to +47.25 degrees, covering a rang of 95 degrees. The mmiber of steering angles, the direction of the steering angles, and the half-power beamwidth of the radar beam 206 may be varied depending on the specific example,

(0042J For example, and further discussed below, a radar beam of a radar unit may be configured to only scan a 60-degree region. If a radar un t can scan a ^' 60-degree region, six radar ^• units may be used ' to sc n a fall 360 azimuth plane. However, if the radar -unit can scan 90 degrees, four radar units may scan the full 360 azimuth plane.

10043} Figure 3 illustrates an example radar unit mounting structure 300, In one example embodiment a radar unit mounting 300 may include a mounting baseplate 302 and an associated mounting location 304 where a radar unit 312 may be mounted to the mounting baseplate 302, The mounting location 304 may be the location where the radar umt 312 may be mounted to the mounting baseplate 302 while performing method 500 of Figure 5. Additionally, during operation of the autonomous vehicle, radar units 312 may be mounted to the mounting baseplate 302 in place of mounting location 304» In even further embodimen ts, the radar unit 312 and the mounting baseplate 302 may be integrated as one un and not separate components,

fi044| A mounting baseplate may be located at each location where a radar unit may be coupled to the autonomous vehicle. For example, each radar imit mounting location (such as locations 104, 106, 108, and 110 -of Figure I) may have a mounting baseplate 302 to which radar unit may be mounted. As shown in Figure 3, the mounting baseplate 302 may include mounting holes (one of which is labeled 308) configured to both align the radar unit when mounted as well as allow the radar unit to be coupled to the mounting baseplate 302 with attachment devices such as screws. The mounting baseplate 302 of Figure 3 is one example of a way the various radar units may be mounted to an ^' autonomous vehicle.

£6045} Whe radar unit 312 is mounted to mounting baseplate 302, the radar unit 312 may not be exactly aligned as designed. This misalignment may manifest as an offset from the desired mounting position. For example, -the radar unit 312 may have an offset when coupled to mounting baseplate 302 in terms of the elevationai angle offset 306a, roll angle offset 306b, and azimuthal angle offset 306c. Additionally, the radar unit 312 may have ^' art offset: when coupled to .mounting baseplate 302 in terms of the X offset, Y offset, and offset hi some embodiments, it may be desirable for a radar unit to be mounted within a threshold rang aroimd a desired direction. If the radar unit mounted within the threshold range, the offset of the radar unit may be -calculated. When the offset, is calculated, the processing system that processes radar signals may be able to account for the offset through data processing. For example,, each of the eievational angle 306a, roll angle 306b, and aziniuthal angle 306c may have a threshold range of ±1 degree from the desired eievational angle, roll -angle, and azimuthal angle. By determining the offsets, the processing system may mathematically compensate for the difference between the desired alignment and the actual alignment,

10046} In some embodiments, a computing device may implement the disclosed methods as computer program ^■ instructions encoded on non-trarisitoiy computer-readable storage media in a machine- readable format, or on other non-transitory media or articles of manufacture, The computing device ma be integrated within the vehicle or it may be a separate .computing device in communication, with the vehicle. Figure 4 is a schematic ilhistrating a conceptual partial view of an example computer program product that includes a computer program for executing a computer process on a computing device, arranged according to at least some embodiments presented herein,

|6047} Figure 4 illustrates a functional block diagram of a computing device 400, according to an embodiment The computing device 400 can be used to perform -junctions in connection with a reeoTjfigurahle mobile device with a balloon network, in particular, the computing device can be used to perform some or all of the functions discussed above in connection with Figures 1-5, {0048} The computing device 400 can be or include arious ^' types of de vices, such as, for example, a server, personal computer, mobile ^' device, cellular phone, or tablet computer. In a basic configuration 402, the computing device 400 can include one or more processors 410 and system memor 420. A memory bus 430 can be used for communicating betwee the processor 410 and the system memory 420. Depending on the. desired configuration, the processor 410 can be of any type, including: a microprocessor (μΡ), a nnerocontrolter (μ£¾ or a digital signal processor (DSP), ^■ .among others. A memory controller 415 can also be used with the processor 410, or iu some implementations, the .memory controller 415 can be an interna l part of the processor 410.

£00493 Depending on the desired configuration, the system memory 420 ca be o any type, including volatile memory (such as RAM) and non- volatile memory ^' (such as ROM, flash memory). The system memory 420 can include one or more applications 422 and program data 424, The application(s) 422 can include an index algorithm 423 that is arranged to provide inputs to the electronic circuits. The program data 424 can include content information 425 that can be directed to any number of types of data. The application 422 can be arranged to operate with the program data 424 on an operating system.

|00S0J The computing device 400 can have additional features or ^• functionality, and additional interfaces to facilitate commiini cation between the basic configuration 402 and an devices and interfaces. For example, data storage devices 440 can be provided including removable storage devices 442, non-removable storage devices 444, or both. Examples of removable storage and non-removable storage devices include magnetic disk devices such as flexible disk, drives and hard-disk drives (HDD), optical disk drives such as compact disk (CD) drives or digital versatile disk (DVD) drives, solid state drives (SSD), and tape drives. Computer storage media ean include volatile and nonvolatile, now-transitory, removable and non-removable media implemented in any method or technology tor storage of information, such as computer readable Instructions, data structures, program modules, or other data. [90511 The system memory 420 and the storage devices 440 are examples of computer storage media. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, D VDs or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium thai can be used to st re the desired information aid that ca» be accessed b the computing device 400.

[0052| The computing device 400 can also include output interfaces 450 that can include a graphics processing unit 452, which can be configured to communicate with various external devices, such as display devices 490 or speakers by way of one or more A/ ^' V ports or a coinmonication interface 470, The communication interface 470 can include a network controller 472, which can be arranged to facilitate communication with one or more other computing devices, over a network ..communication by way of one or more communication ports 474. The communication connection is one e m le of a comm uni cation media.

Communication media can be embodied by computer-readable instructions, data ^" structures, program modules, or other data in a modulated data signal, such as a carrier wave or other transport mechanism, and includes any information delivery media. A modulated data signal can be a signal mat has one or more of its characteristics set or changed in such a manner as to encode information in the signal- By way of example, and not limitation, communication medi can include wired media such a a wired network or direct-wired connection, and wireless media such as acoustic, radio frequency (RF¾ infrared (IB), and other wireless media.

|§Θ531 The computing device 400 can he implemented as a portion of a small -form factor portable (or mobile) electronic device such as a cell phone, a. ersonal data assistant (PDA), a

1.9 personal media player device, a wireless web-watch device, a personal headset device, an application specific device, or a hybrid device that include an of the above functions. The computing device 400 can also be implemented as a personal computer including both laptop computer and non-laptop computer configurations',

|6054| The disclosed methods can be implemented as computer program instructions encoded on a non-transitory computer-readable storage medium in a machine-readable format, or on other non-transitory media or articles of manufacture. The computer program produc t include* a computer program for executing a computer process on a computing device, arranged according to some disclosed implementations _*

|G 55J The computer program product is provided usin a signal bearing medium. The signal bearing medium can include one or more programming instructions that, when executed by one or more processors, can provide functionality or portions of the functionality discussed above in connection with Figures 1-3 and Figure 5. in some implementations, the signal bearing medium can encompass a computer-readable medium such s, but. not limited to, a hard disk drive, a CD, a DVD, a digital tape, or memory _* In ^' seme implementations, the signal bearing medium can encompass a computer-recordable medium such as, but not limited to, memory, read write. R/W) CDs, or R W D VDs, in some implementations, the signal hearing medium can encompass a communications medium such as, but not limited to, a digital or analog communication medium (for example, a fiber optic cable, a waveguide, a wired communications link, or a wireless communication link), Thus, for example, the signal bearing medium can be conveyed by a wireless form of the communications medium (for example, a wireless communications medium conforming with the IEEE 802,1 1 standard or other transmission protocol). [9856] The one or more programming instructions can be, for example,, computer executable instraetkms. A computing device ( such as the computing device 400 of Figure 4) can be configured to provide various -operations in response to the programming instructions conveyed to the computing device by one or more of the computer-readable medium, the computer recordable medium, and the communications medium.

«057 Figure 5 is an example method for radar mounting estimation with unstraetured data. Moreover, the method 500 of Figure 5 will be described in conjunction with Figures 1-4. At block 502, the method 500 includes : ^■ transmitt ng a plurality of radar signals from a plurality of different locations by a radar unit of a vehicle; A vehicular radar system may be configured to interrogate the region around the vehicle via multiple radar units. To interrogate the region around the vehicle, the radar system may transmit the radar beam in a given direction. The transmitted beam may reflect off objects in the region.

|ββ58] At block 504, the method 501) includes receiving a plurality of reflection signals, wherein each reflection signal is associated with one of the transmitted radar signals. The received reflections may allow the radar system and a computer to determiue what objects are located near the vehicle. Not only may objects themselves be determined, but the location (i.e. angle and range to objects), may be determined as well. .In order to operate correctly, radar units of the radar system need to both be placed in -correct locations ami have fairly precise alignment.

[W59] In .some examples, blocks 502 and 504 may be repeated several times. In one examples, the radar path may be divided into many consecutive coherent processing intervals (CP Is). During a CPI the radar unit both transmits and receives a waveform, applies a matched filter bank, and extracts a set of detections that measure range, Doppler, and angle (i azimuth, elevation or both) to t ie various objects that reflect radar. This data may be stored in a memory for further processing. In some examples, several CPis are captured before moving to data processing.

0060] At block 506, the method 500 includes determining, by a processor, at least one stationary object that caused reflection in the plurality of reflection signals, By analyzing the collected data, an algorithm may use a current best guess at radar mounting calibration, determine which measurements arise from stationary objects.

[0061] To determine which object that cause radar reflections are- stationary, a problem is defined as one of parameter estimation where the unknown parameters are radar mounting angle and position with respect to the car body system. Fundamentally, this calcuiation amounts to maximizing the correspondence between scattering, objects in both the Doppler and angle domains.

[0062] A pulse- opp er radar may be able to measure range, Doppler, and 2D bearing to the various- objects that cause radar reflections. In the general case all 3 orientation degrees of freedom (elevationai, rotational, and -an azimuihal) and 3 position degrees of freedom (X, Y, and an Z) are estimated. In some examples, the system may he configured to .only calculated the azimuth offset with respect to the vehicle.

(0063] in one example, the derivation is shown for a 2D car vehicle coordinate system.

Other means of calculating stationary objects and the. offsets may be used as well. The following variables are used to calculate offsets _* v = vehicle velocity

Θ = mounting angle

φ{— ^■ bearing angle to i— th scatterer d;— doppier to ί— tk scatter r

Zi ~ stationary indicator function f or i - th scatterer

N ~ # scatterers

·= denotes a realization of an random variable

N(x; μ, P) evaluates at x a Gaussian with mean μ and variance P 10064] Assume that the the measured reflection signals include independent and identically distributed Gaussian noise. The Gaussian noise may be defined as j and have a variance ¾ . So the bearing angle may be defined as,

¾ = Φ _ί ~ Θ + w¾

[00653 Similarly, the Doppier measurement ma be modeled, including Gaussian noise

«¾ having variance ¾, as,

2

fi _t ™ ίί·Ό5'(Φί) 5ϊη(φ _ί )] (-v) + w ₂

A ^'

f0066| A likelihood t/j may be defined that will, calibrate th system if all seatterers are stationary. On average, if the mounting angle is ideally aligned, then the joint Likelihood of the measurement is at its maximum, where the likelihood defined as,

i

0067) Note that Doppier and bearing measurements are uacorretated since the equations are ^' conditioned on scatterer position and mounting angle. 0068i In practice some unknown subset of the measurements will be made on objects that are in motion. An additional set of bidden random variables, z _f that indicate the stationarity of an. object ma be used. Because the algorithm tterativeiy approximates the unknown parameters, on the n-ih step previous estimate of unknown parameters 0 a d will be denoted.

Via the method of expectation maximization, the iterative estimation eqisation is then,

f0069J By treating the realization ¾ - 0 as one where the likelihood ψ is uninfermative

(since there is no longer a deterministic relationship between. Doppter and bearing) and taking the prior ^ ^>Ψ-$ί as: nnfflformative the equati n may be simplified.

|0070| This stage of processing may make an estimate of # I -¾ ^ * * This estimate may be made because the collection, period is long enough to assure man measurements are made on each object hi the scene. In so doing, the system will have enough data to perform temporal processing. This detection problem may be decomposed into two pieces: assignment and estimation. Assignment is the mapping of measurements to objects as well as the estimation of the number of underlying objects. Estimation is the calculation of the stationarity indicator variable posterior for each object. This probability is then mapped to the measurement associated with that object. hus,, stationary objects may be identified,

}O071] At block 508, the method 500 Includes based on the determined stationary object, determining, by the processor, an offset for the radar unit. Using the determined stationary object, the system may compute a best guess of radar- position and orientation. The radar position and orientation may include at least one offset,

2] A maximum likelihood estimate (MLB) of mounting angle is somewhat complicated by the presence of nuisance variables φ ₍ . The estimate- may be marginalized out or included as unknown parameters in the MLE - we choose here to perform the latter. The MLE objective function as the negative log likelihood,

The minimum of the MLE may be found via Newton's method. The key to the method is calculatin the Hessian (N-t-1 by N+l matrix H) and Jacobian (N+l column vector 3) of the likelihood.

h (fit + Θ - + i _t - [cosC≠i) -?ί (φ _ί )]ν) ¾ ¹ (- l ^' ·.ν ^η ( ,}∞$(φ _ί; )]ί? t ^~ ,Ni · _" ί ¾ø ¾,t « · » ¾¾']

0Θ7 ; Starting with some guess § _> $, for the unknown parameters, they are updated by the following rule until convergence. [§,■$? - [§,$ΐ - γΗ ^'1 ]

[0075} Where γ is some scale factor less than one. In some examples γ Is 1 *! 0 ^Λ -2,

[0076 The calculated MLE gives the offset for the respective radar unit. The previously* diseussed calculation (or different calculations) ma fee used to .determine an Offset for each radar unit of the vehicle. In some examples, a radar unit may have the correct alignment, thus the ^■ determined offset may be 2©ro degrees,

[0077} At block 510, the method 500 includes, operating the radar unit based on the determined offset. Once offsets are determined, the offset may he used by the processing system when locating various objects that cause radar reflections. For example, if an offset is determined to fee 2 degrees .in the positive azimuth direction, the processing system can compensate for this offset. Irs some examples, the processing system may operate the radar unit by applying the the offset to the calculation of a reflective object. Thai is, if the offset is +2 degrees in the azi uth plane, and an object appears to be at +1.5 degrees of azimuth, the system- may adjust for this +2 degree offset, and operate knowing that the reflecting object is really at +17 degrees (i.e. -K15+2 ^~ +17) in the azh ith plane. In some other examples, different mathematical functions may fee used to determine the position of the reflecting objects based on the determined offset

| ^' ββ?8| At block 512, the method 500 controlling an autonomous vehicle based on the radar unit feeing operated with the determined offset. Because the .radar unit is operated with the determined offsets at block 510, this radar information may be used with the navigation system of the vehicle to autonomously control the vehicle. While operating in the autonomous mode, the vehicle may use a computer system to control the operation of the vehicle with li e-to-no human input For exam ple, a human-operator may enter an address into an autonomous ehicle and the vehicle may then be able to drive, without further input from the human (e.g., the human does not have to steer or touch the brake/gas pedals), to the specified destination,

I0079J While the vehicle is operating autonomously, the sensor system ma be receiving dat about the environment of the vehicle from the radar system. The processing system of the vehicle may alter the control of the vehicle based on data received from the various sensors, in some examples, the autonomous, vehicle ma alter a velocity of the .autonomous vehicle in response to data from the various sensors, The autonomous vehicle may change velocity in order to avoid obstacles,, obey traffic laws, etc, When a processing system in the vehicle identifies objects near the auton mous vehicle, the vehicle may he able to change velocity, or alter the movement in another way. The location information used by the vehicle may be provided by the methods and systems disclosed herein.

fyySOJ It should h understood -that arrangements described herei are for purposes of example only. As such, those skilled in the ait will appreciate that other arrangements and other elements {e.g. machines, apparatuses, interfaces, functions, orders, and groupings of functions, etc.) can be used instead, and some elements may he omi tted altogether according to the desired results. Further, many of the elements that are described are functional entities tha may he implemented as discrete or distributed components or in conjunction with other 'components, in any suitable combination ^' and location,

0081 J While various aspects and embodiments have been disclosed herein, other aspects and embodiments will be apparent to those skilled in the art. The various aspects and embodiments disclosed herein are for purposes of illustration and are not intended to be limiting, with the scope being indicated by tlie following claims.