In-vehicle polarimetric radar recognition and / or classification of reflective targets on or near a road

Abstract

A polarimetric radar system is configured to be installed in or on a land vehicle for use with a radar target on a road surface or on a roadside. The radar target is on or near a pathway along which the land vehicle can travel. The polarimetric radar system comprises a radar transmitter, a radar receiver, and a processor. The radar transmitter is configured to transmit sideways from the vehicle one or more polarized transmit signals. The radar receiver is configured to receive sideways from the vehicle one or more polarized reflection signals. The processor, which is coupled to the radar receiver, is configured to determine, based on at least one of said one or more polarized reflection signals, a classification or a type of the radar target.

Description

IN-VEHICLE POLARIMETRIC RADAR RECOGNITION AND / OR CLASSIFICATION OF REFLECTIVE TARGETS ON OR NEAR A ROAD RELATED APPLICATIONS

[0001] This application claims priority to the following United States provisional patent applications: (1) No.63 / 683,882 filed August 16, 2024, entitled “Radar-Based Vehicle Guidance Using Reflectivity of Targets on or Near the Driving Surface”; (2) No.63 / 746,562 filed January 17, 2025 entitled “Vehicle Side-Looking Radar for Guidance and Control”; (3) No. 63 / 807,377 filed May 16, 2025, entitled “Vehicle Guidance Using Polarimetric Radar With Reflective Targets on or Near a Driving Surface”; and (4) No.63 / 862,543 filed August 12, 2025, entitled “Radar-Based Vehicle Guidance Using Series of Short-Segment Arrays of Dihedral Corner Reflectors on a Driving Surface.” The entire disclosures of each of the foregoing priority patent applications are hereby incorporated by reference herein.

[0002] The subject matter of this application also relates to the following patents and patent applications filed by the Applicant: (1) U.S. Patent No. 11,400,956, granted Aug. 2, 2022, entitled “Vehicle Guidance System”; (2) International Patent Appl. No. PCT / US2023 / 028710, entitled “Radar-Based Vehicle Guidance Using Radar Reflectors Embedded in Pavement Markers on a Driving Surface,” filed July 26, 2023; (2) International Patent Appl. No. PCT / US2024 / 027899, entitled “Radar-Based Vehicle Guidance Using Arrays of Reflectors Embedded in or Under a Road Line on a Driving Surface,” filed May 4, 2023; (3) International Patent Appl. No. PCT / US2024 / 050650, Filed October 10, 2024, entitled “Vehicle Guidance Using an IMU and Radar With Reflective Targets on or Near a Driving Surface”; (4) U.S. Provisional Patent Appl. No.63 / 761,223, filed February 21, 2025, entitled “Vehicle Guidance Using a Downward Side-Looking Radar With Road Surface Null Targets on or Near a Driving Surface”; and (5) U.S. Provisional Patent Appl. No.63 / 836,441, filed February 21, 2025, entitled “Radar-Based Lane-Departure Warning.” The entire disclosures of each of the foregoing patents and patent applications are hereby incorporated by reference herein. 50066 / 0501 Page 1 of 64TECHNICAL FIELD

[0003] The field of this disclosure relates generally to control of an automobile or similar land vehicle. More specifically, this disclosure relates to the use of radar (radio detection and ranging) reflectivity of targets on or near a roadway or other driving surface to control or assist in controlling a vehicle, especially its lateral positioning. BACKGROUND INFORMATION

[0004] Radar has been used with automobiles for various purposes such as blind-spot detection and forward-collision warning. However, radar is not conventionally used to perform or assist in performing lateral steering of an automobile. Instead, electronic lateral steering or steering assistance systems typically collect and process visual data. That approach is complicated and limited or unusable in inclement weather (e.g., rain, snow, ice, fog, smoke, and darkness).

[0005] Road lines and pavement markers are commonly deployed on or near a driving surface to demark road or lane boundaries. Curbs, concrete barriers, and other roadside objects also sometimes demark the edge of a road or lane. OVERVIEW OF THE DISCLOSURE

[0006] According to one embodiment, a polarimetric radar system is configured to be installed in or on a land vehicle for use with a radar target on a road surface or on a roadside. The radar target is on or near a pathway along which the land vehicle can travel. The polarimetric radar system comprises a radar transmitter, a radar receiver, and a processor. The radar transmitter is configured to transmit sideways from the vehicle one or more polarized transmit signals. The radar receiver is configured to receive sideways from the vehicle one or more polarized reflection signals. The processor, which is coupled to the radar receiver, is configured to determine, based on at least one of said one or more polarized reflection signals, a classification of the radar target.

[0007] The radar transmitter may be configured to transmit the one or more polarized transmit signals partially downward, and wherein the radar receiver may be configured to receive the one or more polarized reflection signals directed partially upward. 50066 / 0501 Page 2 of 64

[0008] The polarized transmit signals and the polarized reflection signals may be of a type selected from a group consisting of vertically polarized, horizontally polarized, orthogonally polarized, circularly polarized, and elliptically polarized.

[0009] Optionally, the radar transmitter, which is characterized by a transmission beam pattern having a center direction of maximum transmission strength, is configured to be installed in or on the land vehicle so that the center direction of maximum transmission strength is downward and sideways from the land vehicle when the radar transmitter is installed on or in the land vehicle, and the radar receiver, which is characterized by a reception beam pattern having a center direction of maximum reception strength, is configured to be installed in or on the land vehicle so that the center direction of maximum reception strength is upward and sideways generally toward the radar receiver when the radar receiver is installed on or in the land vehicle.

[0010] In one example, the radar target is on the road surface and selected from a group consisting of a raised pavement marker, a snow-plowable pavement marker, a slot in the road surface, a groove in the road surface, a rumble strip, a bott dot, a cat’s eye marker, a curb, a bottom of a concrete barrier, a road surface edge, a vegetation edge, a drain, and a road stud.

[0011] In another example, the radar target is a roadside target selected from a group consisting of a concrete barriers, a guardrail, a building, a sign, a sign post, a lamp, a lamp pos, a fire hydrant, a utility pole, a tree, tall grass, and a vehicle.

[0012] The radar transmitter and the radar receiver may operate, for example, within a frequency range from about 300 MHz to about 1 THz.

[0013] The radar transmitter and the radar receiver, may be, for example, of a frequency-modulated continuous wave type.

[0014] The radar transmitter and the radar receiver may be monostatic or, alternatively, bi-static.

[0015] Optionally, the polarimetric radar system further comprises an antenna array. 50066 / 0501 Page 3 of 64

[0016] Optionally, the radar transmitter and the radar receiver are configured to be placed within a front bumper of the vehicle.

[0017] The radar transmitter and the radar receiver may be packaged within a vehicle light assembly.

[0018] The signal processor may be configured to process just a portion of said at least one of said one or more polarized reflection signals received in a window corresponding to an expected range of a reflection target.

[0019] Optionally, information regarding the radar target is stored in a database, and the system is configured to determine, based on the classification of the radar target and information in the database, a lateral distance between the vehicle and the radar target, wherein the system is further configured to determine a lateral position of the vehicle relative to the pathway based on said lateral distance between the vehicle and the radar target.

[0020] The information regarding the radar target in the database may comprise location information, such as, for example, absolute coordinates, or, alternatively, relative location information with respect to the pathway or a feature of the pathway.

[0021] Information regarding the radar target may be recorded in an automotive navigation database, and the system may be further configured to receive vehicle trajectory information from a GNSS sensor on or in the vehicle, and the system may be configured to determine, based on the classification of the radar target, GNSS-measured coordinates, and information in the database, a lateral distance between the vehicle and the radar target and a direction of vehicle travel.

[0022] According to another embodiment, a method recognizes a radar target on or near a pathway along which a land vehicle travels. The radar targets have different radar reflectivity characteristics compared to the pathway. The radar targets are arranged along the pathway in a direction of vehicle travel. The method transmits sideways at least one polarized radar transmit signal from the land vehicle towards a radar target as the vehicle passes the radar target, receives at the land vehicle at least one polarized radar reflection signal from the radar target as the vehicle passes by the radar target, and determines, based 50066 / 0501 Page 4 of 64on said at least one polarized radar transmit signal and said at least one polarize radar reflection signal, a type of the radar target.

[0023] The steps of transmitting, receiving, and determining may be repeated for substantially every radar target that the vehicle passes.

[0024] The pathway may be one lane of a multi-lane road, and the radar targets may, for example, comprise pavement markers that are distributed along a line demarking one side of the pathway.

[0025] The method may optionally further filter said at least one polarized radar reflection signal based on a range of expected distances between the land vehicle and the radar target.

[0026] The method may optionally further determine, based on one or more said at least one polarized radar transmit signal and one or more of said at least one polarize radar reflection signal, a distance between the vehicle and the radar target.

[0027] The method may optionally further use the distance to guide or assist steering the vehicle laterally to follow the pathway.

[0028] The method may optionally further provide a warning either if the distance exceeds a maximum threshold or is less than a minimum threshold, or if an approximate lateral position of the vehicle is outside of an acceptable range.

[0029] The warning may be an internal warning to a driver of the vehicle or, alternatively, a warning external to the vehicle.

[0030] The method may optionally further perform one or more of a lane-centering assistance function and a lane-keeping function based on the distance.

[0031] The vehicle may comprise a lateral-positioning system independent of the method of claim 24, and the method may further use the distance in conjunction with the independent lateral-positioning system to improve lateral-position determination. The lateral-positioning system may, for example, comprise a vision system viewing painted lines along the pathway. The lateral-positioning system may, for example, comprise a lidar system viewing painted lines along the pathway, recorded targets along the pathway, or both. 50066 / 0501 Page 5 of 64

[0032] The method may optionally further use the classification of the radar target to access information about the radar target in a database.

[0033] The method may optionally further receive GNSS location information for the vehicle and use the GNSS location information, in conjunction with the classification of the radar target, to access the information about the radar target in the database. The information in the database may, for example comprise location information for the radar target.

[0034] The method may optionally further determine, based on the location information for the radar target in the database and a radar-measured distance between the vehicle and the radar target, a lateral position of the vehicle relative to the pathway or a feature of the pathway.

[0035] As one skilled in the art will appreciate in light of this disclosure, certain embodiments of the pavement markers, radar reflectors, and radar disclosed herein may be capable of achieving certain advantages, which will be apparent from the following detailed description of example embodiments, which proceeds with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1A is a front-view diagram of a vehicle equipped with two polarimetric radar transceivers interacting with road surface side and roadside targets on a driving surface of the road, according to one embodiment.

[0037] Figure 1B is a front-view diagram of a vehicle equipped with one polarimetric radar transceiver mounted in the center of the vehicle interacting with road surface side and roadside targets on a driving surface of the road, according to one embodiment.

[0038] Figure 1C is a block diagram of a radar transceiver with circularly polarized and vertically polarized antennas viewing a flat target, according to one embodiment.

[0039] Figure 2A is a front-view diagram of a polarimetric radar transceiver with a circularly polarized antenna viewing a road surface target, such as a raised pavement marker, according to one embodiment. 50066 / 0501 Page 6 of 64

[0040] Figure 2B is a front-view diagram of a polarimetric radar transceiver with a circularly polarized antenna viewing a road surface, according to one embodiment.

[0041] Figure 2C shows relative signal return strength vs. distance for a radar transceiver with circularly and vertically polarized antennas viewing an asphalt road surface and a dihedral corner reflector.

[0042] Figure 3A is a block diagram of a radar transceiver with an AI (Artificial Intelligence) model, according to one embodiment.

[0043] Figure 3B is a block diagram of AI model processing for determining the radar target, according to one embodiment.

[0044] Figure 4A is an isometric diagram of a conventional pavement marker.

[0045] Figure 4B is a top view of a segment of a road on which conventional pavement markers have been placed.

[0046] Figure 5A is a front view of a pavement marker illustrating a radar signal from and a reflection back to a radar transceiver, according to one embodiment.

[0047] Figure 5B is a front view of another pavement marker illustrating a radar signal from and a reflection back to a radar transceiver, according to one embodiment.

[0048] Figure 6A is an isometric diagram of a dihedral corner reflector.

[0049] Figure 6B is an isometric diagram of a pavement marker augmented with two dihedral corner reflectors, according to one embodiment.

[0050] Figure 6C is a top view of the pavement marker of Figure 6B.

[0051] Figure 6D is a side view of the pavement marker of Figure 6B.

[0052] Figure 7A is a cross-section view of a cavity in or below a driving surface, according to one embodiment.

[0053] Figure 7B is a top view of a section of a road in which cavities of Figure 7A have been arranged, according to one embodiment.

[0054] Figure 7C is a top view of a section of a road in which long cavities of Figure 7A have been arranged, according to another embodiment. 50066 / 0501 Page 7 of 64

[0055] Figure 8A is a cross-section view of a pavement marker in a cavity below a driving surface illustrating a radar signal from and a reflection back to a polarimetric radar transceiver, according to one embodiment.

[0056] Figure 8B is a top view of a section of a road in which pavement markers of Figure 8A have been arranged, according to one embodiment.

[0057] Figure 8C is a top view of a section of a road in which pavement markers of Figure 8A have been arranged, according to another embodiment.

[0058] Figure 9 is an isometric view of a snow-plowable pavement marker, according to one embodiment.

[0059] Figure 10A is an isometric view of a curb illustrating a radar signal from and a reflection back to a polarimetric radar transceiver, according to one embodiment.

[0060] Figure 10B another curb illustrating a radar signal from and a reflection back to a polarimetric radar transceiver, according to one embodiment.

[0061] Figure 10C is a front view of a concrete barrier illustrating a radar signal from and a reflection back to a radar transceiver, according to one embodiment.

[0062] Figure 10D is a diagram illustrating pavement markers and roadside vegetation, according to one embodiment.

[0063] Figure 10E is a diagram illustrating pavement markers and a roadside sign, according to one embodiment.

[0064] Figure 10F is a diagram illustrating street curbs, according to one embodiment.

[0065] Figure 10G is a side view of a concrete barrier illustrating a radar signal from and a reflection back to a polarimetric radar transceiver, according to one embodiment.

[0066] Figure 11A is a top-view diagram of two vehicles equipped with polarimetric radar transceivers, illustrating range-based gating interacting with radar-reflective pavement markers and other target structures on a multi-lane road, according to one embodiment. 50066 / 0501 Page 8 of 64

[0067] Figure 11B is a top-view diagram of two vehicles equipped with polarimetric radar transceivers, illustrating range-based gating with vehicle dynamics modelling interacting with radar-reflective pavement markers and other target structures on a multi- lane road, according to one embodiment.

[0068] Figure 11C is a top-view diagram of two vehicles equipped with polarimetric radar transceivers illustrating time-based gating interacting with radar-reflective pavement markers and other target structures on a multi-lane road, according to one embodiment.

[0069] Figure 12A is a plot of a received radar distance vs. time, illustrating amplitude-based and range-based gating, according to one embodiment.

[0070] Figure 12B is a plot of a received radar amplitude vs. time, illustrating amplitude-based and range-based gating, according to one embodiment.

[0071] Figure 12C is a plot of filtered received distance vs. time, illustrating amplitude-based and range-based gating, according to one embodiment.

[0072] Figure 13A is a top view of a vehicle traveling along a roadway, illustrating the vehicle-roadway geometry and the use of a single radar, according to one embodiment.

[0073] Figure 13B is a top view of a vehicle traveling along a roadway, illustrating the vehicle-roadway geometry and the use of a single radar, according to one embodiment.

[0074] Figure 14A is a front view of vehicles, one of which is equipped with a center- mounted polarimetric radar transceiver and GNSS sensor on a road having radar-reflective markers and other target structures, according to one embodiment.

[0075] Figure 14B is a top view of vehicles equipped with radar transceivers on a road having radar-reflective markers and other target structures, according to one embodiment.

[0076] Figure 15A is a table illustrating content of a database of information on road surface targets, according to one embodiment.

[0077] Figure 15B is a table illustrating content of a database of high-definition road surface and roadside targets, according to one embodiment. 50066 / 0501 Page 9 of 64

[0078] Figure 15C is a table illustrating content of a navigational database and heading information, according to one embodiment.

[0079] Figure 16A is a block diagram of a polarimetric radar-based LDW / LKA / LKS system for a vehicle, according to one embodiment.

[0080] Figure 16B is a block diagram of a two-polarimetric radar-based system, each measuring distance to road surface targets, for a LDW / LKA / LKS system for a vehicle, according to one embodiment.

[0081] Figure 17 is a block diagram of a polarimetric radar-based LDW / LKA / LKS system for a vehicle illustrating integration with a navigational map, according to another embodiment.

[0082] Figure 18 is a block diagram of a two-polarimetric radar-based LDW / LKA / LKS system for a vehicle illustrating the integration with a road reference heading map, according to another embodiment.

[0083] Figure 19 is a block diagram of a lateral controller system utilizing polarimetric radar transceivers and a navigational map for steering control of an autonomous vehicle, according to one embodiment.

[0084] Figure 20 is a block diagram of a lateral controller system utilizing polarimetric radar transceivers and a high-definition map for steering control of an autonomous vehicle, according to one embodiment.

[0085] Figure 21 is a top view of a vehicle equipped with radar transceivers, GNSS sensors, and inertial measurement units on a road having pavement markers, according to one embodiment.

[0086] Figure 22 is a block diagram of a system that combines multiple detection, guidance, and / or control systems, including a polarimetric radar-based lateral positioning system, according to one embodiment. 50066 / 0501 Page 10 of 64DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS Preliminary Notes

[0087] Example embodiments are described below with reference to the accompanying drawings. Unless otherwise expressly stated, the sizes, positions, etc., of components, features, elements, etc., as well as any distances therebetween, in the drawings are not necessarily to scale, and may be disproportionate and / or exaggerated for clarity.

[0088] The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting. As used herein, the singular forms “a,” “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should be recognized that the terms “comprise,” “comprises,” “comprising,” “include,” “includes,” “including,” “has,” “have,” and “having,” when used in this document, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. Unless otherwise specified, a range of values, when recited, includes both the upper and lower limits of the range, as well as any sub-ranges therebetween. Unless indicated otherwise, terms such as “first,” “second,” etc., are only used to distinguish one element from another and not to imply any relative order, placement, or ranking. For example, one element could be termed a “first element,” and similarly, another element could be termed a “second element,” or vice versa. The section headings used herein are for organizational purposes only and are not to be construed as limiting the subject matter described.

[0089] Unless indicated otherwise, the terms “about,” “thereabout,” “substantially,” “generally,” “approximately,” etc., mean that amounts, sizes, formulations, parameters, and other quantities and characteristics are not and need not be exact, but may be approximate and / or larger or smaller, as desired, reflecting tolerances, conversion factors, rounding off, measurement error, and the like, and other factors known to those of skill in the art.

[0090] Spatially relative terms, such as “right,” “left,” “forward,” “rearward,” “below,” “beneath,” “lower,” “above,” and “upper,” and the like, may be used herein for ease of description to describe one element’s or feature’s relationship to another element or feature, 50066 / 0501 Page 11 of 64as illustrated in the drawings. It should be recognized that the spatially relative terms are intended to encompass different orientations in addition to the orientation depicted in the drawings. For example, if an object in the figures is turned over, elements described as “below” or “beneath” other elements or features would then be oriented “above” the other elements or features. Thus, the term “below” can, for example, encompass both an orientation of above and below. An object may be otherwise oriented (e.g., rotated 90° or at other orientations), and the spatially relative descriptors used herein may be interpreted accordingly.

[0091] The terms “RF” and “radio” are used herein to refer to or imply electromagnetic radiation in the frequency range that can be detected and processed electronically. For example, the RF spectrum encompasses very low frequencies (VLF, e.g., approximately 3 kHz) and may extend to just below the infrared range (e.g., approximately 300 THz). The RF spectrum includes VLF, low frequency (LF), medium frequency (MF), high frequency (HF), very high frequency (VHF), ultra-high frequency (UHF), super high frequency (SHF), and extremely high frequency (EHF) bands. The RF spectrum also includes microwave and millimeter-wave bands.

[0092] The concepts and innovations described herein with reference to RF or radio radiation and / or radar can also be practiced with other types of radiation or waves, including, for example, infrared, visible, ultraviolet, laser, lidar, and sonar, with suitable modifications to the transceivers and reflectors.

[0093] Unless clearly indicated otherwise, all connections and couplings may be direct (without intermediaries) or indirect (with one or more intermediaries). All operative connections and couplings, unless clearly indicated otherwise, may be electronic via hardware or logical via software. Similarly, unless clearly indicated otherwise, all physical connections and couplings may be rigid or non-rigid, permanent or detachable.

[0094] Like numbers refer to like elements throughout. Thus, the same or similar numbers may be described with reference to other drawings even if they are neither mentioned nor described in the corresponding drawing. Also, even elements that are not denoted by reference numbers may be described with reference to other drawings. 50066 / 0501 Page 12 of 64Additionally, the drawings may include non-essential elements that are included only for the sake of thoroughness. These non-essential elements may be removed entirely or left only in outline form if drawing changes are desired to create greater clarity.

[0095] The embodiments described herein are merely examples, set forth by way of illustration only and not limitation. Those skilled in the art will recognize in light of the teachings herein that there are alternatives, variations and equivalents to the example embodiments described herein and their component parts. For example, other embodiments are readily possible, variations can be made to the embodiments described herein, and there may be equivalents to the components, parts, or steps that make up the described embodiments.

[0096] For the sake of clarity and conciseness, certain aspects of components or steps of certain embodiments are presented without undue detail where such detail would be apparent to those skilled in the art in light of the teachings herein and / or where such detail would obfuscate an understanding of more pertinent aspects of the embodiments.

[0097] The foregoing is illustrative of embodiments of the invention and is not to be construed as limiting thereof. Although a few specific example embodiments have been described, those skilled in the art will readily appreciate that many modifications to the disclosed example embodiments, as well as other embodiments, are possible without materially departing from the novel teachings and advantages of the invention. Accordingly, all such modifications are intended to be included within the scope of the invention as defined in the claims. For example, skilled persons will appreciate that the subject matter of any sentence or paragraph can be combined with subject matter of some or all of the other sentences or paragraphs, except where such combinations are mutually exclusive.

[0098] It will be clear to those having skill in the art that many changes may be made to the details of the above-described embodiments without departing from the underlying principles of the inventions described herein. Downward Side-Looking Polarimetric Radar(s)

[0099] Figure 1A and 1B show two vehicles, 180 and 182, on a typical roadway 300 using polarimetric radars, and Figure 1C illustrates one embodiment of a polarimetric radar. 50066 / 0501 Page 13 of 64In Figure 1A, the vehicle 180 has two side polarimetric radar transceivers 116 and 117 to detect targets on the road surface, e.g., pavement markers, slots in the road, rumble strips, bot dots, grooves in the road, street curbs, etc. and roadside structures such as, for example, road signs, concrete barriers, guardrails, mailboxes, trees, fire hydrants, lamp posts, etc. As illustrated, the polarimetric radar transceiver 117 on the vehicle 180 looks downward and to the side to see a street curb 1420 and looks sideways to view a road sign 1306. The polarimetric radar transceiver 116 on the vehicle 180 views a pavement marker 445 in the center of the roadway 300 and views the vehicle 182 on its side. The polarimetric radar transceiver 116 uses a radar beam 1108 to view the pavement marker 445 and a radar beam 1106 to view the vehicle 182, as illustrated. Similarly, the polarimetric radar transceiver 117 uses a radar beam 1109 to view the street curb and a radar beam 1107 to view the road sign 1306. Alternatively, the polarimetric radar transceiver 116 may have a single wider beam that sees both vehicle 182 and the pavement marker 445, and the polarimetric radar transceiver 117 may have a single wider beam that encompasses both the beam 1107 and the beam 1109. Whether the polarimetric radar transceivers 116 and 117 have single beams or multiple beams, and their beamwidths are parameters that those skilled in the art can select as desired.

[0100] The radar beams 1106, 1107, 1108, and 1109 may be fixed looking to the sides and downward, may alternatively view the targets, may be steered between the different targets, or may be adjusted according to a database of roadside targets. For example, information about a target in a database may inform which radar signal parameters to use to best detect the target. Examples of such parameters include beamwidth, power, polarization, and pointing direction. The polarimetric radar transceivers 116 and 117 may view many different roadside targets and track them simultaneously or effectively simultaneously on or near the road via looking downward and to the side, near the side of the road via looking downward and to the side, and near the side of the road via looking only to the side. The polarimetric radar transceiver 116 on the vehicle 180 may also facilitate determination of a lateral distance 1811 between the vehicles 180 and 182 when the vehicle 180 passes the vehicle 182 or vice versa. The vehicle 182 would typically block the view of the polarimetric radar transceiver 116 on the vehicle 180 from viewing far roadside targets, 50066 / 0501 Page 14 of 64e.g., a concrete barrier. The lateral distance 1811 information, when available, in addition to viewing fixed roadside targets, can provide a more robust control system.

[0101] Figures 1A and 1B also show other sensors such as IMUs (Inertial Measurement Units) 907, a GNSS (Global Navigation Satellite System) receiver 901, and vision sensors 903. These sensors can provide lateral and longitudinal distance information when the radar transceivers 116, 117, and 119 do not view a target.

[0102] Figure 1B illustrates the mounting of a polarimetric radar transceiver 119 in the center front of the vehicle 180 to view the road and roadside targets from two antennas viewing both directions (i.e., opposite lateral directions, right and left or driver-side and passenger-side). In this embodiment, the polarimetric radar transceiver 119 uses radar beams 1106-1109 similar to those in Figure 1A to view the targets. Two advantages of this mounting scheme over mounting separate polarimetric radar transceivers 116 and 117 on the front sides are that (1) having only one unit saves costs and (2) the viewing distances allows the vehicle 180 to cross a target (e.g., pavement marker 445) on or in or in-line with a lane-boundary line while still tracking the target.

[0103] The polarimetric radar transceivers 116 and 117 can simultaneously view multiple road surface targets and track them effectively on or near the road by looking downward and to the side with various polarized antennas, including vertical, horizontal, orthogonal, circular, and elliptical polarizations.

[0104] Figure 1C illustrates an embodiment of a polarimetric radar transceiver 116 with circular polarization antennas 736, 756, and 766, horizontal polarization antennas 726 and 746, and vertical polarization antennas 716 and 776. Figure 1C is a block diagram of a FMCW (Frequency Modulated Continuous Wave) polarimetric radar transceiver with the same right-handedness transmit and receive circular polarized antennas 736 and 756, respectively, and transmit and receive vertical polarized antennas 716 and 776, respectively, transmitting towards the flat surface 3300 and reflecting to the polarimetric radar transceiver 116. As shown in Figure 1C, the transmitted RHCP (Right-Handed Circularly Polarized) signal 7316 from RHCP antenna 736 reflects from the flat surface 3300 as an LHCP (Left-Handed Circularly Polarized) signal 7416 to LHCP antenna 766. The transmitted 50066 / 0501 Page 15 of 64vertically polarized signal 7116 from the vertically polarized antenna 716 reflects as a vertically polarized signal 7216 to the vertically polarized antenna 776. However, the RHCP received antenna 756 receives little reflected signal since it is oppositely polarized.

[0105] Polarimetric radar transceivers 116, 117, and 119 are defined as a type of radar transceiver that has one or more transmitting antennas with the same or different polarizations and one or more receiving antennas with the same or different polarizations. The other polarization states are defined as linear polarization (e.g., vertical, horizontal, or orthogonal), circular polarization (e.g., RHCP or LHCP), and elliptical polarization (e.g., right- hand elliptical polarization (RHEP) and left-hand elliptical polarization (LHEP)). A polarimetric radar transceiver can provide more detailed information about the size, shape, and orientation of targets compared to conventional automotive radars, which typically only measure the range, amplitude, radial velocity, elevation, and azimuth of the returned signal. A polarimetric radar transceiver may be any radar transceiver type, e.g., FMCW, PMCW (Phase-Modulation Continuous Wave), OFDM (Orthogonal Frequency Division Multiplexing), DCM (Digital Code Modulation), multiple-frequency radar, multiple-frequency phase radar, MIMO (Multi-Input Multi-Output) radar, UWB (Ultra-Wide Band) FMCW, UWB impulse radar, FSK (Frequency Shift Keying) radar, monopulse radar, or a combination of any of these types. A polarimetric radar transceiver preferably operates at radio frequencies between 0.3 GHz and 1 THz (300 MHz to 1 THz), but more specifically at radar frequencies: 76-81 GHz, 110-170 GHz, and 220-325 GHz.

[0106] Figures 2A-2C illustrate various configurations of polarimetric radar transceivers viewing a road surface target 445 with sharp corners 448 and the road surface 300. If the polarimetric radar transceiver 116 has vertical-polarized antennas 716, the reflection polarization from the road surface targets 445 and the road surface 300 remains unchanged. No attenuation due to antenna polarization occurs. For example, as shown in Figure 2C, if the road surface target 445 has an aluminum 25 mm dihedral corner reflector, and the road surface 300 is an asphalt road, the separation of received signal amplitude reflected from the dihedral corner reflector is about 20 dB greater than that from the road surface 300. However, if the polarimetric radar transceiver 116 has RHCP antennas 736 and 756 for both transmit and receive antennas, respectively, as shown in Figure 2B, the received 50066 / 0501 Page 16 of 64signal polarity changes upon reflecting from the road surface 300, from a transmitted RHCP signal to a LHCP signal. This is noted as a single bounce, and the circular polarized signal changes polarity from right-hand to left-hand circular polarization or vice versa. However, as shown in Figure 2A for a road surface target 445, the received signal polarity for a sharp corner 448, especially a metallic dihedral corner, changes twice from the sharp corner 448 to the road surface and then returns to the same circular polarity as the transmitted signal. In this example, the received signal is RHCP. As shown in Figure 2A, the polarimetric radar transceiver 116 transmits using an RHCP signal with an RHCP antenna 736. The RHCP signal reflects from the ground surface 300 as an LHCP signal to the side of the sharp corner 448 of the road surface target 445 and changes polarity from LHCP to RHCP from the target side back to the polarimetric radar transceiver 116. The reflection off the sharp corner is noted as a double bounce signal, and the circular polarity comes back as the same as the transmit polarity.

[0107] As shown in Figure 2B, the transmitted RHCP signal reflects from the road surface 300 as an LHCP signal. In theoretical terms, if the received and transmit antennas 736 and 756, respectively, are perfect RHCP antennas, the attenuation is infinite for an LHCP signal. The polarization is the opposite, and therefore, the received LHCP signal is not received by the RHCP antenna 756. In practice, a perfect RHCP antenna does not exist, and the attenuation for practical antennas, depending on the design, is typically around 10-30 dB for an LHCP signal. As shown in Figure 2C, with vertically polarized antennas, the relative signal return amplitude from the asphalt road surface is approximately 80 dB. The returned signal amplitude depends on several parameters, including radar processing parameters, radar processing software, road surface roughness, incident angle, and beamwidth. So, this value is relative to a particular set of test conditions. With circularly polarized antennas using the same polarity for transmit and receive antennas, and the same radar test conditions as the vertical polarized antenna setup, the signal return from the asphalt road surface is approximately 60 dB in this example. For example, as shown in Figure 2C, the relative received road surface signal using vertical polarization is approximately 80 dB, while using circular polarization, it is approximately 60 dB, for a distance of approximately 0.5 meters. The relative signal return amplitude from the metallic dihedral corner reflector 50066 / 0501 Page 17 of 64is similar in both polarizations, vertical and circular, and is around 100 dB at a 0.5-meter distance in this example. The signal-to-noise ratio (S / N) is approximately 20 dB for the vertical polarization antennas, but about 40 dB for the circular polarization antennas, as illustrated in this example. With the higher S / N, the targets on the road surface are more visible to the polarimetric radar transceiver. In this example, the S / N is approximately 62 dB (100 dB signal to 38 dB noise). This value can be achieved if there were no road return backscatter or if the radar were transmitting in the air with no other targets present. The backscattering from the road surface is the dominant factor. In another embodiment, elliptical polarization antennas are used, and the RHEP wave reflects from the road surface 300 as a LHEP wave.

[0108] Using a polarimetric radar transceiver provides a significant advantage in processing road surface targets with sharp corners compared to the road surface itself. Some less RF-reflective targets, such as conventional, prior-art plastic pavement markers, which have a low S / N when using vertically polarized antennas, may not be discernible to the radar transceiver because the signal return from the plastic pavement marker and the road surface signal return are similar in amplitude. However, the plastic pavement markers can be detected by the polarimetric radar transceiver 116 using circular polarization antennas 736 and 756, with a S / N of around 10-30 dB. This is one of the advantages of using a polarimetric radar transceiver, which provides a higher S / N for road surface targets with corners on a rough, flat road surface. Another advantage of using polarimetric radar is to prevent interference between passing vehicles on a multiple-lane highway by using a polarimetric radar transceiver 116 with RHCP antennas on the passenger’s side of the vehicle and a polarimetric radar transceiver 117 using LHCP antennas on the driver’s side.

[0109] Another advantage of using a polarimetric radar is that it can transmit different polarizations at the road surface to determine the characteristics of the road surface and the road conditions, such as snow, wet, dry, or icy.

[0110] Another advantage of the polarimetric radar is that it can better determine the size, shape, orientation, type, and material properties of the target based on the polarization information. In contrast, an automotive radar with vertically polarized antennas measures range, azimuth, elevation, and radial velocity. A polarimetric radar transceiver 116 can detect 50066 / 0501 Page 18 of 64the target and its characteristics by determining scattering coefficients, which may be included in a polarization scattering matrix S, such as the following:

[0111] The polarimetric radar transceiver 116 in Figure 1C transmits both vertically polarized wave 7116 (Ev) and horizontally polarized wave (not shown, Eh) via polarized antennas 716 and 726, respectively. The reflected or scattered radar waves 7216 (Eh and Ev) from the flat plate 3300 are received by vertically and horizontally polarized antennas, 776 and 746, respectively. In this way, each scattering object, in this example, a flat plate 3300, is considered a polarization transformer, and the transformation from a transmitted wave to a received wave can be described as applying a scattering matrix to the signal representing the incident radar wave. Circular and elliptical polarization waves can produce a similar matrix by transmitting both RHCP and LHCP, or RHEP and LHEP, signals and receiving both RHCP and LHCP, or RHEP and LHEP, signals.

[0112] A scattering matrix, for a given set of incident and reflected radar signals and other relevant parameters (e.g., frequency and orientations of radar and target) contains information concerning the scattering properties of the target. For example, a scattering matrix is used to deterministically describe the polarization state of the backscattered radiation from a target by representing the polarimetric backscattering coefficients of the two co-polarized and the two cross-polarized signals as stated above. The diagonal matrix elements are usually called co-polarized, and the non-diagonal elements are typically called cross-polarized. A scattering matrix may contain all the relevant information about the scattering process and the scatterer itself. Elements of the scattering matrix or an equivalent matrix, for instance, the known Covariance matrix and the Coherency matrix, are observable power terms. Different relevant matrix formalisms are used in radar polarimetry, including the Jones Matrix, S-matrix, Müller M-matrix, and Kennaugh K-matrix. By measuring the 50066 / 0501 Page 19 of 64scattering matrix, its elements, or an equivalent, the strength and polarization of the scattered radar wave for an arbitrary polarization of the incident wave can be computed. Decomposition of the polarimetric scattering matrix can define the type of object. Objects can be classified as a single bounce (odd), e.g., a flat plate, trihedral corner reflectors, and spheres, as a double bounce (even), e.g., dihedral corner reflectors viewed between 0-90 degrees, as a cross-polar scattering, e.g., dipole or dihedral viewed rotated by 45 degrees, as an asymmetric back scattering, e.g., grass or a combination depending on the radar beam viewing angle.

[0113] Figure 3A illustrates an embodiment of the polarimetric radar 116, polarization determination 9013, and an AI model 9010 that processes the radar information received from the polarized antennas 700 to classify the target at the output 9100. The polarimetric radar transceiver 116 provides lateral distance, target amplitude, target elevation, velocity, and azimuth of the road surface or roadside target, and polarized data. The AI model 9010 utilizes machine learning and / or deep-learning techniques to distinguish different roadside targets from the 4D information from the radar transceiver, e.g., range, velocity, azimuth, and elevation, and from the polarization scattering matrix. Road surface and roadside targets can be classified with 4D information and polarization scattering matrix. As shown in Figure 3A, the polarimetric radar 116 measures the polarization characteristics of the reflected radar signal using various polarized antennas 700, providing information about the object’s shape, orientation, and material properties. This information can be represented in forms such as the scattering matrix (S-matrix) or the polarimetric coherence matrix (T-matrix) 9015. To extract meaningful features, polarimetric decomposition methods, 9017 as Freeman-Durden or reflection symmetric decomposition (RSD), can be applied to break down the radar signal into components representing different scattering mechanisms (e.g., surface scattering, double bounce, volume scattering). Features, like the power values of different polarization channels (horizontal, vertical, left-hand, right- hand) or their combinations, derived from the polarimetric data, are used as inputs for machine learning models as shown in Figure 3A. AI models 9010 utilize machine learning approaches, including traditional machine learning algorithms such as Support Vector Machines (SVM), Decision Trees, and Random Forests, which can be trained on polarimetric 50066 / 0501 Page 20 of 64radar features to classify objects. Deep learning, particularly Convolutional Neural Networks (CNNs), excels at automatically extracting hierarchical features and has shown significant potential in classifying objects from polarimetric radar images. Figure 3B demonstrates a deep neural network architecture. Combining polarimetric features extracted through decomposition with deep learning models (e.g., CNNs or Long Short-Term Memory networks) is another embodiment which will enhance the overall accuracy on classifying targets.

[0114] Figure 3A shows an embodiment using the AI model 9010 with radar sensor signal processing blocks 9016 and polarization processing blocks 9015 & 9017. The polarimetric radar transceivers 116, 117, or 119 can determine the distance, velocity, amplitude, azimuth angle, and elevation angle to the road surface or roadside target by using well-known signal processing techniques. One technique is to apply the FFT (Fast Fourier Transform) to the ADC (Analog-to-Digital Converter) data to provide a range versus amplitude. The second stage of this embodiment, as illustrated in Figure 3A, utilizes the AI model 9010. The model 9010 extracts data via block 9011 from both the radar data and the polarimetric data, and classifies the roadside target via block 9012. The radar data provided to the AI model 9010 can be 1D (range vs. amplitude), 2D (elevation, range, and amplitude), 3D (elevation, azimuth, range, and amplitude) or 4D (elevation, azimuth, range, amplitude, and velocity) with the polarimetric data to classify road surface or roadside targets in addition to providing lateral distance and other information. The azimuth and elevation information can be collected via beam steering from right to left and up and down, using multiple radar antennas or sensors, or employing synthetic aperture radar (SAR) and interferometric synthetic aperture radar (ISAR) techniques as the vehicle passes by the road target. The elevation data may be collected with multiple radar sensors or TX / RX antennas mounted in the vertical direction.

[0115] Figure 3B illustrates an embodiment of a deep-learning model processing radar data through the input layer 9021. This model utilizes convolutional neural networks, recurrent neural networks (RNNs), and max pooling techniques (blocks 9023 through 9029) to extract features. The extracted features are then classified via the 9031 block to determine the roadside target at the output 9100. Road surface and roadside targets are collected to 50066 / 0501 Page 21 of 64train the AI model 9010 and to test the model with real data. Using the AI model 9010, the road surface and roadside targets can be determined, rather than being inferred from the latitude and longitude information in the database. The GNSS sensor 901 may not be necessary since the road surface and roadside targets are identified via the AI model and can then be synchronized with the database, which contains the exact location of the road surface or roadside targets.

[0116] The polarimetric radar transceiver with an AI model can classify, for example, the following types of road surface targets: raised pavement markers, snow-plowable pavement markers, slots in the road, grooves in the road, rumble strips, milled-in rumble strips, bott dots, cat’s eye markers, curbs, the bottom of concrete barriers, road edges, grass edges, drains of the road, drains in the curbs, and road studs. A polarimetric radar transceiver with an AI model can classify, as an example, the following types of roadside targets: sides of concrete barriers, guardrails, buildings, signs, lamp posts, fire hydrants, utility poles, trees, sides of tall grass, and vehicles. Targets on the Surface and the Side of the Road

[0117] Figures 4A and 4B show typical pavement markers 445 that are currently used on today’s roads. The pavement marker 445, shown in Figure 4A, is rectangular; however, pavement markers come in many different shapes and sizes. The pavement marker 445 can be placed on top of the road surface, in which case they are called raised pavement markers (RPMs), or can be installed at or slightly below the pavement surface level (e.g., snow- plowable markers), or below the roadway surface 300 in cut grooves, divots, or slots. The pavement markers 445 are typically installed along or in line with lane-boundary lines, such as painted lines 305, 307, and 309 on the roadway 300, as shown in Figure 4B. The painted line 307 is typically a yellow center line, and the painted white lines 305 and 309 mark the sides of the road. The primary function of the pavement marker 445 is to reflect light waves, allowing the vehicle’s headlights or sunlight to be reflected to the driver, enabling the driver to visually see the location of the road lines, especially in inclement weather. The pavement markers have a visible reflector 275 on one or more sides as shown in Figure 4A. Additionally, the raised pavement reflectors may act as a rumble guard when the vehicle crosses the roadway lines or pavement markers 445, as the markers 445 are typically 25 mm 50066 / 0501 Page 22 of 64above the roadway surface and cause the vehicle’s tires to impact the raised pavement markers. Raised pavement markers have undergone numerous modifications over the years, featuring various visible reflective properties, including cat’s eyes, LED (light-emitting diode) lights, and lidar (light detection and ranging) reflection patterns. The pavement markers 445 are very reflective to visible light, and they may or may not be reflective to radio frequency (RF) radiation coming from radar transceivers. However, they have traditionally not been purposefully designed for RF reflectivity. The pavement markers are typically used on roadways, as well as in parking lots, shipping yards, shipping docks, and other paved and unpaved areas.

[0118] Figures 5A and 5B are front views of pavement markers 445 and 447, respectively, illustrating the polarimetric radar transceiver 116 transmitting RF signals 1119 with different polarizations, typically vertical and circular polarization, and receiving RF signals 1118, which are reflections from the pavement markers 445 and 447. Figure 5A shows the polarimetric radar transceiver 1161 with pavement marker 445, and Figure 5B shows another type of pavement marker 447 with curved sides. Both pavement markers 445 and 447 contain an optical reflector 275. If the polarimetric radar transceiver 116 has only vertical polarized antennas for both transmit and receive antennas, the reflection from the road surface 300, a single reflection, and the pavement marker 445 and 447 sides, double reflection (ground and then the pavement marker side) is received with no polarization attenuation by the antenna since both are vertically polarized. However, as described earlier, if the polarimetric radar transceiver 116 has circular or elliptical polarization antennas for both transmit and receive antennas of same polarity (RHCP, LHCP, RHEP, or LHEP), the reflection from the road surface 300 changes the right-handedness to left- handedness of the circular or elliptical polarization, but the side of the pavement marker 445 and 447 sides provides double-bounce reflection, one from the ground surface 300 and one from the side of the pavement markers 445 and 447. It reflects the same polarity as transmitted. In these two embodiments, as shown in Figure 5A and 5B, the road surface signal is either attenuated or not received by the same-handedness circular-polarity antenna. So, the S / N is much greater using circular or elliptical polarization than using vertical or horizontal polarization. The same polarimetric radar transceiver 116, using 50066 / 0501 Page 23 of 64circularly polarized antennas, views the pavement markers 445 or 447 on the road surface 300 significantly better, using the same radar processing software as the vertically polarized antennas. In Figure 5A, the side is shaped as an angle, and some of the signal reflections 1118 will be sent off past a single antenna radar transceiver receiver. If multiple vertically spaced antennas is used, more of the oblique angle side reflection can be captured. If the side were a right corner angle, the signal reflections 1118 can be captured using a single antenna. In Figure 5B, the side corner of the pavement marker 447 is somewhat rounded on the top of the marker. The return signal from the top is less than the signal return from the low section of the pavement marker 447.

[0119] Figures 6B-6D illustrate a pavement marker 284 in which is embedded a type of radar reflector called a dihedral corner reflector 287, as shown in Figure 6A. A dihedral corner reflector 287 (and 286) consists of two plane reflectors intersecting along an axis where a dihedral angle is formed. The dihedral angle may be 90° or approximately 90° (i.e., at a right angle, orthogonal, perpendicular, or normal). Incident waves entering the aperture so formed with a direction of incidence perpendicular to the edge (i.e., at a direction of incidence that bisects the 90° angle between the two planes), are returned parallel to their direction of incidence. As described earlier, a dihedral corner reflector will not change the handedness of circularly or elliptically polarized waves since it has a double bounce, one from the bottom of the reflector and one from the reflector side. Therefore, using the same right-handed or left-handed circularly, or elliptically polarized antennas for both the transmit and receive antenna(s), there will be no attenuation of the signal of the dihedral corner reflector due to the antenna polarization. However, the road surface (not shown in this figure) changes the handedness of the transmitted signal, and therefore, the received signal from the road surface via the polarized antenna is significantly attenuated. In theory, the signal is attenuated infinitely at the opposite polarization antenna; however, in practice, the incorrect polarization attenuates the signal by approximately 10-30 dB.

[0120] Figure 6B is an isometric view of the pavement marker 284 with dihedral corner reflectors 286 and 287 on opposite sides of the pavement marker 284. The dihedral corner reflectors 286 and 287 can be constructed in various ways. For example, the dihedral corner reflector 287 may be molded or extruded and cut into segments. The dihedral corner 50066 / 0501 Page 24 of 64reflector 287 may be, for example, formed by welding or otherwise connecting two slats or flat plates, or by bending a flat plate. In these embodiments, a circularly or elliptically polarized radar will have a significant S / N for a dihedral corner reflector on the road surface.

[0121] Figure 6B shows the incorporation of the dihedral corner reflectors 286 and 287 within a pavement marker 284, which also has an optical reflector 275 for human vision, cameras, and lidar. In this embodiment, the dihedral corner reflectors 286 and 287 are attached to the sides of the pavement marker 284 with axes of the the dihedral corner reflectors 286 and 287 substantially parallel to the direction of vehicle travel. The method of attachment of the dihedral corner reflectors 286 and 287 to the central body portion of the pavement marker 284 may be, for example, glue, adhesive, snap fit, friction fit, fasteners (e.g., screws, rivets, nails, staples), etc. Optionally, the pavement marker 284 may have a bottom plate (not shown).

[0122] A dihedral corner reflector is a passive device used to reflect radio waves directly back toward the source of emission. The dihedral corner reflectors 286 and 287 can be made of a right-angle metal, such as aluminum, galvanized steel, copper, brass, or magnesium, or any other RF-reflective material. The dihedral corner reflectors 286 and 287 can also consist of metallic or other RF-reflective paint on a non-metallic surface, a metallic or other RF-reflective foil, or electroplating a conductive or other RF-reflective material on a non-metallic or non-RF-reflective surface such as plastic. A dihedral corner reflector can also be formed with one side being a roadway surface (e.g., asphalt, concrete), and the other side being a metallic surface, as the road surface may be sufficiently reflective to RF radiation. For example, when a radar signal is transmitted toward the dihedral corner reflectors 287 downward from the side (right, as illustrated in Figure 6A), and the radar signal is directed at a downward angle less than 45°, e.g., 25° (where perfectly horizontal toward the horizon is considered 0° and vertical straight into the ground is 90°), and the radar source mounted low from the ground, e.g., 0.3 m, the ground surface acts like a smooth reflective surface to the RF waves, and the road surface acts as one side for the dihedral corner reflector. The road surface is more reflective when wet.

[0123] The reflectivity of the road surface can be enhanced by coating portions of the road surface with an RF-reflective adhesive (e.g., conductive epoxy) or RF-reflective paint 50066 / 0501 Page 25 of 64(e.g., copper, carbon, nickel, and silver conductive paint) below and / or around the pavement marker. For example, MG Chemicals 841AR is an acrylic nickel conductive paint, and MG Chemicals 841ER is a nickel conductive epoxy. An RF-reflective adhesive may also be used to adhere the pavement marker, which contains the other surface of the dihedral corner reflector, to the road surface. A dihedral corner reflector can also be formed by painting or coating pertinent sides of an existing pavement marker with perpendicular or nearly perpendicular sides and the road surface near sides with an RF-reflective coating. A dihedral corner reflector can also be formed by painting milled grooves, divots, or cutouts in a roadway surface with RF-reflective coating.

[0124] Pavement marker 284 features an optical reflector 275 similar to current pavement markers, providing optical reflections on the front side that face oncoming vehicles. These optical reflectors in the pavement marker can be used by vision cameras or lidar sensors for further redundancy for lateral guidance.

[0125] Figures 6C and 6D show the top and side views of the pavement marker 284, respectively. The typical dimensions of a current raised pavement marker installed on roadways are 100 mm x 100 mm x 25 mm. Using these dimensions, the dihedral corner reflectors 286 and 287 can be around 25 mm x 25 mm x 100 mm. These dimensions are for typical current raised pavement markers; longer or higher raised pavement markers can be constructed and installed on roadways, which may have a larger radar cross-section (RCS).

[0126] The pavement marker 284 (and the other pavement markers described herein) may be affixed to any driving surface for any vehicle, including roadways such as one-way, bidirectional or multi-lane roads, highways, thoroughfares, lanes, avenues, boulevards, parkways, pathways, byways, streets, paths, trails, and the like, as well as on bridges, in tunnels, parking lots, parking structures, shipping yards, shipping docks, transportation depots, loading / unloading bays, warehouse lots or yards, etc., whether paved (e.g., asphalt, concrete, brick, stone) or unpaved. The pavement marker 284 (and the other pavement markers described herein) is preferably affixed to the surface so that the longitudinal axis of a dihedral corner reflector (i.e., the axis where the two reflective surfaces intersect) within the pavement marker 284 is generally parallel to the direction of vehicle travel and along one or both sides of the path of travel of the vehicle. 50066 / 0501 Page 26 of 64

[0127] Figures 7A-7C and 8A-8C illustrate embodiments of pavement markers that may be mounted below the surface in rumble strips, parallel cut slots, or grooves. Figure 7A illustrates a front side view of a rumble strip 310, which is a cut groove formed in the road surface 300 to produce a sound when a tire rides over the rumble strip 310 to alert the driver. The polarimetric radar transceiver 116 transmits circularly or elliptically polarized signals 1119 and receives the same polarized returned signals 1118 from the side and bottom of the rumble strip 310. In this configuration, the side and bottom of the rumble strip form a dihedral corner reflector, providing a double bounce. In this embodiment, the S / N is significantly higher for circularly or elliptically polarized antennas, but with less return than a metallic corner reflector, since the pavement material is typically non-metallic.

[0128] Figure 8A shows a pavement marker 394 placed inside a rumble strip 310. In this embodiment, the returned radar signal 1118 will be altered by the presence of the pavement marker 394. The radar return-signal characteristics for this arrangement are different from an empty rumble strip 310. The polarimetric radar transceiver 116 with circularly or elliptical;y polarized antennas can detect the side of the rumble strip and then detect the existence of the pavement marker 394, depending on the pavement marker design. If the pavement marker 394 is designed with a T-shaped metallic dihedral corner reflector in the center of the pavement marker 394, the pavement marker 394 is substantially more reflective than the rumble strip sides; however, if no such addition is made and the marker is a typical plastic pavement marker, then it is less reflective than the sides of the rumble strip.

[0129] Figure 7B shows a top view of rumble strips 310 along a roadway with a painted line 910. Figure 8B shows a top view of the pavement marker 394 installed in rumble strips 310 along a roadway with a painted line 910. Figure 7C shows the top view of slots 312 cut in the highway. Here, the polarimetric radar transceiver 116 views the slots 312 for a more extended period, allowing the onboard controller to determine the lateral position and / or the vehicle heading by taking multiple lateral distance measurements from the long slot 312. Figure 8C shows another embodiment where the pavement marker 394 is installed in the slots 312, where existing pavement markers are typically installed, so that snow plows or other equipment will not tear them off the roadway surface. To improve the 50066 / 0501 Page 27 of 64reflectivity of these slots, rumble strips, or grooves, the milled rumble strips may be modified to form dihedral corner reflectors by using sprayed conductive paint. Figure 7A shows the milled rumble strip 310, which has been coated with conductive paint on its sides and bottom to increase the reflectivity of the rumble strip. The longer length milled highway slots 312, as shown in Figure 7C, can be painted with conductive paint along their entire length to provide a larger RCS for a polarimetric radar transceiver to view.

[0130] Figure 9 shows a front view of an installed snow-plowable pavement marker 1290 and the polarimetric radar transceiver 116—a snow plow rides on the top of sides 1206 and 1208, which are typically metal. An optical reflector 1275, generally made of plastic, is installed in the middle of the frame of the snow-plowable pavement marker 1290, allowing drivers or vision sensors to see the marker 1290. The polarimetric radar transceiver 116 transmits signals 1119 to the metal surface 1285 and receives signals 1118 back from a polarimetric radar transceiver 116 on one side of the pavement marker 1290. The radar return signal 1118 is stronger for the pavement marker 1295 than for the other raised pavement markers described earlier, as the pavement marker 1295 is made of metal and some of its sides are at right angles. Snow-plowable markers 1290 may be mounted in a center or interior lane line to be viewed by polarimetric radar transceivers mounted on vehicles in two different lanes. The snow-plowable marker 1290 is usually mounted in an interior road line. When the snow plowable marker 1290 is covered with snow, it is not visible to the human eye, cameras, or lidar. However, the marker 1290 is visible to the polarimetric radar transceiver 116 despite being covered with snow. Using a circularly or elliptically polarized polarimetric radar transceiver rather than the typical radar transceiver with vertical polarization antennas, the S / N is significantly higher, and the snow plowable pavement markers 1290 are highly visible by radar on the road surface 300.

[0131] For road surface targets that are along the painted roadway line, e.g., pavement markers, rumble strips, road edges, street curbs, grooves in the road, and bot dots, the polarimetric radar can determine the target type by transmitting both RHCP and LHCP transmit signals and receiving the copolar and cross-polar RHCP and LHCP backscattering signals from the target to form a polarization scattering matrix S. The target type can also be determined by transmitting vertically and horizontally polarized transmit signals and 50066 / 0501 Page 28 of 64receiving the copolar and cross-polarized backscattering signals from the target, which can form a polarization scattering matrix S. Either of these two techniques can differentiate between different road surface targets. By utilizing 3D (range, azimuth, & elevation) or 4D (range, azimuth, elevation, and velocity) radar information, along with a polarization S matrix, an AI model can accurately determine the target types. Knowing the target type can help determine the road or street by using a lookup database of navigational maps that includes road-surface structures. Concrete Barriers, Curbs, Streets, Road Edges, and Road Media Barriers

[0132] Figures 10A-10G illustrate structures along a roadside on the road surface or a short distance away from the lane center. Figure 10A shows a sidewalk 1422 and a curb 1420 next to a road surface 300. The polarimetric radar transceiver 116 transmits circularly or elliptically polarized signals 1119 to the curb 1420 and receives the same polarity reflective signals 1118 from the curb 1420. The curb 1420 acts like a dihedral corner reflector but has lower reflectivity than a metallic surface. The reflectivity is greater if the concrete is wet. Using circularly or elliptically polarized polarimetric radar transceiver 116, the S / N is significant higher than using vertical polarization, and the curb 1420 is highly visible to the polarimetric radar transceiver based on the curb 1420 is typically 4 inches in height (10 cm) and makes a good large radar cross section. Plus the road surface 300 signal return is significantly less because the reflection from the road reverses the handedness of the polarized wave and is the opposite handedness of the radar receiver circular or elliptical polarization antennas.

[0133] Figure 10B shows another surface of the concrete curb 1424. In Figure 10B, the curb has a slight curvature in the center, which will lessen the signal return.

[0134] Figure 10C shows a front view of a concrete barrier 1426 and the polarized polarimetric radar transceiver 116. In this embodiment, the polarized polarimetric radar transceiver 116 treats the concrete surface as a right-angle reflector along the bottom of the concrete barrier 1426 and as a flat surface along the top of the concrete barrier 1426. A RHCP transmit signal 1119 is transmitted at the bottom of the barrier, and the signal is received as a RHCP received signal 1118 at the corner side of the barrier. On the upper 50066 / 0501 Page 29 of 64section of the concrete barrier, a vertical polarized signal 1719 is transmitted at the side of the barrier 1426, and the signal 1718 is reflected and received as a vertical polarized signal. The polarimetric radar transceiver 116 in this embodiment has various polarized transmit and receive antennas.

[0135] Figure 10D shows the roadway 300 with pavement markers 284 on one side and a road edge 305 and off-road surface 1305 on the other side. In this embodiment, a polarimetric radar transceiver can transmit and receive a vertically or circularly polarized wave at different angles to detect the road edge, as the road surface 300 provides a distinct return compared to the off-road surface 1305.

[0136] Figure 10E is similar to Figure 10D, but here the off-road surface 305, which is smoother, can be detected by the polarimetric radar transceiver. Additionally, Figure 10E features a side structure, such as a sign 1306, which is visible to the polarimetric radar transceiver using vertically polarized antennas. The roadside structure can be treated as a flat reflection or single bounce reflection, and vertical polarization antennas can be used. Figure 10F shows an isometric view of a center curb 1928 and a street-side curb 1420. The polarimetric radar transceiver 116 (not shown in the figure) can view the structures as corner reflectors to receive strong return signals and achieve a high S / N using circularly or elliptically polarized antennas.

[0137] Figure 10G shows the polarimetric radar transceiver 116 viewing a guardrail 1432. Both vertical and circular polarization antennas can be used to determine the structure type, e.g., a guardrail.

[0138] There can be many items, markers, or targets along the roadside that have specific radar characteristics that allow an onboard processor to classify or recognize the target based on factors such as return signal strength, polarization backscattering signals, the number of markers, and previously recorded data (e.g., in a database) for those markers (e.g., approximate location, distance, expected RCS characteristics, etc.).

[0139] For roadside targets that are along the pathway, e.g., guardrails, concrete barriers, road signs, lamp posts, buildings, utility poles, etc., the polarimetric radar can determine the target type by transmitting both RHCP and LHCP transmit signals and 50066 / 0501 Page 30 of 64receiving the copolar and cross-polar RHCP and LHCP backscattering signals from the target to form a polarization scattering matrix S. The target type can also be determined by transmitting vertically and horizontally polarized transmit signals and receiving the copolar and cross-polarized backscattering signals from the target, which form a polarization scattering matrix S. Either of these two techniques differentiate between different roadside targets. A combination of using both circularly and vertically polarized signals provides a robust method for determining target type. By utilizing 3D (range, azimuth, & elevation) or 4D (range, azimuth, elevation, and velocity) radar information, along with the polarization S matrix, an AI model can accurately determine the target types. Knowing the target type, whether discrete or semi-continuous, can help determine the road, street, and location on the road / street by using a lookup database of navigational maps that includes roadside structures. A navigational map with road surface and roadside targets, along with their geographical location within the map, can provide not only lateral distance information from the side target but also the location along the roadway. Processing Polarimetric Radar Transceiver Signals

[0140] Range-based gating, time gating, amplitude gating, polarization gating, and other filtering methods (e.g., Kalman filtering) can be employed to exclude reflection signals from objects that are not part of the road surface or roadside targets of interest. Several methods of gating or filtering the received signal can be utilized to detect pavement markers or other target structures, as illustrated in Figures 11A-11C.

[0141] In Figure 11A, a range gate 3030 is illustrated, and the range is gated based on the expected width of roads and lanes. The expected width may be stored in a database. Lanes are typically 2.7 to 4.6 m (9 to 15 ft.) wide. The vehicles 180 and 182, equipped with polarimetric radar transceivers 116 and 117, view the pavement markers 284 and other target structures 2841. The pavement markers 284 are denoted in each lane, but only labeled for the top lane. Recorded data, along with location information (e.g., GNSS-based), can be used to determine the approximate location of a vehicle in terms of street, lane, and other relevant details. With this information and the expected lane width or expected distance to the target structure, a suitable range gate 3030 can be determined, as illustrated in Figure 11A. 50066 / 0501 Page 31 of 64

[0142] In Figure 11B, a Kalman filter tracks the target structures based on a vehicle lateral movement model, vehicle speed, and expected target structure spacings from a navigational database. The range gate 3032 can be smaller due to the known vehicle model and the Kalman filter, allowing for more effective filtering of other road debris. Figure 11B also shows a short section of a concrete barrier 1426, which can also be detected by Vehicle #2’s polarimetric radar transceiver 116 and filtered for lateral distance.

[0143] In Figure 11C, a time gate 3083 on the expected pavement marker 284 or other target structure 284' is based on the vehicle speed and known spacings of the target structures from a database. Figure 11C shows the time gating on the expected concrete barrier 1426.

[0144] Figures 12A-12C illustrate polarimetric radar transceiver processing and filtering of both received signal amplitude (denoted in dB) and range to the target to obtain a more reliable distance to the target structure, and no distance if there is no marker visible to the polarimetric radar transceiver. Figure 12A shows the radar-measured distance vs. time to the strongest targets with clutter from a vehicle to the side targets on the side of the road while the vehicle is moving along the roadway. The side targets include three RF- reflective markers 1100, 1102, and 1104. Figure 12B shows the radar amplitude vs. time for the same setup. The three RF-reflective markers, 1100, 1102, and 1104, have the highest returned signal strength, as seen in Figure 12B. An amplitude threshold of 1106 can be set to filter the radar marker data from clutter. With both amplitude filtering and range gating, Figure 12C shows the results where the radar distance to the three markers 1100, 1102, and 1104 are noted with the clutter removed.

[0145] These techniques can have static range gating and amplitude threshold values, or the thresholds and range can be dynamic (e.g., determined from previously collected radar data). The range, time, and amplitude can be updated with proper gates. Range gates are typically based on the road width, lane widths, and the number of lanes. A map database can provide that information based on the vehicle’s location, as determined by GNSS or another navigational sensor. Time gating between pavement markers can be based on the marker spacing and vehicle speed, where the time is calculated as the marker spacing divided by the vehicle speed. The received signal strength (amplitude) gate can be based on the average 50066 / 0501 Page 32 of 64received signal strength over previous (e.g., the last few) target passings, plus a pre- determined threshold value. Polarization gating utilizes different antenna polarizations to minimize the visibility of specific targets while enhancing the visibility of others. As previously described, circular polarization gating can eliminate reflected signals from flat or round targets, e.g., road surface or pop cans or bottles (a single bounce), but allow reflected signals from corner surfaces, e.g., pavement markers, curbs, and concrete barriers (a double bounce) by using same circular or elliptical polarity of transmit and receive antennas. Circular polarization gating can eliminate road surface targets, such as pavement markers, curbs, and concrete barriers, while allowing road surface reflections by using opposite circular polarities (e.g., TX RHCP and RX LHCP) antennas. Using circular polarization gating in this manner could determine the road surfaces and road conditions.

[0146] Figures 13A and 13B are top views of diagrams illustrating the geometry of a vehicle traveling along a road with road surface markers being viewed by an onboard polarimetric radar 116 and with an onboard GNSS sensor 901. Figure 13A is a top view of a diagram illustrating a vehicle 180 traveling along a road. The dynamics of the vehicle 180 show that the vehicle 180 is travelling in the y direction along a lane with a velocity of v. The wheel angle δ of the front wheels 804 is shown from the vehicle 180 going straight ahead. The x direction is the lateral movement of the vehicle 180. The yaw angle of the vehicle 180 is nonzero when travelling around a curve as compared to straight ahead. The vehicle 180 has a polarimetric radar transceiver 116 mounted on the side of the vehicle 180, measuring a distance d1between one RF-reflective pavement marker 445 and the radar transceiver 116. The spacing 194 of the RF-reflective pavement markers 445 is denoted as a distance dm. Thus, the polarimetric radar transceiver 116 measures the lateral distance to the passing RF-reflective pavement markers 445 every time the vehicle 180 passes a marker 445.

[0147] Figure 13A also shows another polarimetric radar transceiver 1016 mounted similarly to the radar transceiver 116, but on the rear of the vehicle 180, on the same side as the radar transceiver 116. Likewise, two radar sensors can be mounted on the opposite sides and measure the distance from the RF-reflective pavement markers (not shown). Radar transceiver 116 measures distance d1 from one of the RF-reflective pavement markers 445, and at a similar time or as soon as it passes a RF-reflective pavement 445, the radar 50066 / 0501 Page 33 of 64transceiver 1016 measures the distance d2to another of the RF-reflective pavement markers 445. Besides providing the distances from the polarimetric radar transceiver to the RF- reflective pavement markers 445, a vehicle heading angle relative to the RF-reflective pavement markers can be derived by the formula: Ψ(r)= tan-1[(d1-d2) / L] where Ψ(r) represents the heading angle of the vehicle relative to the markers using radar (r), d1 is the radar measurement taken from the radar transceiver 116, where d2 is the radar measurement taken from the polarimetric radar transceiver 1016 at about the same time, and where L is the length between the polarimetric radar transceivers 116 and 1016. The lateral distance d1 and the estimated vehicle heading angle ψ(r) of the vehicle 180 relative to the pavement markers 445 can be determined from these two radar distance measurements. One embodiment features two polarimetric radar transceivers spaced the same or multiple lengths apart from each other, such as the pavement markers 445, with a separation distance of dm194. This arrangement allows the polarimetric radar transceiver 116 to pass a pavement marker 445 at approximately the same time as the polarimetric radar transceiver 1016 passes another pavement marker 445. The specific radar transceiver spacing L minimizes concerns that vehicle motion might cause issues in measuring the vehicle heading angle, ψ(r). For instance, if the pavement marker separation distance of 194 is 1.5 meters, then thetransceivers should be spaced at 1.5, 3.0, 4.5, or 6.0 meters apart for both radar transceivers to view pavement markers 445 at approximately the same time.

[0148] The vehicle 180 is equipped with a GNSS sensor 901 that provides similar information as above, but using a navigational map. As the vehicle travels forward, the GNSS sensor 901 measures locations g1 and then measures g2 a short-traveled distance later. Waypoints w1 and s2 are two waypoints on the navigational map that indicate the direction the vehicle should follow to stay in the lane. If the location g1 and w1 are about the same, then the vehicle heading angle ψ(g) relative to the navigational map GNSS information can be derived by the formula: Ψ(g)= tan-1{[(w1-g1) -(w2-g2)] / y(g)} 50066 / 0501 Page 34 of 64where Ψ(g) represents the heading angle of the vehicle relative to the GNSS waypoints on the map using GNSS sensor (g), and where w1-g1 is the distance between the waypoint w1 and the measured GNSS location g1, w2-g2 is the distance between the waypoint w2 and the measured GNSS location g2 at a specified time later, and y(g) is the distance travelled by the vehicle.

[0149] Using both the polarimetric radar transceiver’s vehicle heading angle Ψ(r) and the GNSS-derived heading angle Ψ(g), one can employ a Kalman filter to combine these two heading angles and provide a precise vehicle heading angle error within the lane since the pavement markers 445 and the navigational map are in alignment. Therefore, the vehicle’s heading error, along with the lateral position error from the lane center, can be determined.

[0150] Figure 13B is a top view of another embodiment illustrating a vehicle 180 travelling along the road 300 using only a front polarimetric radar. Figure 13B differs from Figure 13A in that the vehicle 180 in Figure 13A has two side-mounted radar sensors 116 and 1016, whereas the vehicle 180 in Figure 13B has just one side-mounted radar sensor 116. The dynamics of the vehicle 180 show that it is traveling in the y-direction along the lane with a velocity of v. The wheel angle δ of the front wheels 804 is shown as the vehicle 180 navigates a straight section of road. The x-direction is the lateral movement of the vehicle. The yaw angle of the vehicle 180 is nonzero when traveling around a curve versus a straight segment of the road. The vehicle 180 has a radar transceiver 116 mounted on the side of the vehicle measuring a distance dX2 between one RF-reflective pavement marker 284 or side target and the radar transceiver 116. The spacing 194 of the RF-reflective pavement markers 284 is denoted as a distance dmand stored in a database for side targets that are not spaced equally. The radar transceiver 116 measures the passing RF-reflective pavement markers 284 every time the radar transceiver 116 passes a marker 284. The time tRbetween pavement markers is derived from the set marker 284 separation distance dm stored in a database, divided by the average vehicle velocity v during the transition between pavement markers 284: tR = dm / v(average). For example, suppose the vehicle 180 is travelling at a constant velocity of 25 m / s, and the distance between markers is 3 m. In that case, the update time of the radar transceiver distance measurements tR is 3 m / 25 m / s or 120 ms. 50066 / 0501 Page 35 of 64

[0151] Figure 13B shows only one side-mounted polarimetric radar that uses the travel displacement to determine the vehicle heading angle relative to the pavement markers. The forward movement of the vehicle 180 from one position to another position 180' is shown in the dashed lines. The same polarimetric radar sensor 116', at a later time tR will measure the lateral distance dX1 to the next pavement marker 284. Besides providing the distances from the polarimetric radar sensor to the pavement markers 284, a deviation angle ψ2(r) can be derived by the formula: Ψ2(r) = sin-1[(dx1-dx2) / dm] where Ψ2(r) represents the heading angle of the vehicle relative to the markers using radar (r), and where dX2 is the radar measurement taken at time tR, the time between pavement markers, before dX1, the measurement, where dx2is the radar measurement taken at the current time, and where dm is the spacing of the pavement markers. The lateral distance dXof the vehicle 180 and the vehicle deviation angle Ψ2(r) relative to the pavement markers 284 can be determined by taking multiple radar distance measurements from several pavement markers 284.

[0152] A vehicle equipped with a GNSS sensor 901 provides similar information as above, but using a navigational map. As the vehicle travels forward, the GNSS sensor 901 measures locations g1 and then measures g2 a short distance later. Waypoints w1 and w2 are two waypoints on the navigational map that indicate the direction the vehicle should follow to stay in the lane. If the location g1 and w1 are about the same, then the vehicle heading angle ψ2(g) relative to the navigational map GNSS information can be derived by the formula: Ψ2(g)= tan-1{[(w1-g1) -(w2-g2)] / y(g)} where Ψ2(g) represents the heading angle of the vehicle relative to the GNSS waypoints on the map using GNSS sensor (g), and where w1-g1 is the distance between the waypoint w1 and the measured GNSS location g1, w2-g2 is the distance between the waypoint w2 and the measured GNSS location g2 at a specified time later, and y(g) is the distance travelled by the vehicle. 50066 / 0501 Page 36 of 64

[0153] Using both the polarimetric radar transceiver’s vehicle heading angle Ψ2(r) and the GNSS-derived heading angle Ψ2(g), one can employ a Kalman filter to combine these two heading angles and provide a precise vehicle heading angle error within the lane since the pavement markers 284 and the navigational map are in alignment. Therefore, the vehicle’s heading error, along with the lateral position error from the lane center, can be determined. Polarimetric Radar and Target Location Databases

[0154] Figure 14A shows a front view of two vehicles, 180 and 182. Figure 14A illustrates the mounting of a polarimetric radar transceiver 119 in the center front of the vehicle 180 to view the road surface and roadside targets from polarized antennas viewing both directions (i.e., opposite lateral directions, right and left or driver-side and passenger- side). In this embodiment, the polarimetric radar transceiver 119 uses radar beams 1106- 1109 to view the targets. Figure 14A shows that other sensors, e.g., GNSS sensor 901, vision sensor 903, and IMU 907, can be fused with the polarimetric radar 119. There are different sensors (not shown in Figure 14A), including lidar, road speed sensors, and wheel angle sensors, that can also be integrated. The GNSS sensor 901 provides vehicle location information based on a satellite constellation 1901. There are several global navigation satellite systems currently in use: the United States’ GPS (Global Position System), Russia’s GLONASS, the European Union’s Galileo, and China’s BeiDou. The accuracy of standalone GPS is around 3-5 meters. A GNSS sensor 901 tracking all the satellites and satellite signals can achieve sub-meter accuracy. If corrections are available using DGPS (Differential Global Positioning System), RTK-GNSS (Real-Time Kinematic - Global Navigation Satellite System), PPP-GNSS (Precise Point Positioning - Global Navigation Satellite System), or RTK-PPP (Real- Time Kinematic – Precise Point Positioning), centimeter-level accuracy can be achieved in an open-sky environment. IMU 907 can be integrated with GNSS sensor 901 and / or polarimetric radar 119 for better location information.

[0155] The polarimetric radar transceiver 119 can simultaneously or effectively simultaneously track road surface targets and roadside targets using various polarized antennas. Radar beams 1108 and 1109 can have same-polarity transmit and receive circularly polarized antennas (e.g., RHCP to transmit and RHCP to receive). Same-circularly- 50066 / 0501 Page 37 of 64polarized antennas can detect the sharp corners of the pavement markers 445 or curbs 1420 (double-bounce reflections) while attenuating the road return (single-bounce reflection). Radar beams 1106 and 1107, on the other hand, can have vertically polarized transmit and receive antennas or oppositely circularly polarized transmit and receive antennas (e.g., RHCP to transmit and LHCP to receive). The side targets produce single-bounce reflections and can be easily viewed by the polarimetric radar 119 via either vertical, horizontal, or circular polarization, or a combination of these for classifying the target. This capability to see side targets can achieve better location accuracy and avoid vehicles or obstacles if the vehicle is changing lanes or drifting from the lane.

[0156] Figure 14B shows a typical roadway 300 with two vehicles 180 and 182 using side radar transceivers 116 and 117 to detect road surface and roadside targets. The radar transceiver 117 on vehicle 180 views the concrete barrier 1426, and radar transceiver 116 on the same vehicle 180 views the pavement markers 284 in the center of the roadway 300, as well as the far-side objects on the passenger side, e.g., curbs 1420. The radar transceiver 117 on the vehicle 182 views the pavement markers 284 in the center of the roadway 300 and the concrete barrier 1426 on the far driver’s side. The radar transceiver 116 on vehicle 182 views the road edge 305, the signpost 1306, curbs 1420, drainage outlet 1421, lamp post 1309, and the roadside surface 1305. The radar transceivers 116 and 117 can view many different roadside targets and track them simultaneously or effectively simultaneously on or near the road via looking downward and to the side, near the side of the road via looking downward and to the side, and near the side of the road via looking only to the side as shown in the previous figures on the radar transceiver viewing roadside targets. The radar transceiver 116 on vehicle 180 may also view the lateral distance 1811 from the vehicle 182 when the vehicle 180 passes the vehicle 182. The vehicle 182 would typically block the view of the radar transceiver 116 on the vehicle 180 from viewing the far roadside targets, e.g., the side post 1306. This added lateral distance 1811 information, in addition to viewing the markers 284 on the road, may be used to provide a more robust control system. Likewise, the radar transceiver 117 on the vehicle 182 can view the same lateral distance 1811 when the vehicle 180 passes by. 50066 / 0501 Page 38 of 64

[0157] Figures 15A-15C illustrate three example tables of different databases of target information. Figure 15A is a road surface target database 8066 containing target information on road surface targets. Figure 15B is a high-definition (HD) database 8866 providing location information and other relevant characteristics on targets. Figure 15C is a typical navigational database, such as Google Maps or OpenStreetMap, which includes route information via nodes, links or road segments, and surrounding buildings. The heading reference information can be calculated from nodes and links or road segments to provide the heading direction that the vehicle should follow.

[0158] Figure 15A illustrates some example data in a database 8066 of various types of road surface targets, including their placement relative to the lane center, length, height, and spacing. For example, the raised pavement markers are typically spaced along the painted line with a distance from the lane center of about 1.6 m. Therefore, the road surface target database 8066 and the radar distances from these targets provide lateral guidance and control information similar to that of the vision or lidar sensors tracking painted lines. The length of the target as the vehicle passes by is 100 mm, and the height is 20 mm above the road surface. It is typically spaced about 3 m around curves or about 12.2 to 24.4 m on straight sections of the road. These road surface targets for each road are generally stored in the databases of DOTs (departments of transportation). The DOTs of states, provinces, or countries have specified these items for their roads. Some DOTs or third-party databases even have the precise locations of some of these road surface or roadside targets. A radar sensor can determine the lateral distance based on range measurement using current radar processing techniques and select the type of target by utilizing polarization and beam steering / MIMO antenna techniques, with comparison to known targets on the road surface or roadside. The length, height, and spacing can be determined from the database as the polarimetric radar sensor passes by the targets. Other radar processing techniques, besides filtering on magnitude and range, can be used to distinguish the pavement markers from the road surface and other objects. For example, polarization can distinguish a dihedral corner reflector from the road surface and other objects on or near the road. Also, fixed or approximately fixed spacing of the pavement markers can be used to filter out false signals based on time. For example, a reflection arriving between two pavement markers may be 50066 / 0501 Page 39 of 64excluded as a probable false target. As described above, a Kalman filter, complementary filter, or other known filtering techniques can be used to filter out reflection signals based on the fixed spacing / time of the pavement markers and / or other known or predictable attributes of the pavement markers. The pavement markers 284 are therefore distinguishable from other objects, and the many pavement markers 284 on a driving surface may be uniformly identical or similar with similar dihedral corner reflectors and similar RCSs.

[0159] Figure 15B illustrates some example data in an HD navigational database 8866 with accurate location information for the target, including target type, length, roadside or road surface target, spacing, and other relevant details. The HD navigational database 8866, along with a GNSS sensor and a polarimetric radar transceiver, can not only locate the lateral distance to the road surface target or roadside structure but also provide GNSS coordinates for dead reckoning landmarks to update the GNSS sensor for accurate tracking of longitudinal and lateral information. Figure 15B can contain the road reference heading and road surface targets illustrating not only the location of the target, the target type, the length of the target, radar beam position and polarization, the type of road surface, the typical target spacing for similar targets, and the heading reference angle of the road segment or trajectory. The database 8866 may be created and updated based on the following: (1) Current navigational maps, e.g., Google Maps, Google Earth, Microsoft Maps, OpenStreetMap, etc. (2) Repeated route data, collected from travelling the route and using SLAM (simultaneous localization and mapping) techniques. (3) Surveying current road surface or roadside targets and their characteristics via polarimeter radar and GNSS.

[0160] The HD navigational database 8866 shows a table of some items. Still, the table could also include one or more radar sensors for side target location, mileage marker information, lane information for multiple-lane highways, types of targets, radar mounting side, spacing of the targets, estimated target RCS, radar transceiver settings to obtain maximum or at least desirable returned signals from the targets, the best radar antenna 50066 / 0501 Page 40 of 64polarization for target type, the radar beam angle or angles for multiple views, radar mounting height or height of the best antenna pair with the highest RF return, and potential S / N or confidence level in returned signal. Some targets are more similar to a dihedral corner reflector, e.g., curbs 1420; others are similar to raised pavement markers 445 or snow plowable markers 1290, etc. Therefore, a circular polarized radar antenna works well on those roadside targets since a dihedral corner reflector is a double-bounce target and returns an RHCP wave as an RHCP wave, for example. Other roadside targets, e.g., signpost 1306, lamp post 1309, road edge 305, etc., work well with vertically polarized radar since the RF return signal is a signal-bounce signal. Other roadside targets may require multiple types of polarization to provide the best target return and, therefore, the best target recognition.

[0161] A database like the one partially illustrated in Figure 15B (8866) could include RF-reflective pavement markers with metallic dihedral corner reflectors, which typically have the highest returned signals and provide the vehicle to be controlled laterally with the radar transceivers 116 with the highest confidence. A database like the one partially illustrated in Figure 15B (8866) can be recorded using a GNSS sensor (like 901) with centimeter accuracy, e.g., the u-Blox XPLR-HPG-2 Unit; radar sensors 116 & 117; IMU 907; and vehicle speed sensor 8008. Lidar, vision, and forward-looking radar may also be recorded in the database to allow for fusion of side-looking radar information with the other onboard sensors. The database 8866 may also include or be used with trained AI models using machine learning and / or deep-learning techniques for classifying radar targets, lidar targets, and vision targets. The data in the database 8866 may be used to train an AI model. Typical road surface and roadside targets include curbs 1420, sign posts 1306, shoulder edges 305, lamp posts 1309, raised pavement markers 284, slots 310 & 312 in the road, telephone poles, manhole covers, mileage marker signs, curb drainage openings 1315, snow plowable markers 1290, trees with and without leaves, guardrails, concrete barriers 1426, city buildings, tunnel sides, bridge sides, construction cones, etc.

[0162] As shown in Figure 15B, a polarimetric radar sensor 116 on a vehicle tracks the raised pavement markers 445, one type of target. It expects the next pavement marker, 445, to be approximately 10 m away (time to next marker = marker spacing × vehicle speed). Therefore, when the next marker 445 is determined from the polarimetric radar sensor 116, 50066 / 0501 Page 41 of 64the database 8866 can indicate the stored latitude and longitude information for the marker 445. The difference between the actual GNSS readings and the HD database 8866 readings can be used to update the GNSS sensor 901. The vehicle position relative to the pavement marker 445 can be determined both laterally (the radar-measured distance) and longitudinally (the maximum return signal is received when the vehicle starts to pass or passes the known location of pavement marker 445). The lateral distance between the pavement marker 445 and the vehicle can determined by the polarimetric radar transceivers on each vehicle, as they measure the lateral distance to the pavement marker 445 as each vehicle passes it.

[0163] The HD database 8866 can store a desired lateral distance between the vehicles in the center of each lane and the pavement marker 445. The difference between the stored, desired lateral distance (which could be obtained, for example, by recording radar-measured distances when the vehicles are known to be in or near the center of the respective lanes) and the currently measured lateral distance is an offset distance of the vehicles versus the center of their respective lane. The longitudinal locations of the vehicles along the roadway are determined via the same radar sensors 116 and 117, respectively, on each vehicle, detecting the maximum amplitude of the returned signal from the pavement marker 445 as it passes by the pavement marker 445 and using the HD database 8866 for the known location of that particular pavement marker 445. The HD database 8866 can store locations (e.g., GNSS coordinates) of a vehicle in the center of the lane and / or the pavement marker 445. As the vehicles pass the pavement marker 445, the peak amplitude of the returned radar sensor signal indicates the center of the pavement marker 445. The stored location for the pavement marker 445 then determines the current longitudinal location of the vehicle based on the peak radar signal return as the vehicle passes the pavement marker 445. Alternatively, the derivative of the return signal can be monitored, and a change in that derivative signal corresponds to the beginning of a pavement marker 445. In the above embodiment, the IMU 907 can be used between targets to update the GNSS sensor 901. However, the concrete barrier 1426 provides a continuous lateral position from the polarimetric radar on the vehicle, and the IMU 907 may not be needed for lateral position 50066 / 0501 Page 42 of 64updates. The longitudinal updates from concrete barrier 1426 could be provided via IMU 907, vehicle speed 8008, and intersections between the concrete barriers.

[0164] Several different embodiments provide vehicle location information, integrating a polarimetric radar transceiver with a GNSS sensor, vision sensor, and / or lidar sensor: (1) Using the GNSS (RTK-GNSS and PPP corrections) sensor for accurate location, given the necessary satellites, and the polarimetric radar acts as a backup to the GNSS. (2) GNSS (without corrections) for general vehicle location with the polarimetric radar providing accurate location updates. The polarimetric radar updates the GNSS sensor with precise latitude and longitude information, providing the vehicle’s location. (3) In GNSS-denied areas, e.g., tunnels, bridges, urban areas with tall buildings, mountainous regions, and dense forests, the polarimetric radar provides accurate vehicle location both laterally and longitudinally. (4) Vision sensors provide vehicle lateral location via painted lane lines; the polarimetric radar acts as a backup during inclement weather or when the vision sensor becomes blinded. (5) Lidar sensors provide vehicle lateral location via painted lines and stored boundaries, and the polarimetric radar acts as a backup during inclement weather.

[0165] Figure 15C illustrates some example data in a road navigational database 8006, such as Google Maps or OpenStreetMap, which includes route information via nodes, links, or road segments, surrounding interest sites, and various attributes. The heading reference information can be calculated from nodes and links or road segments to provide the heading direction that the vehicle should follow. The road database is a basic model of road geometry. Still, it can contain many other layers and models, such as a lane model with lane information and a location model with landmarks, concrete barriers, and other features. Street names, numbers, and house numbers, as well as points of interest, are denoted as 50066 / 0501 Page 43 of 64waypoints or nodes and coded as GNSS coordinates. This enables users to find a desired destination by street address or by geographic coordinates. Each node within a map graph represents a point location of the surface of the road and is represented by a pair of longitude and latitude coordinates. Each link or road segment represents a stretch of road between two nodes, and is represented by a line segment (corresponding to a straight section of road) or a curve having a shape that is generally described by intermediate points along the link. However, curves may also be represented by a combination of points or nodes, along with a radius and polar coordinates, to define the boundaries of the curve. Vector maps store geographic features as geometric shapes with attributes. These features are composed of coordinates (latitude / longitude). The change in latitude and longitude coordinates between two points can be used to calculate the direction or heading of a segment, for example, a road segment. The heading of a vector can be determined using trigonometric functions like the arctangent (atan2) of the change in y coordinates divided by the change in x coordinates (dy / dx). This yields the angle of the vector relative to a horizontal axis. Figure 15C shows the nodes, their latitude / longitude locations, heading directions (derived from the vehicle’s travel along the road path), and road segments. Different Configurations of Polarimetric Radar and Lateral Controller

[0166] Figure 16A illustrates a block diagram of the lane guidance or control system 1850, which can be utilized for a lane departure warning (LDW), lane assistance system (LAS), and / or lane keeping system (LKS). The guidance controller 830 receives inputs from the polarimetric radar transceiver 116, the road surface target database 8066, and one or more sensors, such as the IMU 907, the wheel speed sensor 8008, the steering wheel angle sensor 8028, and the turn signal 822. From these inputs, the guidance controller 830 outputs sound to a speaker 824, a visible display 826 of the vehicle within the lane, and / or a control signal to a steering mechanism 802 to turn wheels 804. The polarimetric radar transceiver 116 measures the distance dR from the road surface target, in this example, the distance to a pavement marker 445. As described earlier, the polarimetric radar transceiver 116 transmits an RHCP signal, for example, signal 1119, using circularly polarized antennas to the pavement marker on the road surface. It receives back an RHCP signal 1118 from the pavement marker 445 and an LHCP signal (not shown) from the road surface. Since the 50066 / 0501 Page 44 of 64received antenna (in this example) is also RHCP, the pavement marker reflection signal is well received, while the road surface reflection is attenuated. However, to determine the type of target, i.e., a pavement marker, both copolar and cross-polar circularly polarized signals are transmitted and received from the backscatter of the target. The S matrix can be derived, and polarization decomposition can be performed on the S matrix to determine the target type using machine learning techniques. Existing road surface targets on the road can be distinguished using polarization, allowing these targets to be used in the guidance controller system 1850 without any changes to the road infrastructure. Since most roads have pavement markers, bot dots, rumble strips, or grooves in the road, using polarimetric radar to view these targets and measure the distance to them is similar to using vision or lidar sensors with the painted lines. The radar operates in all-weather conditions, whereas vision and lidar are impaired by snow, fog, or heavy rain. However, if the pavement markers or road surface targets are made more RF-reflective, e.g., by adding a dihedral metallic reflector as suggested earlier and spacing them closer together, a more robust controller / infrastructure system can be created.

[0167] Figure 16B illustrates another block diagram of a lane guidance or control system 950, according to another embodiment, which can be utilized for LDW, LAS, and / or LKS. Figure 16B differs from Figure 16A in that Figure 16B features two polarimetric radar transceivers, 116 and 117, for tracking, for example, both the pavement marker 445 on one side of the vehicle and the curb 1402 on the other side. By providing two lateral measures to the controller 830 and using filtering techniques to combine the two radar measurements dX1and dX2, respectively, a more robust lateral guidance and control system can be achieved.

[0168] Figure 17 illustrates an embodiment of lateral guidance and navigation schemes. Figure 17 is a block diagram of a guidance / navigation system 880 that warns the driver if the vehicle leaves its lane, providing visual, audible, haptic, and / or vibration warnings, and provides navigational guidance information to the driver. This embodiment represents a LDW and / or LKA / LKS system integrated with a navigational system database 8006, which may contain attributes of radar-based targets with location information. The polarimetric radar transceiver 116 transmits a signal 1119 in the direction of a side target 284 and receives a signal 1118 reflected from the target structure 284. The distance 50066 / 0501 Page 45 of 64between the road surface target, e.g., pavement marker 284, and the polarimetric radar transceiver 116 is measured by the radar transceiver 116 using well-known radar techniques and provided to the lateral guidance controller 880. The polarimetric radar transceiver 116 measures the distance dRfrom the road surface target 284. The actual lateral distance from the vehicle to the pavement marker 284 is related to the radar’s measured distance dR and the geometry of the mounting of the radar transceiver 116 attached to the vehicle. The radar transceiver 116 distance can be fused with a GNSS sensor 901, IMU 907, and navigational database 8006. The navigational database 8006 contains the coordinates (e.g., GNSS) of the road geometry, as well as optional attributes of the road surface or roadside target location and characteristics. When the polarimetric radar transceiver 116 passes the marker 284, it measures the distance dRto the road surface marker 284. The guidance controller 880 is programmed to utilize the current GNSS sensor 901 and IMU 907 information to determine the vehicle’s current location and select the nearest marker in the database 8006 or the nearest marker that acceptably matches the radar characteristics and location information provided by the polarimetric radar transceiver 116 and the GNSS sensor 901, respectively. Once a match is found, the exact location information from the database 8006 is used to update the navigation information. The guidance controller 880 fuses these sensors (GNSS 901, IMU 907, vehicle speed 8008, and radar 116) using, for example, an extended Kalman filter. The output of the extended Kalman filter provides more accurate position and speed information under various external conditions, such as the blockage of navigation / positioning satellites while in the tunnel. Suppose the global positioning (e.g., GNSS) satellites are blocked. In that case, the IMU sensor 901 can provide relative information on the vehicle location, and the polarimetric radar transceiver 116 can provide absolute details every time the vehicle passes the road surface marker 284.

[0169] With the fusion of these types of sensors, the controller 880 in Figure 17 can provide an early warning if the vehicle deviates from its lane of travel. If the vehicle travels too close to or too far away from the target 284, indicating a lane departure, the guidance controller 830 can cause a sound to be emitted by a speaker 824 (or plural speakers) to alert the driver of the lane departure, cause the display 826 to alert the driver of the lane departure visually, and / or cause vibration of the driver seat, steering wheel, or other vehicle 50066 / 0501 Page 46 of 64part. The guidance controller 880 can disable lane-departure alerts if a turn signal 822 is activated, indicating the driver’s intention to initiate a lane change or turn in the direction of the detected lane departure. Whether lane-departure alerts are enabled or disabled by a prior turn signal can be a configurable setting set by the driver, factory, or a service technician. The guidance controller 880 can also be configured to activate the turn signal indicator (e.g., blinker) on the side of the vehicle corresponding to the detected lane departure when a lane departure is detected. That can serve as a warning to other drivers, vehicles, pedestrians, cyclists, etc. that a lane departure is occurring or might occur even if the driver has not actuated a turn signal. In this way, safety can be enhanced for others as compared to the case in which no lane-change indication is provided.

[0170] Figure 17 also shows a navigational display 888 to provide the driver with information regarding the road the vehicle is traveling on and instructions to follow to reach the driver’s destination. The guidance system 850 can warn the driver if the vehicle is leaving its lane with visual, audible, and / or haptic warning, and, if the driver takes no corrective response, the system 850 can automatically take steps to maintain the vehicle in its lane. The database 8006 in Figure 17 can be upgraded via wireless connection to the Internet 500 to keep the database 8006 current with updates reflecting road changes and maintenance.

[0171] In addition to providing lateral distance to the road surface or roadside target, the guidance system 880 can also provide the vehicle’s heading angle relative to the road surface targets and / or relative to road navigational map waypoints using the navigational database 8006 and the GNSS sensor 901. The navigational database can provide the road geometry ahead of the vehicle using the current GNSS coordinates so that curvature in the roads can be previewed and anticipated by the controller 880.

[0172] Figure 18 illustrates a block diagram of another embodiment 950, featuring two polarimetric radar transceivers on the same side, which measure distances dX1 and dX2, respectively, from the road surface targets, e.g., pavement markers 445. These distances are inputs to a lateral guidance controller 890. In addition to the radar input, the guidance controller 890 receives inputs from the vehicle speed sensor 8008, GNSS 901, wheel speed sensor 8008, and the IMU 907. In addition, the guidance controller 890 has access to a high- 50066 / 0501 Page 47 of 64definition HD database 8866 containing road reference heading and route information, as well as a reference table of road surface targets and roadside targets, along with their corresponding locations. The vehicle speed sensor 8008 provides the vehicle’s forward velocity. The IMU 907 provides the vehicle’s acceleration, angular rate, and orientation, using a combination of accelerometers, gyroscopes, and, in some cases, a magnetometer. The lateral guidance controller 890 can accept additional sensor information, such as lateral acceleration and compass direction, to provide a more robust lateral guidance control system. The lateral guidance controller 890 sends a signal to the steering motor or controlled device in the steering mechanism 802. The steering mechanism 802 controls the steering wheels 804. The steering wheel 804 and the steering mechanism 802 are for road vehicles, but the vehicle could be an off-road vehicle, and the controller could utilize hydraulic actuators. The desired lateral distance that the vehicle maintains from a road surface target, e.g., pavement marker 445, is controlled by adjusting the steering wheels 804. Note that in some cases, there may be only a single steerable wheel, such as a bike, trike, or three-wheeled vehicle.

[0173] The polarimetric radar transceivers 116 and 1016 may generate a continuous radar transmission signal and / or monitor continuously for a reflection, or the polarimetric radar transceivers 116 and 1016 may periodically and repeatedly transmit a discrete radar signal (e.g., a pulse or a chirp) and monitor for reflections, depending on the type of radar techniques employed. The radar transmission signal is generally directed partially downward and sideways from the vehicle 180, and the pavement marker 445 is designed to reflect the incident radar signal partially upward and sideways back toward the radar signal source (i.e., the transceiver 116 and / or 1016) on the vehicle 180.

[0174] The polarimetric radar transceivers 116 and 1016 and / or the lateral guidance controller 890 may filter reflections based on range, polarization, and / or magnitude to distinguish the target from other objects. The targets (e.g., pavement markers 445) are typically within a known range from the vehicle.

[0175] Figure 19 is a block diagram of a lateral guidance control system 970 that controls the vehicle’s steering to maintain its position in the lane. This embodiment represents an autonomous lateral control system. The front left and right polarimetric radar 50066 / 0501 Page 48 of 64transceivers, 116 and 117, respectively, transmit polarized signals in the direction of the targets and receive polarized signals reflected from them. The distances between the targets and the polarimetric radar transceivers are measured by transceivers 116 and 117, respectively, using well-established radar techniques. The outputs of the polarimetric radar transceivers 116 and 117 are combined with an IMU 907, providing the lateral guidance controller 8104 with the measured lateral distance error dERROR (dC-dX). The desired control distance from the target is subtracted from the radar-measured distance, dc, to determine an error, dERROR, which is calculated by the lateral guidance controller 8104. The vehicle deviation angle ψ relative to the navigational path trajectory is calculated from the projective trajectory of the navigational map and the measured positions by the GNSS sensor 901. Vehicle speed 8008 and road geometry preview information from the navigation database 8006 are also sent to the lateral guidance controller 8104. For redundancy, the IMU 907 and GNSS sensors 901 are used by the guidance controller 8104 to provide a more robust steering output. There may be additional sensors, such as multiple IMUs, RTK-GNSS, multiple vision cameras, and laser radars (not shown), that can be used to augment the lateral control system 970.

[0176] The output of the guidance controller 8104 controls a steering mechanism 8020 via a desired steering wheel command, dCMD, which determines the angle of the wheels. Along with the vehicle / wheel dynamics 8040, this command controls the lateral position of the vehicle in the lane relative to the targets. The lateral guidance controller system 970 can have various embodiments for control algorithms that control the vehicle’s steering along the pavement markers. Some examples include fuzzy control algorithms, proportional- integral-derivative (PID) control algorithms, adaptive PID control algorithms, Stanley control algorithms, model predictive control (MPC) control algorithms, neural network control algorithms, modified sliding mode control algorithms, multi-rate control algorithms, and linear quadratic regulator (LQR) control algorithms.

[0177] For example, the Stanley control algorithm can be utilized for the control algorithm within the lateral controller 8104. The lateral controller 8104 calculates the vehicle heading angle ψ of the road trajectory using the GNSS sensor 901 and the navigation database 8006 and the position error dERROR(t) (the desired distance minus the radar- 50066 / 0501 Page 49 of 64measured lateral distance dxfrom the polarimetric radar transceivers 116 and 117). The controller 8104 uses this information to control the steering wheel angle δ via the command wheel angle dCMD. The formula for the Stanley controller is as follows: dCMD(t)=ψ(t)+tan-1[ks*dERROR(t) / (ksv+vy(t)] where dCMD(t) is the steering wheel angle command a specific time t, ψ(t) is the vehicle heading angle related to the path trajectory, dERROR(t) is the position error related to desired path, vy(t) is the vehicle velocity, and ks and ksv are constants that can be tuned for desired performance.

[0178] The desired control distance, dC, from the target may be a predetermined distance based on the assumed lane width, vehicle width, and the location of the polarimetric radar transceiver on the vehicle. Alternatively, the desired control distance dCfrom the target may depend on the vehicle’s approximate location to account for different lane widths and / or varying target locations on various roads. For example, the database 8006 may store as attribute a set of desired control distances for different roads or geographic locations, and the database 8006 may be queried based on the geographic location of the vehicle, such as determined, for example, by the GNSS sensors 901, sufficient to decide on which road or set of nearby roads the vehicle is traveling on. Additionally, the desired control distance dC from the target may be dependent on the vehicle’s direction of travel, which may be determined from compass sensors and / or GNSS measurements, as lane widths and / or target locations may not be the same in different directions on the same road. The desired control distance dCmay be offset from the true lane center or varied slightly (e.g., with a small random component) to distribute tire wear across the lane more evenly.

[0179] In another embodiment, the desired control distance dCmay be dynamically determined as the vehicle travels down the road by measuring distances to different, laterally displaced target structures, such as the road surface or roadside targets. For example, the lane width can be indirectly measured by (a) measuring distance(s) to target structures marking an adjacent lane boundary on a given side of the vehicle, (b) measuring distance(s) to target structures marking the next lane boundary on the same side of the 50066 / 0501 Page 50 of 64vehicle, (c) subtracting those measured distances to calculate the adjacent lane width, and assuming that the vehicle’s current lane width is the same.

[0180] As shown in Figure 19, the vehicle is equipped with right-looking and left- looking polarimetric radar transceivers, two desired control distances (right and left) can be used, and the control strategy can be to equalize the two distances (e.g., using a zero difference between the right and left distances as the nominal setpoint) so that the vehicle is centered in its lane.

[0181] Figure 20 is a block diagram of a controller guidance system 960, which shows an HD database 8866, a controller 8004, a GNSS sensor 901, and several polarimetric radars 116, 117, 1016, and 1017. The database 8866 of road surface and roadside targets, along with their locations, is shown in more detail in Figure 15B and was described earlier. The database 8866 can be onboard the vehicle or external to the vehicle and accessed via wireless Internet 500. The database 8866 may contain details regarding the road targets along the trajectory for use by the onboard controller 8004 to guide or control the lateral direction of the vehicle moving along the road pathway. This embodiment represents an autonomous lateral control system. The polarimetric radar transceivers 116, 117, 1016, and 1017, mounted on the front and rear of both sides of the vehicle, respectively, transmit polarized signals in the direction of the target structures and receive polarized signals reflected from them. The distances between the target structures and the polarimetric radar transceivers 116, 117, 1016, and 1017 are measured by the radar transceivers using well- known radar techniques and provided to the lateral guidance controller 8004. The polarimetric radar transceivers 116 and 117, mounted on the front of the vehicle, measure the distance dX1(a combination of the two polarimetric radar transceivers) from target structures near the front of the vehicle. The polarimetric radar transceivers 1016 and 1017 measure the distance dX2 from different target structures from the rear of the vehicle. The desired control distance, dC, from the target structure is subtracted from the radar-measured distance, dX1, to determine the cross-track error, dERROR, which is then input to the lateral guidance controller 8004. Data from one or more of a vehicle speed sensor 8008, HD database 8866, IMU 907, vision sensor 903, GNSS sensor 901, and wheel angle 8028 may be sent to the lateral guidance controller 8004. There may be additional sensors, such as 50066 / 0501 Page 51 of 64multiple IMUs, RTK-GNSS, multiple vision cameras, and laser radar (not shown), that can be used to augment the lateral control system 960.

[0182] The output of the guidance controller 8004 controls a steering mechanism 8020, which in turn controls the angle of the wheels. Along with the vehicle dynamics 8040, this mechanism controls the lateral position of the vehicle in the lane relative to the pavement markers. The two radar transceivers 116 and 117 provide cross-track error, dERROR, and the front and rear radar distance measurements provide heading angle error relative to the targets. The lateral guidance controller system 960 can have various embodiments for control algorithms that control the vehicle’s steering along the pavement markers. Some examples include fuzzy control algorithms, proportional-integral-derivative (PID) control algorithms, adaptive PID control algorithms, Stanley control algorithms, model predictive control (MPC) control algorithms, neural network control algorithms, modified sliding mode control algorithms, multi-rate control algorithms, and linear quadratic regulator (LQR) control algorithms.

[0183] In this embodiment, the HD database 8866, which contains road trajectories, and the GNSS sensor 901 can provide both lateral and longitudinal information of the vehicle’s location based solely on the HD database 8866 and the GNSS sensor 901. Combining GNSS sensor / HD database information with polarimetric radar transceiver information provides a robust, accurate, and reliable control system. Variations

[0184] Figure 21 shows another embodiment in which multiple polarimetric radar transceivers, 117, 1017, and 1117, along with one or more IMUs 907 can be installed at several locations on the left side of the vehicle 180. Similarly, one or more polarimetric radar transceivers, such as 116, 1016, and 1116, can be installed along with one or more IMUs such 907 on the right side of the vehicle 180. The radar transceivers 116, 1016, 1116, 117, 1017, and 1117 can be installed in side marker lights, headlights, license plates, fog lights, daytime running lights, or turn signals on the vehicle, or under covers or within enclosures for the same, for example, for minimum impact on the exterior design of the vehicle 180. The radar transceivers can be mounted on or in a bumper (e.g., front bumper) of the vehicle 180. The 50066 / 0501 Page 52 of 64radar transceivers can be mounted behind the vehicle’s body exterior (e.g., behind a body panel) made of RF-transparent material, typically plastic and / or fiberglass. In another embodiment, one or more radar transceivers 117, 1017, and 1117 see pavement markers 1284 on the left, and one or more radar transceivers 116, 1016, and 1116 see pavement markers 284 on the right. Additionally, a radar transceiver 119 and 1019, along with an IMU 907, can be mounted in the center front or center rear of the vehicle 180, allowing it to view both directions, as well as the right and left sides of the vehicle 180, given the presence of pavement markers 284 and 1284. These additional radar transceivers can provide redundancy, heading angle information, and more timely distance measurements. There can be many other combinations of mounting one or more radar transceivers on the vehicle 180 to provide for more robust lateral guidance control.

[0185] The IMUs 907 located around the vehicle 180 can be part of the radar transceivers 116, 117, and 119 or separate standalone units. A separate IMU 907 is also shown in the center of the vehicle 180. The IMUs 907 can provide a change in distance measurements from the last side target that the radar transceivers viewed. Therefore, when IMU information (i.e., inertial measurements) is fused with the radar transceiver information, a distance measurement between side targets, e.g., pavement markers 284 and 1284, can be continuously, quasi-continuously, or approximately continuously determined or determined at discrete points between side targets. As used herein, the phrase “inertial measurement” refers to any data that an IMU can produce. Examples of inertial measurements include acceleration, yaw rate, heading, and changes of the foregoing.

[0186] Figure 21 also shows that a GNSS sensor 901 is mounted on the top and can be mounted on the front left and right sides. There can be one or more GNSS sensors or antennas mounted on the vehicle to provide better visibility to the satellites, for determining heading information by using two GNSS antennas spaced apart, and for redundancy.

[0187] Another embodiment (not shown) utilizes polarimetric radar transceivers to determine the position of a trailer on a roadway. Besides having a polarimetric radar transceiver on the front of the tractor unit to control the lateral movement of the tractor, the positions of the trailers can be determined using multiple side-looking and downward- 50066 / 0501 Page 53 of 64looking polarimetric radar transceivers on both sides of the trailers, which view the road surface or roadside targets.

[0188] For redundancy and / or higher reliability, the radar guidance techniques described herein can be combined with other guidance techniques based on different technologies. For example, the polarimetric radar transceivers 116 and 117 can be integrated with other sensors such as, one or more rear side or rear mounted polarimetric radar transceivers 1016, a GNSS receiver 901, lidar sensor 902, vision sensor 903, forward- looking radar 904, IMU 907, steering wheel angle sensor 8028, vehicle speed sensor 8008, or other location sensors 905, as shown in Figure 22. Additionally, an HD database 8866 of past routes, navigational maps, road surface targets, and roadside targets, as well as a navigation database 8006, can be integrated via the controller 437 for enhanced accuracy. As another example, the processing of signals from various sensors, including the radar sensors described herein, can be performed by a single processor or a computer (which may have multiple processors). Joint processing of various sensor types can enhance the performance of any single sensor type. For example, the output of an onboard processor 437 using, for example, one or more Kalman filters (KF), Linear Kalman Filter (LKF), Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF), and Particle Filter (PF) can provide more reliable location information 914, steering information 912, and speed information 918.

[0189] A typical FMCW radar sensor, such as the Texas Instruments AWR1843, can determine object range by performing an FFT on the processed received signal, which corresponds to the beat frequency. A radar transceiver can provide velocity by using a second FFT across several sequential received signals (chirps with ramped transmitted frequencies), detecting the change in phase from one chirp to the next. A radar transceiver can provide direction using a third FFT across several received signals from different antennas, utilizing AoA (Angle-of-Arrival). Adding polarization to the antennas, the orientation and target type can also be determined. This results in a radar cube comprising range, velocity, and direction, and an additional table for object type and orientation. Accurate lateral distances to a polarimetric radar transceiver can be determined using a single FFT on the first frequency ramp, which is typically 10-250 microseconds long, known 50066 / 0501 Page 54 of 64as a fast chirp in FMCW modulation. This fast chirp allows for distance measurements in short time intervals, e.g., less than one millisecond. Typically, the lateral velocity measurement is small and near zero; therefore, a second FFT is not required for targets that are lateral to the polarimetric radar transceiver. The radar processor can accurately measure lateral distances to the road surface or roadside targets when the radar beam is directed at them. Then, process distance and velocity on multiple receive antennas for the detection of other objects surrounding the vehicle when the radar beam sweeps vertically along the side with one or more polarization antennas. For example, blind spot detection can be accomplished via the radar cube with vertical polarization antennas when the radar beam sweeps along the side of the vehicle.

[0190] The radar beam transmitted by the polarimetric radar transceiver preferably has a single or dominant lobe, characterized by a specific beam direction and shape. The beam direction is typically characterized by a vector along the center or centroid of the main lobe. The radar transceiver is preferably placed and pointed so that the beam direction is toward where the road surface targets are expected to be when the vehicle is centered in its lane. The beam shape may be, for example, Gaussian and characterized by a beam width, typically measured from the half-power points on opposite sides of the beam center. The beam may be circularly symmetrical about the center direction vector or may be asymmetrical. Range-based gating or filtering can be employed to ignore reflection signals from objects outside of the range of interest.

[0191] The beam shape and / or direction may be static or variable. Well-known beam-shaping techniques can be applied to alter the beam either dynamically or non- dynamically. For example, when the beam is generated from an antenna array, well-known electronic beam steering and beam shaping techniques (e.g., delay-and-sum, windowing) can be employed to electronically alter (or steer) the beam. For example, the beam can switch between a relatively narrow width, pointed at a road surface target with circularly polarized antennas, and a relatively broader width, pointed more upwardly, for blind spot detection using vertically polarized antennas.

[0192] In some embodiments, mechanical beam shaping devices can be employed. For example, an RF lens can be used to alter a beam’s shape, including its width. If the radar 50066 / 0501 Page 55 of 64transceiver is located within a vehicle’s light assembly, then the exterior light cover (or a relevant portion thereof) can be shaped and constructed of a suitable material to serve as an RF lens. As another example, an RF reflector (e.g., horn) may also alter a beam’s shape and / or width.

[0193] Another embodiment combines a current automotive radar transceiver that provides object detection on the front, side, and / or rear of the vehicle with the ability to detect the distance to pavement markers or other road surface or roadside structures. The polarimetric radar transceivers can be used for both side-object detection (detecting a vehicle on the side) and ranging of on-road or roadside target structures. Not only can the same polarimetric radar transceivers be used for blind-spot detection, but also for cross- traffic alerts, lane-change assistance, and traffic-jam assistance using, for example, known beam steering techniques with the multiple antennas on or with the same radar transceiver unit. This could be, for example, a dual-purpose side radar transceiver that processes side objects, such as vehicles in blind spots, and processes on-surface objects, such as reflective pavement markers, for lateral distance measurements.

[0194] There are various types of antenna polarization. Vertical and horizontal polarized antennas are the most common and are currently used on automotive radar. However, there are orthogonally, linearly, circularly, and elliptically polarized antennas. GNSS sensors usually use circularly polarized antennas. In some embodiments of the polarimetric radar transceiver, circularly polarized antennas are used, both RHCP and LHCP. There are different constructions of circularly polarized antennas, including, for example, microstrip patch antennas, wire antennas helical antennas, dielectric resonator antennas, helix antenna, slot antennas, gap waveguide technology antennas, meta surface, horns, cross dipoles, spiral antennas, QHA / PQHA antennas, circular polarized antenna arrays, and dual- frequency circular polarization antenna waveguide technology, meta surface, horns, cross dipoles, spiral antennas, CP (circularly polarized) Patch, QHA / PQHA antennas, and dual- 50066 / 0501 Page 56 of 64frequency circular polarization antennas. Another embodiment of CP antennas is an array of multiple CP antennas spaced to provide a narrow and steerable radar beam. CONCLUSION

[0195] The foregoing embodiments, descriptions, and terms are set forth by way of illustration and example only and are not meant as limitations. For example, while the foregoing description often refers to a pavement marker as a specific example of a road surface target, this is merely an illustration. The principles presented herein are equally applicable to other types of road surfaces, and other road surface or roadside targets can be used in place of pavement markers in the examples given above. The scope of the invention should be determined only by the following claims, claims presented in continuing patent applications or a post-issuance proceeding, and equivalents to such claims. 50066 / 0501 Page 57 of 64

Claims

CLAIMS 1. A polarimetric radar system configured to be installed in or on a land vehicle and for use with a radar target on a road surface or on a roadside, wherein the radar target is on or near a pathway along which the land vehicle can travel, the polarimetric radar system comprising: a radar transmitter configured to transmit sideways from the vehicle one or more polarized transmit signals; a radar receiver configured to receive sideways from the vehicle one or more polarized reflection signals; and a processor, coupled to the radar receiver, configured to determine, based on at least one of said one or more polarized reflection signals, a classification of the radar target.

2. A polarimetric radar system according to claim 1, wherein the radar transmitter is configured to transmit the one or more polarized transmit signals partially downward, and wherein the radar receiver is configured to receive the one or more polarized reflection signals directed partially upward.

3. A polarimetric radar system according to claim 1, wherein the polarized transmit signals and the polarized reflection signals are of a type selected from a group consisting of vertically polarized, horizontally polarized, orthogonally polarized, circularly polarized, and elliptically polarized.

4. A polarimetric radar system according to claim 1, wherein the radar transmitter is characterized by a transmission beam pattern having a center direction of maximum transmission strength, and the radar transmitter is configured to be installed in or on the land vehicle so that the center direction of maximum transmission strength is downward and sideways from the land vehicle when the radar transmitter is installed on or in the land vehicle; and the radar receiver is characterized by a reception beam pattern having a center direction of maximum reception strength, and the radar receiver is configured to be installed in or on the land vehicle so that the center direction of maximum reception strength is 50066 / 0501 Page 58 of 64upward and sideways generally toward the radar receiver when the radar receiver is installed on or in the land vehicle.

5. A polarimetric radar system according to claim 1, wherein the radar target is on the road surface and selected from a group consisting of a raised pavement marker, a snow- plowable pavement marker, a slot in the road surface, a groove in the road surface, a rumble strip, a bott dot, a cat’s eye marker, a curb, a bottom of a concrete barrier, a road surface edge, a vegetation edge, a drain, and a road stud.

6. A polarimetric radar system, according to claim 5, wherein the radar target is a roadside target selected from a group consisting of a concrete barriers, a guardrail, a building, a sign, a sign post, a lamp, a lamp pos, a fire hydrant, a utility pole, a tree, tall grass, and a vehicle.

7. A polarimetric radar system according to claim 1, wherein the radar transmitter and the radar receiver operate within a frequency range from about 300 MHz to about 1 THz.

8. A polarimetric radar system according to claim 1, wherein the radar transmitter and the radar receiver are of a frequency-modulated continuous wave type.

9. A polarimetric radar system according to claim 1, wherein the radar transmitter and the radar receiver are monostatic.

10. A polarimetric radar system according to claim 1, wherein the radar transmitter and the radar receiver are bi-static.

11. A polarimetric radar system according to claim 1, further comprising: an antenna array.

12. A polarimetric radar system according to claim 1, wherein the radar transmitter and the radar receiver are configured to be placed within a front bumper of the vehicle.

13. A polarimetric radar system according to claim 1, wherein the radar transmitter and the radar receiver are packaged within a vehicle light assembly. 50066 / 0501 Page 59 of 6414. A polarimetric radar system according to claim 1, wherein the signal processor is configured to process just a portion of said at least one of said one or more polarized reflection signals received in a window corresponding to an expected range of a reflection target.

15. A polarimetric radar system according to claim 1, wherein information regarding the radar target is stored in a database, and wherein the system is configured to determine, based on the classification of the radar target and information in the database, a lateral distance between the vehicle and the radar target, wherein the system is further configured to determine a lateral position of the vehicle relative to the pathway based on said lateral distance between the vehicle and the radar target.

16. A polarimetric radar system according to claim 15, wherein the information regarding the radar target in the database comprises location information.

17. A polarimetric radar system according to claim 16, wherein the location information comprises absolute coordinates.

18. A polarimetric radar system according to claim 16, wherein the location information comprises relative location information with respect to the pathway or a feature of the pathway.

19. A polarimetric radar system according to claim 1, wherein information regarding the radar target is recorded in an automotive navigation database, the system further configured to receive vehicle trajectory information from a GNSS sensor on or in the vehicle, and wherein the system is configured to determine, based on the classification of the radar target, GNSS-measured coordinates, and information in the database, a lateral distance between the vehicle and the radar target and a direction of vehicle travel.

20. A polarimetric radar system according to claim 1, wherein the processor is further configured to determine the classification of the radar target by determining one or more polarization scattering properties of the target. 50066 / 0501 Page 60 of 6421. A polarimetric radar system according to claim 20, wherein the processor is further configured to determine the classification of the radar target by determining a polarization scattering matrix for the target.

22. A method for recognizing a radar target on or near a pathway along which a land vehicle travels, the radar targets having different radar reflectivity characteristics compared to the pathway, the radar targets being arranged along the pathway in a direction of vehicle travel, the method comprising: transmitting sideways at least one polarized radar transmit signal from the land vehicle towards a radar target as the vehicle passes the radar target; receiving at the land vehicle at least one polarized radar reflection signal from the radar target as the vehicle passes by the radar target; and determining, based on said at least one polarized radar transmit signal and said at least one polarize radar reflection signal, a type of the radar target.

23. A method according to claim 22, wherein the steps of transmitting, receiving, and determining are repeated for substantially every radar target that the vehicle passes.

24. A method according to claim 22, wherein the pathway is one lane of a multi-lane road, and wherein the radar targets comprise pavement markers that are distributed along a line demarking one side of the pathway.

25. A method according to claim 22, further comprising: filtering said at least one polarized radar reflection signal based on a range of expected distances between the land vehicle and the radar target.

26. A method according to claim 22, further comprising: determining, based on one or more said at least one polarized radar transmit signal and one or more of said at least one polarize radar reflection signal, a distance between the vehicle and the radar target. 50066 / 0501 Page 61 of 6427. A method according to claim 26, further comprising: using the distance to guide or assist steering the vehicle laterally to follow the pathway.

28. A method according to claim 26, further comprising: providing a warning either if the distance exceeds a maximum threshold or is less than a minimum threshold, or if an approximate lateral position of the vehicle is outside of an acceptable range.

29. A method according to claim 28 wherein the warning is an internal warning to a driver of the vehicle.

30. A method according to claim 28, wherein the warning is a warning external to the vehicle.

31. A method according to claim 26, further comprising: performing one or more of a lane-centering assistance function and a lane-keeping function based on the distance.

32. A method according to claim 26, wherein the vehicle comprises a lateral-positioning system independent of the method of claim 24, the method further comprising: using the distance in conjunction with the independent lateral-positioning system to improve lateral-position determination.

33. A method according to claim 28, wherein the lateral-positioning system comprises a vision system viewing painted lines along the pathway.

34. A method according to claim 28, wherein the lateral-positioning system comprises a lidar system viewing painted lines along the pathway, recorded targets along the pathway, or both. 50066 / 0501 Page 62 of 6435. A method according to claim 22, further comprising: using the classification of the radar target to access information about the radar target in a database.

36. A method according to claim 35, further comprising: receiving GNSS location information for the vehicle; and using the GNSS location information, in conjunction with the classification of the radar target, to access the information about the radar target in the database.

37. A method according to claim 33, wherein the information in the database comprises location information for the radar target.

38. A method according to claim 37, further comprising: determining, based on the location information for the radar target in the database and a radar-measured distance between the vehicle and the radar target, a lateral position of the vehicle relative to the pathway or a feature of the pathway.

39. A method according to claim 22, wherein determining the type of the radar target comprises determining one or more polarization scattering coefficients for the radar target.

40. A method according to claim 39, wherein determining one or more polarization scattering coefficients for the radar target comprises determining a polarization scattering matrix for the target. 50066 / 0501 Page 63 of 64

Images