Photometric registration of installed RFID tag readers

Abstract

RFID tag readers can be used to interrogate and locate RFID tags, which can be useful for tracking items to which the RFID tags are attached. In operation, an RFID tag reader measures the angle of arrival and / or amplitude of the tag's reply, which can then be used to estimate the tag's location relative to the RFID tag reader. This location estimate can be translated to a global coordinate frame. But the accuracy of the location estimate in the global coordinate frame depends in large part on how accurately the locations and orientations of RFID tag readers are known in the global coordinate frame. Registering the locations of the installed RFID tag readers to an accurate representation of the environment improves the accuracy of the RFID tag readers' location estimates of the RFID tags.

Description

Photometric Registration of Installed RFID Tag ReadersCROSS-REFERENCE TO RELATED APPLICATION(S)

[0001] This application claims the priority benefit, under 35 U.S.C. 119(e), of U.S. Application No. 63 / 619,984, filed on January 11, 2024, which is incorporated herein by reference in its entirety for all purposes.BACKGROUND

[0002] Radio-frequency identification (RFID) tags are low-cost devices that can be attached to objects and offer the promise of automated tracking, locating, sales check-out, and inventory of the objects among other commercial and medical applications. There are passive, semiactive, and active types of RFID tags that can be wirelessly interrogated by an RFID tag reader, also called an RFID reader or simply a reader, and emit wireless RF replies to the reader. Each RFID tag’s reply can include information stored in that RFID tag, such as a tag identification number or alpha-numeric sequence and electronic product code (EPC). Other information may be included with the reply.

[0003] Passive RFID tags have no battery and are therefore typically lower cost than semiactive and active RFID tags. Passive RFID tags essentially modulate and backscatter energy from an RF interrogation pulse sent by the RFID reader to transmit a reply. Although passive RFID tags are lower in cost (e.g., on the order of l / 10ththe cost of an active tag) than active and semi -active RFID tags, reply signals from passive RFID tags are weaker at longer ranges. In addition, the read range of a passive RFID tag can be significantly shorter than the read range of an active tag. For example, the read range of a passive tag may be limited to 100 meters in a line-of-sight environment where there are no intervening objects to scatter the interrogation and reply signals, whereas the read range of an active tag may be over 500 meters in the same environment.

[0004] When an RFID reader is deployed in a setting with a dense population of RFID tags (e.g., at least 10 RFID tags per square meter) and that includes objects which scatter RF signals, the RFID reader may not be able to read all of the tags even if the tags are located within 100 meters of the reader. Such settings occur in warehouse and retail sales facilities, for example. In such cases, it may be necessary to install multiple RFID readers throughout the facility in order to read all of the tags.SUMMARY

[0005] An RFID tag reader can be installed or deployed at an installation site as follows. Before installation, an installer, inspector, surveyor, or another person obtains a first representation (e.g., a three-dimensional lidar scan) of the installation site (e.g., with a portable lidar scanner). The installer installs the RFID tag reader at the installation site. After installation, the installer, inspector, surveyor, or other person obtains a depth map of the RFID tag reader and at least a portion of the installation site surrounding the RFID tag reader (e.g., with a smart phone camera or other suitable camera). A computer processor derives a second representation of the installation site from the depth map, registers the second representation to the first representation, and determines a position and / or an orientation of the RFID tag reader with respect to a coordinate system of the installation site based on registration of the second representation to the first representation.

[0006] Installing the RFID tag reader in the installation site may include adjusting a pitch, a yaw, and / or a roll of the RFID tag reader to within 0.25° of a desired angle.

[0007] Registering the second representation to the first representation can include warping, translating, rotating, and / or scaling the second representation.

[0008] Before the second representation is registered to the first representation, the processor may elide or remove the RFID tag reader from the second representation. Eliding the RFID tag reader from the second representation can include obtaining an image (e.g., a color image) of the RFID tag reader and at least a portion of the installation site surrounding the RFID tag reader after installation of the RFID tag reader at the installation site, segmenting the RFID tag reader in the image, and removing points corresponding to segments of the RFID tag reader from the second representation. The processor can segment the RFID tag reader by matching the image to a computer-aided design (CAD) model of the RFID tag reader. Alternatively, the processor can elide the RFID tag reader from the second 3D representation by registering a model of the RFID tag reader to the depth map and removing points corresponding to the model of the RFID tag reader from the second representation (e.g., points within a predetermined distance of the model of the RFID tag reader).

[0009] Determining the position and / or the orientation of the RFID tag reader with respect to the coordinate system of the installation site may include translating the second 3D representation with respect to the first representation based on an image of the RFID tag reader and at least a portion of the installation site surrounding the RFID tag reader after installation.

[0010] If desired, the processor can determine a yaw orientation of the RFID tag reader based on at least one fiducial mark on the RFID tag reader. For instance, the processor can determine the yaw orientation of the RFID tag reader by matching a template of the fiducial mark(s) to an image of the RFID tag reader and at least a portion of the installation site surrounding the RFID tag reader after installation.

[0011] Once the RFID tag reader has been deployed, it can be used to interrogate and locate an RFID tag as follows. First, the RFID tag reader transmits a signal to an RFID tag. Then the RFID tag reader receives a reply to the signal from the RFID tag. The RFID tag reader or an appliance or controller operably coupled to the RFID tag reader estimates a location of the RFID tag with respect to the installation site based on the reply and the position and / or the orientation of the RFID tag reader with respect to the coordinate system of the installation site.

[0012] In some cases, the RFID tag reader can use channel estimates or signatures to locate the RFID tag. In these cases, the RFID tag reader derives a channel estimate for a communications channel to an RFID tag and compares that channel estimate to one or more previously obtained channel estimates. These other channel estimates can be obtained after the RFID tag reader has been installed as follows. A camera, such as on a smart phone or tablet, acquires an image of a portion of the installation site surrounding a (reference) RFID tag. A processor, either in the smart phone or tablet or in a cloud-based server or another device, determines a position of the RFID tag in the coordinate system of the installation site based on the image of the portion of the installation site surrounding the RFID tag. The RFID tag reader receives a reply from an RFID tag and determines, based on the reply, a channel estimate representing a communications channel between the RFID tag reader and the RFID tag. This channel estimate is associated with the position of the RFID tag in the coordinate system of the installation site and stored for future use.

[0013] The RFID tag can also be mounted to a drone, in which case a camera mounted to the drone can acquire the image of the portion of the installation site. In these cases, determining the position of the RFID tag can include using visual inertial odometry to determine the position of the drone. Determining the position of the RFID tag in the coordinate system of the installation site can also include registering the image of the portion of the installation site surrounding the RFID tag to the first representation of the installation site.

[0014] In other cases, the RFID tag can be associated with a quick response (QR) code or bar code. In these cases, acquiring an image of the QR code or bar code triggers interrogation ofthe RFID tag by the RFID tag reader. The image of the QR code or bar code can be the same image that shows the portion of the installation site and is used to determine the RFID tag’s location in the coordinate system of the installation site. The RFID tag reader can receive the reply from the RFID tag as part of this interrogation.

[0015] (Channel estimates can be acquired in this fashion regardless of how the RFID tag reader is installed or deployed.)

[0016] An RFID tag reader deployed in a fixed position in this way forms part of a system for locating RFID tags that also includes an interrogator controller. The interrogator controller is operably coupled to the RFID tag reader, which may be shifted and / or rotated with respect to is desired position and / or orientation. The interrogator controller is configured to estimate an angle of arrival at the RFID tag reader of a reply received by the RFID tag reader in a coordinate system centered on the RFID tag reader. The interrogator controller translates the angle of arrival from the coordinate system centered on the RFID tag reader to a coordinate system of the environment based on registration of a position and / or an orientation of the RFID tag reader with respect to the coordinate system of the environment. The interrogator controller can also estimate a location of the RFID tag in the coordinate system of the environment based on the angle of arrival in the coordinate system of the environment.

[0017] All combinations of the foregoing concepts and additional concepts discussed in greater detail below (provided such concepts are not mutually inconsistent) are contemplated as being part of the inventive subject matter disclosed herein. In particular, all combinations of claimed subject matter appearing at the end of this disclosure are contemplated as being part of the inventive subject matter disclosed herein. The terminology explicitly employed herein that also may appear in any disclosure incorporated by reference should be accorded a meaning most consistent with the particular concepts disclosed herein.BRIEF DESCRIPTIONS OF THE DRAWINGS

[0018] The skilled artisan will understand that the drawings primarily are for illustrative purposes and are not intended to limit the scope of the inventive subject matter described herein. The drawings are not necessarily to scale; in some instances, various aspects of the inventive subject matter disclosed herein may be shown exaggerated or enlarged in the drawings to facilitate an understanding of different features. In the drawings, like referencecharacters generally refer to like features (e.g., functionally similar and / or structurally similar components).

[0019] FIG. 1 depicts an RFID environment with several ceiling-mounted RFID tag readers, also called readers or sensors, for querying and locating passive RFID tags.

[0020] FIG. 2A shows a bottom perspective view of a reader with a four-element, square antenna array.

[0021] FIG. 2B depicts a perspective cross-sectional view of the reader of FIG. 2A.

[0022] FIG. 2C depicts a mounting arrangement for a reader in a drop ceiling.

[0023] FIG. 2D depicts an arrangement for mounting a reader from a drop ceiling.

[0024] FIG. 2E depicts a reader mounted with an adjustable mount.

[0025] FIG. 3 A depicts a reader that is aligned properly in azimuth and lateral position.

[0026] FIG. 3B depicts a reader that is misaligned in azimuth and the error in tag position due to the azimuthal misalignment.

[0027] FIG. 3C depicts a reader that is misaligned laterally and the error in tag position due to the lateral misalignment.

[0028] FIG. 4A depicts a reader that is aligned properly in elevation angle and height.

[0029] FIG. 4B depicts a reader that is misaligned in elevation angle and the error in tag position due to the elevation angle misalignment.

[0030] FIG. 4C depicts a reader that is misaligned in height and lateral position and the error in tag position due to the height and lateral misalignment.

[0031] FIG. 5 illustrates a process for measuring and registering the actual location and orientation of an installed reader.

[0032] FIG. 6A illustrates a process for registering an installed reader using a color image and a depth map captured with a suitable camera.

[0033] FIG. 6B illustrates an alternative process for registering an installed reader using a color image and a depth map captured with a suitable camera.

[0034] FIG. 7A shows a point cloud of the ceiling of an installation site before any readers have been installed.

[0035] FIG. 7B illustrates a portion of the pre-installation point cloud of FIG. 7 A.

[0036] FIG. 8A shows an image (e.g., a color or grayscale image) of an installed reader next to an HVAC duct in the installation site of FIG. 7.

[0037] FIG. 8B shows a depth image of the installed reader of FIG. 8 A.

[0038] FIG. 9 illustrates a computer-aided design (CAD) model of the sensor registered to a point cloud derived from the depth image of FIG. 8B.

[0039] FIG. 10 illustrates a point cloud derived from the image of FIG. 8 A and / or depth map of FIG. 8B with the sensor removed from the point cloud.

[0040] FIG. 11 illustrates the sensor-less, post-installation point cloud of FIG. 10 registered to a section of the pre-installation point cloud of FIGS. 7A and 7B.

[0041] FIG. 12 illustrates a collection of sensor-less, post-installation point clouds derived from depth maps of different installed sensors registered to the pre-installation point cloud of FIGS. 7A and 7B.

[0042] FIG. 13 illustrates determination of the yaw angle of an installed sensor from the alignment of fiducials on the CAD model of the sensor to the corresponding fiducials in the image or depth map of the installed sensor.

[0043] FIG. 14 illustrates a process of collecting signatures / channel estimates with installed sensors from panels with embedded or attached RFID tags.

[0044] FIG. 15 illustrates a process of collecting signatures / channel estimates with installed sensors from a drone with an embedded or attached RFID tag.DETAILED DESCRIPTION

[0045] FIG. 1 depicts a multipath RFID environment 100 with a dense population of passive RFID tags 101 and one or more RFID readers 150, also called RFID tag readers, tag readers, interrogators, or sensors, which communicate with the RFID tags 101. The RFID tags 101 can be attached to objects, which can be identified and tracked using the RFID tags 101 and sensors 150. The RFID environment 100 can be in a retail store or warehouse, for example, though other settings are possible. There can be furnishings 120 in the environment that affect RF signals (e.g., block, attenuate, and / or scatter the RF signals). The furnishings 120 can include shelves, racks, cabinets, etc. that may be used to hold the objects to which at least some of the RFID tags are attached. For example, some furnishings 120 may comprise metal shelving that holds one or more retail items (not shown in FIG. 1) that are tagged with the RFID tags 101.The furnishings 120 can be arranged in rows in some settings, with aisles separating the rows to allow access to objects tagged with the RFID tags 101. The RFID environment 100 can be bounded by the ceiling, floor, and walls 110, 112, any of which can reflect or scatter RF signals from the sensors 150 and / or RFID tags 101. There may be one or more cameras 130 installed in the RFID environment to capture images of at least portions of the environment. The sensors 150 and cameras 130 can be communicatively coupled to a interrogator controller 140, also called a controller, central controller, or appliance, which can receive and process data from the sensors 150 and the cameras 130.

[0046] The sensors 150 are preferably installed in the RFID environment 100 such that together they can communicate with every RFID tag 101 in the RFID environment 100. The sensors 150 may communicate via one or more ethernet cables (not shown) with the interrogator controller 140, which may also supply electrical power from the interrogator controller 140 to the sensors 150. The interrogator controller 140 may be a computer, laptop, smart phone, or purpose-built device with a processor and non-volatile computer-readable memory adapted to communicate with the RFID readers 150 and issue commands recognizable to the RFID readers 150. The interrogator controller 140 can also receive signals from the RFID readers 150.RFID Readers / Sensors

[0047] RFID tag readers 150 can be used in retail stores, warehouses, supply rooms, libraries, museums, galleries, or other environments for tracking objects with RFID tags 101. Generally, the readers 150 are mounted or hung from the ceiling, e.g., at a height in a range from 8 feet to 20 feet above the floor, so that they emit interrogation signals downward, toward RFID tags 101 on shelves, tables, clothing racks, or other storage units. The RFID readers 150 are arranged so that they provide adequate coverage of the RFID environment 100, e.g., on a grid with a spacing chosen so that all or substantially all of the volume of the RFID environment 100 can be interrogated by the RFID tag readers 150. The RFID tag readers 150 can be arrayed in a regular (periodic) fashion as shown in FIG. 1 or distributed irregularly across the RFID environment.

[0048] FIGS. 2A and 2B depict an example of an RFID reader 150 that includes a phased antenna array having four antenna elements 205. The illustration of FIG. 2A is a perspective view from the bottom of the reader 150. The reader 150 also includes a ground plane 220 inside an enclosure or housing 280 that extends across and behind the antenna elements 205 withrespect to the emission direction of an RF beam from the reader. FIG. 2A also shows the reader’s pitch, roll, and yaw rotational axes.

[0049] The enclosure 280 is roughly disc shaped and can be made of acrylonitrile butadiene styrene (ABS), polyvinyl chloride (PVC), vinyl, or another suitable material that is substantially transparent at the carrier frequencies of the interrogation signals and replies. The enclosure 280 may include thinner sections 282 adjacent to the antenna elements 205. These thinner sections 282 can be 1 / 8” thick or less so that they transmit in-band RF energy to and from the antenna elements 205 without significant attenuation of the transmitted RF energy. The enclosure 280 may also have visible fiducial markings or may have a shape (e.g., an asymmetric shape) that facilitates determining the installed reader’s pose or orientation as explained below.

[0050] FIGS. 2C and 2D illustrate two techniques for mounting the RFID reader 150 with respect to a drop ceiling. A drop ceiling, also called a dropped ceiling, T-bar ceiling, or suspended ceiling, is a secondary ceiling that is suspended from the main ceiling or other overhead structure. Drop ceilings are very common in retail and office environments. A drop ceiling typically has a suspension grid 230 that hangs from the main ceiling or overhead structure and defines square and / or rectangular openings or cells. The suspension grid can be made from metal or plastic and is configured to hold ceiling tiles or fluorescent lights in the openings. In the United States and Canada, these openings are typically 24 inches by 24 inches (610 mm by 610 mm) or 24 inches by 48 inches (610 mm by 1220 mm) and accommodate ceiling tiles and light fixtures. In Europe, the grid opening size is typically 600 mm by 600 mm or 600 mm by 1200 mm. The ceiling tiles and fixtures can be slightly (5 mm) smaller than the grid opening size (e.g., smaller by 5 mm in Europe at 595 mm by 595 mm or 595 mm by 1195 mm).

[0051] The antenna elements 205 can be mounted in, above, or below an opening in the suspension grid 230 of the drop ceiling. In FIG. 2C, the reader 150 is housed in a box 240 that fits within an opening and is held in place by the suspension grid 230. The antenna elements 205 may be approximately the same height as the grid 230 and the ground plane 220 may be located above the grid 230. In some cases, an aesthetic panel 232 can be used to cover all or a portion of the RFID reader 150 and the opening in the grid 230. The aesthetic panel 232 can be made from plastic, such as ABS, PVC, Vinyl, or a blend thereof, and be thin enough so as to insignificantly attenuate RF signals propagating to and from the antenna elements 205.

[0052] An alternative mounting arrangement is shown in FIG. 2D, where the RFID reader 150 is mounted so that it extends below the grid 230. Such a mounting arrangement may be used when the suspension grid 230 comprises metal that may interfere with RF signals to and from the antenna elements 205. In either mounting arrangement, the box 240 can contain and / or couple to other components, such as an adjustable mount 210 shown in FIG. 2E, and Ethernet cables to communicate with the interrogator controller 140 and other readers 150.

[0053] One way to aid in alignment of the RFID readers 150 is to mount them with an adjustable mount 210, as depicted in FIG. 2E. The adjustable mount 210 may attach to the ceiling, box 240 as in FIG. 2C or FIG. 2D, or to another overhead structure 250, and the RFID reader 150 can mount to the adjustable mount 210. The adjustable mount can provide three or more degrees of freedom (e.g., three rotational degrees of freedom and one translational (vertical) degree of freedom) to adjust the location and / or orientation of the RFID reader 150. In some implementations, the adjustable mount can be used to adjust the RFID reader’s pitch, yaw, and roll about the rotational axes shown in FIG. 2A. These degrees of adjustment can allow the RFID reader to be leveled e.g., with respect to the floor of the RFID environment) and aligned with the x, y, z coordinate system of the RFID environment 100 (e.g., as shown in FIG. 1). The adjustable mount 210 can also be used to raise or lower the reader 150 and possibly to translate the reader 150 in a plane parallel to the floor or ceiling.Locating RFID Tags with RFID Readers / Sensors

[0054] Referring again to FIG. 1, the interrogator controller 140 and the readers 150 can locate the RFID tags 101 based on the tags’ replies to signals from the readers 150. For example, the interrogator controller 140 can issue a command to inventory the RFID tags 101 (and attached items) in the RFID environment 100 or to locate one or more RFID tags 101 (and attached item(s)) in the RFID environment. In response to these commands, one or more of the readers 150 transmit signals to the RFID tags 101 and receive the tags’ replies. The RFID readers 150 transmit the tags’ replies or indications of the tags’ replies (e.g., received signal strength indicators (RSSIs), angles of arrival (AOAs), channel estimates, EPCs or other information derived from or encoded in the replies, etc.) to the interrogator controller 140. The interrogator controller 140 uses this information to identify and locate the RFID tags 101 that replied to the readers 150.

[0055] The sensors 150 and / or interrogator controller 140 locate the tags 101 in an environmental coordinate system, also called a coordinate frame, frame of reference, orreference frame, fixed with respect to the environment 100. In FIG. 1, the environmental coordinate system is a Cartesian coordinate system with the origin in one comer and x and y axes (lower left) extending parallel to the floor and ceiling. The z axis (not shown) extends perpendicular to the x and j’ axes, i.e., into and out of the plane of FIG. 1. Other coordinate systems are also possible, including Cartesian coordinate systems centered at other points within the environment or cylindrical or spherical coordinate systems with appropriately selected origins. The interrogator controller 140 provides information to sales associates, managers, customers, and / or other users about the estimated location(s) of the tag(s) 101 in the environmental coordinate system or with reference to points registered to the environmental coordinate system (e.g., entrances, exits, sales counters, etc.).

[0056] The accuracy of the estimated locations depends in part on how accurately the sensors’ locations and orientations are known in the environmental coordinate system. This is because the sensors 150 and / or interrogator controller 140 use location estimation methods based on measurements of the tags’ locations relative to the sensors 150. For instance, the interrogator controller 140 can triangulate a given tag’s location from the AO As of replies from that tag 101 as measured by two or more different sensors 150. The interrogator controller 140 can also estimate a tag’s location from the intersection of a ray along the AOA measured by a single sensor 150 and a plane that contains or is assumed to contain the tag 101, e.g., at a height of 1 meter above the ground. The interrogator controller 140 can estimate the tag’s location using multilateration from estimates of distance (e.g., as represented by RSSIs) from the tag to several different sensors 150 or from an estimate of distance between the tag 101 and one sensor 150 and an assumption about the tag’s height. Signature matching involves correlating a channel estimate derived from a reply detected by a given sensor 150 to previously recorded channel estimates between that sensor 150 and known locations in the environment 100 as described in greater detail below.

[0057] If the RFID readers’ locations and orientations in the coordinate system are known, then the appliance 140 or reader(s) 150 can translate tag location estimates relative to the readers’ locations into absolute tag location estimates — that is, tag location estimates given in coordinates of the environmental coordinate system. Translating the tag location estimates into the environmental coordinate system enables the appliance 140 to supply the coordinates of a tag’s estimated location to an app on a smartphone, tablet, or other device that shows the tag’s estimated location on a map of the environment 10, or relative to a wall, doorway, fixture, or other readily identifiable reference point in the environment 100. Providing absolutecoordinates or coordinates with respect to a fixed reference point instead of a reader 150 makes it easier to corroborate or average tag location estimates derived from measurements by different sensors 150 and to locate tags 101 and any corresponding objects.

[0058] FIGS. 3 A-3C and 4A-4C illustrate how errors in the orientation and position of a reader can produce errors in estimated tag locations. FIG. 3 A is a plan view of a reader 150 that is mounted to a ceiling of an RFID environment, such as a retail store, warehouse, or manufacturing facility. The reader 150 aligned to the environmental coordinate system 300, which in this case is a Cartesian coordinate system. (Cylindrical and spherical coordinate systems are also possible.) This coordinate system 300 is fixed with its origin at a known, readily identifiable location, such as a corner of a room or indoor space that forms (part of) the environment. The coordinate system 300 provides a suitable reference frame for a person to find a tag 101 located by the reader 150 and the interrogator controller 140 (FIG. 1).

[0059] For example, the reader 150 and / or the interrogator controller 140 can generate an estimated location 301 for the tag from the AO A 303 of a reply from the tag 101 to the reader 150 and an assumption or knowledge about the tag’s height or the AOA between the tag 101 and another reader (not shown). Because the reader 150 — and hence the reader’s coordinate system or reference frame — are properly aligned in azimuth and lateral position with respect to the coordinate system 300, the tag’s estimated location 301 coincides with the tag’s actual position (indicated by tag 101).

[0060] Misalignment of the reader 150 in azimuth (yaw) or lateral position produces errors in the location estimates produced by the interrogator controller 140 from the reader’s measurements of the tag’s replies. In FIG. 3B, the reader 150 is rotated in azimuth with respect to the coordinate system 300. Put differently, the coordinate system 300’ centered on the reader 150 is rotated with respect to and therefore does not coincide with the coordinate system 300 of the RFID environment. As a result, the reader 150 measures an AOA 303’ that is rotated with respect to the environmental coordinate system 300. Unless this rotation is undone or compensated, the sensor 150 or interrogator controller 140 will estimate a tag location 301’ that is offset from the tag’s true location from the AOA 303’. Similarly, a shift in the reader’s lateral position with respect to the environmental coordinate system 300 (i.e., a shift in the reader’s coordinate system 300”) yields a shifted AOA 303” and a shifted estimated location 303” for the tag 101 as shown in FIG. 3C.

[0061] Misalignment of the reader 150 in elevation angle (roll or pitch) or vertical position also produces errors in tag location estimates. FIG. 4A shows a profile view of a reader 150 aligned properly to the environmental coordinate system 300 as well as a plane 400 at an assumed height (e.g., about 1-2 meters above the floor) of the tags 101. The AOA 403 measured between the tag 101 and the reader 150 intersects the plane 400 at the tag’s true position, yielding an accurate location estimate 401 as shown in FIG. 4 A. Tilting the reader 150 in elevation angle (pitch or roll) tilts the reader’s coordinate system 300’” with respect to the environmental coordinate system 300, yielding a measured AOA 403’” and a location estimate 401’” that are tilted and shifted, respectively, with respect to their nominal values as shown in FIG.4B. Likewise, FIG. 4C shows that shifting the reader 150 in height and lateral position shifts the reader’s coordinate system 300”” with respect to the environmental coordinate system 300, yielding similarly shifted measured AOA 403”” and location estimate 401””.

[0062] In many cases, the misalignments or rotational and positional errors between the readers and the environment are unknown. These errors can also vary from reader to reader in an uncorrelated fashion. For example, one reader may be shifted horizontally to accommodate a vent, whereas another reader may be tilted and shifted vertically to fit around a nearby pipe or conduit. The errors in sensor positions and orientations produce error or uncertainty in the tag location estimates produced from the sensors’ measurements of tag replies.

[0063] The interrogator controller 140 and / or readers 150 compensate for these misalignments, rotational errors, and / or positional errors by translating the measurements of AO As and other quantities made in the readers’ relative frames of reference into the common environmental frame of reference. These translations are based on measurements of the readers’ actual positions and orientations — that is, the positions and orientations of the installed readers 150 rather than their desired positions and orientations. Registering the readers’ actual positions and orientations to the common environmental frame of reference or coordinate system as described below with respect to FIGS. 5-13 yields the rotations and / or translations that the interrogator controller 140 and / or readers 150 can apply, e.g., as appropriate rotation and / or translation matrices, to the AO As and other quantities measured by the readers 150.

[0064] The interrogator controller 140 can use AO As measured by different readers 150 in their respective relative coordinate frames and translated into the common environmental coordinate frame to locate RFID tags more accurately. To see how the translation improves accuracy, consider three readers 150 that are offset and / or rotated with respect to their intendedpositions and orientations and that measure AO As to the same tag. Without any correction or compensation for the readers’ offset and / or rotational misalignment, the AO As, when projected into the common environmental reference frame, will not intersect at the tag’ s location. Instead, they may intersect at one or more different points or may not intersect at all. Compensating for the readers’ offset and / or rotational misalignment when projecting the AO As into the common environmental reference frame causes the projected AO As to intersect at the tag’s actual position, neglecting other error(s).Deploying and Installing RFID Readers at an Installation Site

[0065] FIG. 5 illustrates a process 500 for measuring and compensating for misalignment or error in a reader’s actual position with respect to its desired or intended position in an environmental coordinate system / frame of reference, e.g., with an origin at one corner of the installation site. Because each sensor locates RFID tags with respect to itself, knowing each sensor’s actual location and orientation in the environmental coordinate system / reference frame makes it possible to translate that sensor’s relative RFID tag location measurements accurately and precisely onto the environmental coordinate system. Following this process when installing or deploying readers can reduce or eliminate errors in tag location estimates made from AO As, RS Sis, channel estimates, or other measurements of tag replies made with the readers.

[0066] The process 500 begins with an initial survey or measurement of the installation site, which may be a retail store, warehouse, office, library, healthcare facility, or other site that stores or houses RFID tags. This survey can include obtaining a two-dimensional (2D) or three- dimensional (3D) representation or model of the installation site; for example, a surveyor can take one or more lidar scan(s) (502) of the installation site with a portable lidar scanner or a sequence of color or grayscale photographs or videos with a camera. The representation(s) (e.g., lidar scan(s), image(s), and / or video(s)) should show the ceiling, walls, floor, and fixtures of the installation site and can include both depth information and color and / or shading. Red- green-blue-depth (RGB-D) images or videos are especially useful for registering the installed sensor to tiled ceilings and other surfaces without much 3D texture. The lidar scanner or camera used for the initial site survey may include an accelerometer or inertial measurement unit (IMU) for sensing the direction of gravity while performing the initial site survey.

[0067] In some cases, though not all, the representation(s) may show the installation site in enough detail and with fine enough spatial resolution (e.g., 6 inches, 3 inches, 1 inch, or finer)to identify and locate vents, sprinkler heads, and other potential obstructions on the ceiling, walls, and floor. The lidar scans, image, videos, and / or other representations can be used to generate a map or plan of the locations at which the sensors are to be installed, optionally after the lidar scans have been aligned properly or otherwise registered or validated. For more information on generating this map or plan, please see International Application No. PCT / US2023 / 068002, entitled “Deploying RFID Readers in Environments Having a Dense Population of RFID Tags,” which is incorporated herein by reference in its entirety for all purposes.

[0068] Once the map or plan of desired locations for readers has been created, an installer can use it to install the readers (504) in the desired locations at the installation site. Ideally, the readers are installed exactly as specified in the map or plan. In practice, however, some or all of the readers may not be installed exactly as specified in the map or plan. They may be shifted, rotated, and / or tilted with respect to their desired positions for any of a variety of reasons, including but not limited to obstacles unforeseen during site survey and design, installation difficulties, inaccuracies in the installation map or site survey data (e.g., images or lidar scans), or errors in installation. Each reader may be installed in or on an adjustable mounting bracket and / or include an internal bracket that can be used to adjust the reader’s height, lateral position, pitch, yaw, and / or roll within a limited range (e.g., a few millimeters or centimeters of lateral and / or vertical travel and / or by ± 0.25° in pitch, yaw, and roll). Deviations from the desired pitch, yaw, and roll angles can degrade sensor performance, especially for sensors that are installed higher above the floor.

[0069] Next, an installer or inspector performs photometric registration of the installed sensors, e.g., as described below with respect to FIGS. 6-13. Photometric registration can be carried out with a camera that can acquire color or grayscale images and depth images or depths maps. The camera may optionally be equipped with an accelerometer or IMU for measuring the camera’s orientation when acquiring the image(s) and depth map(s). Most tablets and smartphones, including iPads and iPhones, are equipped with suitable cameras and accelerometers. Photometric registration identifies and can be used to compensate for deviations between the sensor’s actual pose and position and the sensor’s desired pose and position.

[0070] To perform photometric registration, the installer or inspector takes color or grayscale images and depth images or depths maps of the installed sensors with a smartphone or other suitable camera or sensor (e.g., Microsoft Kinect sensor) (506). If so equipped, thesmartphone’s accelerometer or IMU may also measure the acceleration due to gravity, or gravity vector, as the smartphone acquires the images and depth maps. Optionally, the gravity vector can be used to calibrate the roll and pitch of the installed sensor as explained below. Alternatively, the gravity vector can be obtained from the initial lidar scan of the installation site.

[0071] The images and depth maps acquired with the smartphone are images or image channels that contain information relating to the distance of the surfaces of scene objects from a viewpoint (e.g., the camera’s location). Each image and depth map shows some or all of an installed sensor and at least a portion of the installed sensor’s surroundings, including one or more features discernible in the initial lidar scan(s) of the installation site. The surroundings provide context for registering the images and depth maps to the representations of the installation site acquired before sensor installation. Ideally, though not necessarily, the camera is positioned to capture each image and depth map from a distance where the accuracy of depth is optimal (this results in better data) and from right underneath the installed sensor to reduce or minimize perspective distortion. For open ceilings and tiled ceilings, images and depth maps captured this way contain enough features to be registered to the pre-installation representation of the installation site.

[0072] The installer or inspector uses a computer, smartphone, tablet, or other suitable device to register the images and depth maps of the installed sensors to the initial representation(s) of the installation site (508). For example, the installer or inspector can upload the images and depth maps to the cloud, at which point a script executing on a remote processor picks up the uploaded data and performs the registration automatically. Alternatively, the installer or inspector can register the images and depth maps of the installed sensors to the pre-installation representation(s) locally, e.g., on a smartphone, tablet, or other computing device. In either case, once the installer or inspector has acquired the images and depths maps, one or more local or cloud processors can carry out the remaining steps of the process 500 automatically.

[0073] Because the sensors don’t appear in the initial representation(s) of the installation site, they are not used for registration and can be elided or omitted from the depth images during registration. (As described in greater detail below, a sensor can be removed from a depth image by matching the depth image to a computer-aided design (CAD) model of the sensor, then segmenting the sensor and the surrounding points from the depth image or by ignoring points belonging to the sensor when performing point cloud registration.) Removing the installed sensor from the depth map prevents the sensor from interfering with registration of the depthmap to the pre-installation representation and permits estimation of the installed sensor’s roll, pitch, and yaw apart from the pre-installation representation.

[0074] Instead, the features (e.g., edges, comers, or sides of walls, ceilings, ceiling panels, vents, etc.) that appear in both the depth images and the lidar scan(s) are used to register the depth images to the lidar scan(s). Registration may include translating, rotating, scaling, warping, stretching, or otherwise distorting or transforming the depth images in one or more dimensions so that they are properly aligned to the lidar scan(s). If the ceiling, vents, and other features in the depth images and lidar scans have sufficient texture — for example, if the ceiling is an open ceiling with exposed pipes, conduits, and ducts — then it may be possible to align point clouds derived from the depth images to the lidar scan point clouds. If the ceiling around an installed sensor has little to no texture (e.g., a tiled ceiling), then the shading or color variations in a color or grayscale image of the installed sensor and its surroundings can be used to register the installed sensor in translation (x and j’ axes) and yaw to color and / or shading in the pre-installation representation of the installation site. Once the depth images have been registered to the lidar scan(s), the installer or inspector can use the processor to estimate the actual angular orientations and positions of the installed sensors with respect to their desired angular orientations and positions, respectively (510, 512), from the registered depth images. For instance, each sensor may have one or more fiducial features or marks, such as a logo, character, and / or dot, on its housing (enclosure 280, FIG. 2A) indicating the sensor’s yaw with respect to its surroundings as explained in greater detail below. (The sensor housing can also be shaped asymmetrically to provide cues or information about the sensor’s roll, pitch, and yaw.) The housing’s orientation and position are known and fixed with respect to the sensor’s antenna elements (e.g., to within the negligible manufacturing tolerances), so the orientation of this fiducial mark with respect to one or more walls, ceiling panels, vents, sprinklers, etc. in the installation can be used to estimate the sensor’ s yaw angle with respect to the environmental coordinate system or reference frame.

[0075] The sensor’s roll and pitch angles can be estimated from the registered depth images as follows. First, a processor fits a plane to the sensor enclosure or housing and estimates the plane normal (i.e., the vector orthogonal to the plane of the sensor enclosure). The registration of the depth image to the pre-installation representation gives the plane normal in the environment coordinate frame. In some cases (but not every case), the processor uses the gravity vector measured by the smartphone’s accelerometer to correct any roll or pitch error introduced by the sensor not being level or mounted flush to the ceiling. The processor can also estimate thegravity vector from an IMU in the lidar scanner used to obtain the initial lidar scan of the installation site.

[0076] Alternatively, because the positions and orientations of the sensor’s antenna elements with respect to the sensor’s enclosure are fixed and known, registering the sensor’s enclosure to the sensor’s surroundings / environmental coordinate system makes it possible to determine the positions of the antenna elements in the environmental coordinate system. The coordinates of the corners of the antenna elements can be used to determine the pitch and roll angles of the antenna elements in the environmental coordinate system. Knowledge of the installed sensor’s actual position and actual orientation in the environmental coordinate system / reference frame make it possible to compensate for any misalignment or deviation between the installed sensor’s actual position and actual orientation and its desired position and desired orientation (514).

[0077] Once the position and orientation of the sensor (or of the sensor’s antenna elements) in the environmental coordinate system are known, the sensor can be used to locate RFID tags at the installation site. The sensor queries a tag within range and detects its reply (516). The sensor or an appliance coupled to the sensor determines the AOA, RSSI, channel estimate, and / or other information for the reply and uses that information along with the sensor’s position and orientation in the environmental coordinate system to estimate the tag’s location in the environmental coordinate system (518). More specifically, a sensor can locate a tag in its own coordinate system (i.e., a coordinate system centered on the sensor), then transform the coordinates in its own coordinate system to the environmental coordination system using a transformation matrix derived during the process of registering the sensor (and its coordinate system) to the environmental coordinate system.Sensor Pose Estimation via Photometric Registration

[0078] FIGS. 6A-13 illustrate photometric registration (e.g., steps 506, 508, 510, and 512 of the process shown in FIG. 5) in greater detail. As explained above, photometric registration involves acquiring an image and a depth map of the installed sensor and its immediate surroundings with a smartphone or tablet. A processor generates a point cloud of the installed sensor and its immediate surroundings from the depth map, then registers the point cloud (minus the sensor points) to a lidar point cloud or other 3D representation of the installation site. More specifically, photometric registration can involve estimating the pose or position and orientation of an installed sensor using:(1) a 3D representation of the installation site, e.g., a lidar point cloud like those shown in FIGS. 7 A and 7B;(2) a color or grayscale image of the installed sensor, e.g., FIG. 8A, which is an image of an installed reader next to an HVAC duct in the installation site of FIGS. 7 A and 7B;(3) a depth map of the installed sensor, e.g., FIG. 8B, which shows a depth image of the installed reader of FIG. 8 A;(4) a 3D model of the installed sensor (e.g., a point cloud of the sensor sampled uniformly from a CAD model of the sensor; during processing, the 3D model is centered at the origin of the relevant coordinate system and oriented such that the normal of the sensor’s main surface is parallel to the -z axis);(5) a binary template of the sensor (e.g., where the surface is white and the fiducial markers are black) for yaw estimation;(6) information about the desired or expected pose of the installed sensor (e.g., a rough estimate of the installed sensor’s position);(7) the focal length of the camera (e.g., in a smartphone) used to acquire the image and depth map; and(8) optionally, the gravity vector experienced by the camera (e.g., measured by the smartphone’s accelerometer) when acquiring the image and depth map (the gravity vector can be used to correct any perspective distortion in the RGB image, much like a homography warping, to improve matching of the fiducial template during yaw estimation).This information can be used to estimate the installed sensor’s pose in six dimensions — up / down, left / right, forward / backward, roll, pitch, and yaw. And the estimated pose can be used for registering the installed sensor to the coordinate system in which the sensor locates RFID tags.

[0079] FIG. 6A illustrates a first process 600 for estimating the pose of an installed sensor. This process can be carried out automatically by one or more local or cloud-based processors. (For simplicity, the process 600 is explained here as though performed by a single processor, though different steps and / or different portions of each step can be performed by different processors in a distributed fashion.)

[0080] In this process 600, a processor finds and segments the sensor in the color (red-green- blue or RGB) or grayscale image (602). The processor can segment the sensor by matching the image against a CAD model of the sensor, for instance, by using OpenCV SimpleBlobDetectorto find a point on the sensor, then using a watershed algorithm to grow the sensor region to the bounds of the sensor in the image.

[0081] In (604), the processor determines the yaw orientation of the installed sensor. The yaw orientation can be estimated by adaptively thresholding the region of the sensor in the RGB image and then sliding a sensor template of variable scale and rotation over this region. The estimated yaw rotation is that of the template with the highest or maximal cross-correlation. Because yaw estimation is based on the RGB image, not the depth image, the processor can determine the sensor’s yaw separately from (e.g., before, after, or while) eliding the installed sensor from the depth image (described below).

[0082] The processor can elide or remove the installed sensor from the depth image as follows. It starts by applying the segmentation from the RGB image to the depth map as a binary mask and segments the pixels lying in the masked region of depth map (i.e., the region of the depth map that contains the sensor) (606). The processor converts the segmented pixels in the depth map into a point cloud representing the installed sensor (“installed sensor point cloud”) and projects depth values into the installed sensor point cloud derived from the focal length of the camera (608). The processor determines the centroid of the sensor points in the installed sensor point cloud (610) and translates the installed sensor point cloud so the centroid is at the origin (612). The processor registers the installed sensor point cloud to a point cloud of the sensor derived from the 3D model (e.g., CAD model) of the sensor using the Iterative Closest Point (ICP) process or another suitable process for reducing or minimizing the difference(s) between a pair of point clouds (614). For instance, FIG. 9 illustrates a computer-aided design (CAD) model of the sensor registered to a point cloud derived from the depth image of FIG. 8B. The processor rotates the installed sensor point cloud so that the sensor’s yaw orientation is 0 degrees (616), then removes the points corresponding to the sensor from the installed sensor point cloud to yield a sensor-less point cloud (618). FIG. 10 illustrates a sensor-less point cloud derived from FIG. 9.

[0083] By eliding the sensor, the processor can register the depth map to the pre-installation 3D representation of the installation site, then determine the installed sensor’s position and orientation based on the registration. The processor does this by translating the sensor-less point cloud to the expected or desired sensor location in the pre-installation 3D representation (620). FIG. 11 illustrates the sensor-less point cloud of FIG. 10 registered to a section of the pre-installation point cloud of FIGS. 7 A and 7B. And FIG. 12 illustrates a collection of sensor-less, post-installation point clouds derived from depth maps of different installed sensors to the pre-installation point cloud of FIGS. 7 A and 7B.

[0084] The processor finds the pose of the sensor-less point cloud that provides the highest or maximal correspondence to the corresponding portion of the pre-installation 3D representation by performing sliding window matching along the x, y, and z axes and rotation about the z axis (622). When the sensor-less point cloud has significant depth texture, the processor uses the ICP process or another suitable process to register the sensor-less point cloud in three dimensions to the pre-installation 3D representation at the pose that provides the maximal correspondence (624). If the sensor-less point cloud lacks sufficient depth texture for ICP, but the RGB images of the installed sensor and the installation site have sufficient color texture, then the processor can perform a correlation analysis based on color. (This correlation analysis can be seen as a sliding window matching based on color.)

[0085] Then the processor extracts the translation and orientation of the installed sensor based on the registration transformation (626).

[0086] FIG. 6B illustrates an alternative process 650 for estimating the pose of an installed sensor. Unlike the process 600 in FIG. 6A, the process 650 in FIG. 6B does not use OpenCV SimpleBlobDetector to find a point on the sensor or a watershed algorithm to grow the sensor region to the bounds of the sensor in the image and thus avoids potential problems associated with failing to detect a point on the sensor or the watershed region extending beyond the edges of the sensor. Instead, the process 650 involves estimating the sensor midpoint via template matching, aligning the sensor model to the sensor midpoint in the depth map, and extracting the more precise yaw of the installed sensor from a perspective-corrected synthetic image of the installed sensor. The process 650 does not use any rotational degrees of freedom and incorporates information about the sensor’s appearance when finding the pose of the installed sensor.

[0087] To start, the processor overlays a model of the sensor onto the depth image of the installed sensor as follows. The processor finds the sensor midpoint in the RGB image of the installed sensor (652), for example, by matching a template of the fiducial marks on the sensor housing to the RGB image. The processor converts the depth map of the installed sensor into a point cloud (654) and projects depth values into the point cloud using the (known) focal length of the camera that was used to acquire the depth map. The processor backprojects the sensor midpoint from the 2D RGB image into three dimensions using depth information fromthe depth map (656). Next, the processor registers a model of the sensor (e.g., a point cloud of the sensor derived from a CAD model of the sensor) to the depth image by translating the center of the model to the 3D sensor midpoint in the depth map and performing ICP processing or another suitable process (658). The processor estimates the normal of the sensor housing via random sample consensus (RANSAC)-based plane estimation or another suitable technique (660) and reorients the model of the sensor (e.g., the sensor point cloud) so that the normal of the sensor housing is parallel to the gravity vector of the camera or device that acquired the depth image (662).

[0088] Once the model of the sensor (sensor point cloud) has been registered to the depth image, the processor can remove the sensor from the depth image (664). More specifically, the processor removes the points in a point cloud derived from the depth image corresponding to the registered sensor point cloud and as well as points within a predetermined distance (e.g., 2 cm) of the registered sensor point cloud (e.g., using a KDTree-based radius search).

[0089] Next, the processor uses the extracted points to find the yaw of the installed sensor. The processor does this by generating a mesh of these extracted points (e.g., via Poisson surface reconstruction) (666) and uses the mesh to render a synthetic image of the sensor housing (668). The processor estimates the yaw orientation of the installed sensor by matching the synthetic image to a template of the fiducial marks on the sensor housing (670). For instance, FIG. 13 shows a synthetic image of the installed sensor being rotated in yaw to match a template of the housing — a dot on one side of the housing and a stylized R on the other side of housing.

[0090] The processor translates the sensor-less point cloud to the expected sensor location in pre-installation 3D representation (e.g., point cloud) of the installation site (672). The processor performs sliding window matching (in x, y, and z coordinates), finding pose with maximal correspondence between the sensor-less point cloud and the pre-installation 3D representation (674). The processor registers the sensor-less point cloud to the pre-installation 3D representation using the ICP process or another suitable process (676) and extracts or determines the installed sensor’s translation and orientation based on the registration transform.Continuous Signature / Channel Estimate Collection

[0091] The installed sensors can locate tags using AOA measurements, RSSI measurements, or other techniques, including channel estimates or (multi-path) signatures. As explained briefly above, a channel estimate or signature represents the scattering, fading, interference, and other effects that affect signals propagating along a communications channel or linkbetween a particular sensor and a particular tag (or more precisely, a tag in a particular orientation at a particular location). These effects are functions of tag position, tag orientation, sensor position, carrier frequency (frequency channel), and sensor antenna configuration (e.g., the sensor beamforming sector, or simply sector) and affect each communications channel differently. Each channel estimate represents these effects, parameterized by sensor, carrier frequency, sensor beamforming sector, interrogation signal amplitude, and / or other degrees of freedom associated with the interrogation signal, for a given communications channel.

[0092] If each channel estimate is unique and the communications channels remain relatively static and independent of the individual tags’ characteristics, then a channel estimate can be used as a fingerprint that identifies the corresponding communications channel. And if the locations of the endpoints of each communications channel — i.e., the locations of the sensor and tag — are known, then the channel estimate can be mapped to those locations. In other words, for each sensor, a unique channel estimate for each set of parameters can be assigned to each possible tag location within range of that sensor. The sensors and / or interrogator controller can store these parameterized channel estimates for locating tags quickly: if a tag’s channel estimate has not changed since the tag was last interrogated, then the tag’s location (likely) has not changed since the tag was last interrogated. Similarly, if a new or recently moved tag’s channel estimate matches a previously stored channel estimate, then that new or recently moved tag is (likely to be) at the location associated with the previously stored channel estimate. For more on locating RFID tags using channel estimates, please see International Application No. PCT / US2024 / 020357, filed March 18, 2024, and entitled “Channel Estimation for Locating RFID Tags,” which is incorporated herein by reference in its entirety for all purposes.

[0093] FIGS. 14 and 15 illustrate two methods of collecting channel estimates and associating the channel estimates with locations during deployment and / or operation of the installed sensors at an installation site 1403, 1503. As shown in FIG. 14, one or more panels 1402 with attached or embedded RFID tags 101 are distributed around the installation site. Each panel 1402 is also printed with a unique quick response (QR) code 1404 and / or bar code. An installer 1401 takes a picture of each panel 1402 and its surroundings with a smart phone 1410, tablet, or other device, which either registers the picture to the pre-installation representation of the installation site or transmits the picture to a server or other device (e.g., the interrogator controller 140) that registers the picture to the pre-installation representation. Either way, this registration effectively locates the corresponding panel 1402 in the coordinate frame of theinstallation site 1403. The smart phone 1410 transmits the panel’s location to the sensors 150 and / or interrogator controller 140, which store the panel’s location in local memory.

[0094] Before, during, or after taking the picture of the panel 1402 and its surroundings, the installer 1401 reads the QR code 1404 on the panel 1402 with the smart phone 1410. In response to reading the QR code 1404, the smart phone 1410 triggers interrogation of the panel’s RFID tag 101 by one or more of the installed sensors 150. At least one of the sensors 150 transmits a query or command to the RFID tag 101, and the sensors 150 within range detect the tag’s reply to the query or command. Each sensor 150 derives a channel estimate for the communications channel between itself and the tag 101 and either associates the channel estimate with the panel’s location in local memory or transmits the channel estimate to the interrogator controller 140, which associates the channel estimate with the panel’s location in local memory. The installer 1401 can repeat this process with a single panel 1402 at many different locations, with many panels 1402 at respective locations, or with many panels 1402 each moved to many different locations.

[0095] FIG. 15 shows how to collect channel estimates with an RFID tag 101 mounted on a drone 1502 (here, a quadcopter) or remote-control vehicle that includes a camera 1504 and an inertial measurement unit (IMU) 1506, accelerometer, or other device that measures the drone’s motion. The drone 1502 travels around the installation site 1503, it uses the camera 1504 to acquire images of its surroundings and the IMU 1506 to measure its specific gravity and angular rate. The drone 1502 uses these images and IMU measurements to determine its position at different times (e.g., every so many seconds) as it flies through the installation site 1503. At the same time, the sensors 150 installed in the installation site 1503 interrogate the RFID tag 101 attached to the drone 1502, forming channel estimates for each reply from the RFID tag 101 and recording the time at which they receive each reply. The drone 1502 reports its position as a function of time to a cloud-based processor or to the interrogator controller 140, which matches the drone’s position to the corresponding channel estimates from the sensors 150 using the time stamps for the positions and channel estimates. The interrogator controller 140 stores the channel estimates and associated positions in local memory and / or distributes them to the sensors 150 for use in locating RFID tags during operation.Conclusion

[0096] While various inventive embodiments have been described and illustrated herein, those of ordinary skill in the art will readily envision a variety of other means and / or structures forperforming the function and / or obtaining the results and / or one or more of the advantages described herein, and each of such variations and / or modifications is deemed to be within the scope of the inventive embodiments described herein. More generally, those skilled in the art will readily appreciate that all parameters, dimensions, materials, and configurations described herein are meant to be exemplary and that the actual parameters, dimensions, materials, and / or configurations will depend upon the specific application or applications for which the inventive teachings is / are used. Those skilled in the art will recognize or be able to ascertain, using no more than routine experimentation, many equivalents to the specific inventive embodiments described herein. It is, therefore, to be understood that the foregoing embodiments are presented by way of example only and that, within the scope of the appended claims and equivalents thereto, inventive embodiments may be practiced otherwise than as specifically described and claimed. Inventive embodiments of the present disclosure are directed to each individual feature, system, article, material, kit, and / or method described herein. In addition, any combination of two or more such features, systems, articles, materials, kits, and / or methods, if such features, systems, articles, materials, kits, and / or methods are not mutually inconsistent, is included within the inventive scope of the present disclosure.

[0097] Also, various inventive concepts may be embodied as one or more methods, of which an example has been provided. The acts performed as part of the method may be ordered in any suitable way. Accordingly, embodiments may be constructed in which acts are performed in an order different than illustrated, which may include performing some acts simultaneously, even though shown as sequential acts in illustrative embodiments.

[0098] All definitions, as defined and used herein, should be understood to control over dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0099] The indefinite articles “a” and “an,” as used herein in the specification and in the claims, unless clearly indicated to the contrary, should be understood to mean “at least one.”

[0100] The phrase “and / or,” as used herein in the specification and in the claims, should be understood to mean “either or both” of the components so conjoined, i.e., components that are conjunctively present in some cases and disjunctively present in other cases. Multiple components listed with “and / or” should be construed in the same fashion, i.e., “one or more” of the components so conjoined. Other components may optionally be present other than the components specifically identified by the “and / or” clause, whether related or unrelated to thosecomponents specifically identified. Thus, as a non-limiting example, a reference to “A and / or B”, when used in conjunction with open-ended language such as “comprising” can refer, in one embodiment, to A only (optionally including components other than B); in another embodiment, to B only (optionally including components other than A); in yet another embodiment, to both A and B (optionally including other components); etc.

[0101] As used herein in the specification and in the claims, “or” should be understood to have the same meaning as “and / or” as defined above. For example, when separating items in a list, “or” or “and / or” shall be interpreted as being inclusive, i.e., the inclusion of at least one, but also including more than one, of a number or list of components, and, optionally, additional unlisted items. Only terms clearly indicated to the contrary, such as “only one of’ or “exactly one of,” or, when used in the claims, “consisting of,” will refer to the inclusion of exactly one component of a number or list of components. In general, the term “or” as used herein shall only be interpreted as indicating exclusive alternatives (i.e., “one or the other but not both”) when preceded by terms of exclusivity, such as “either,” “one of,” “only one of,” or “exactly one of.” “Consisting essentially of,” when used in the claims, shall have its ordinary meaning as used in the field of patent law.

[0102] As used herein in the specification and in the claims, the phrase “at least one,” in reference to a list of one or more components, should be understood to mean at least one component selected from any one or more of the components in the list of components, but not necessarily including at least one of each and every component specifically listed within the list of components and not excluding any combinations of components in the list of components. This definition also allows that components may optionally be present other than the components specifically identified within the list of components to which the phrase “at least one” refers, whether related or unrelated to those components specifically identified. Thus, as a non-limiting example, “at least one of A and B” (or, equivalently, “at least one of A or B,” or, equivalently “at least one of A and / or B”) can refer, in one embodiment, to at least one, optionally including more than one, A, with no B present (and optionally including components other than B); in another embodiment, to at least one, optionally including more than one, B, with no A present (and optionally including components other than A); in yet another embodiment, to at least one, optionally including more than one, A, and at least one, optionally including more than one, B (and optionally including other components); etc.

[0103] In the claims, as well as in the specification above, all transitional phrases such as “comprising,” “including,” “carrying,” “having,” “containing,” “involving,” “holding,”“composed of,” and the like are to be understood to be open-ended, i.e., to mean including but not limited to. Only the transitional phrases “consisting of’ and “consisting essentially of’ shall be closed or semi-closed transitional phrases, respectively, as set forth in the United States Patent Office Manual of Patent Examining Procedures, Section 2111.03.

Claims

CLAIMS1. A method of deploying a radio-frequency identification (RFID) tag reader in an installation site, the method comprising: obtaining a first representation of the installation site before installation of the RFID tag reader; installing the RFID tag reader at the installation site; obtaining an image and / or depth map of the RFID tag reader and at least a portion of the installation site surrounding the RFID tag reader after installation of the RFID tag reader at the installation site; deriving a second representation of the installation site from the image and / or depth map; registering the second representation to the first representation; and determining a position and / or an orientation of the RFID tag reader with respect to a coordinate system of the installation site based on registration of the second representation to the first representation.

2. The method of claim 1, wherein obtaining the first representation of the installation site comprises performing a lidar scan of the installation site.

3. The method of claim 1, wherein the first representation is a first three-dimensional (3D) representation and the second representation is a second 3D representation.

4. The method of claim 1, wherein installing the RFID tag reader comprises adjusting a pitch, a yaw, and / or a roll of the RFID tag reader to within 0.25° of a desired angle.

5. The method of claim 1, wherein registering the second representation to the first representation comprises at least one of warping, translating, rotating, or scaling the second representation.

6. The method of claim 1, further comprising, before registering the second representation to the first representation: eliding the RFID tag reader from the second representation.

7. The method of claim 6, wherein eliding the RFID tag reader from the second representation comprises:obtaining an image of the RFID tag reader and at least a portion of the installation site surrounding the RFID tag reader after installation of the RFID tag reader at the installation site; segmenting the RFID tag reader in the image; and removing points corresponding to segments of the RFID tag reader from the second representation.

8. The method of claim 7, wherein segmenting the RFID tag reader comprises matching the image to a computer-aided design (CAD) model of the RFID tag reader.

9. The method of claim 6, wherein eliding the RFID tag reader from the second representation comprises: registering a model of the RFID tag reader to the image and / or depth map; and removing points corresponding to the model of the RFID tag reader from the second representation.

10. The method of claim 9, further comprising: removing points within a predetermined distance of the model of the RFID tag reader from the second representation.

11. The method of claim 1, wherein determining the position and / or the orientation of the RFID tag reader with respect to a coordinate system of the installation site comprises translating the second representation with respect to the first representation based on an image of the RFID tag reader and at least a portion of the installation site surrounding the RFID tag reader after installation.

12. The method of claim 1, further comprising: determining a yaw orientation of the RFID tag reader based on at least one fiducial mark on the RFID tag reader.

13. The method of claim 12, wherein determining the yaw orientation of the RFID tag reader comprises matching a template of the at least one fiducial mark to an image of the RFID tag reader and at least a portion of the installation site surrounding the RFID tag reader after installation.

14. The method of claim 1, further comprising:transmitting, from the RFID tag reader, a signal to an RFID tag; receiving, by the RFID tag reader, a reply to the signal from the RFID tag; and estimating a location of the RFID tag with respect to the installation site based on the reply and the position and / or the orientation of the RFID tag reader with respect to the coordinate system of the installation site.

15. The method of claim 1, further comprising: acquiring an image of a portion of the installation site surrounding an RFID tag; determining a position of the RFID tag in the coordinate system of the installation site based on the image of the portion of the installation site surrounding the RFID tag; receiving, by the RFID tag reader, a reply from the RFID tag; determining, based on the reply, a channel estimate representing a communications channel between the RFID tag reader and the RFID tag; and associating the channel estimate with the position of the RFID tag in the coordinate system of the installation site.

16. The method of claim 15, wherein acquiring the image of the portion of the installation site surrounding the RFID tag comprises photographing the RFID tag and the portion of the installation site surrounding the RFID tag.

17. The method of claim 15, wherein the RFID tag is mounted to a drone, acquiring the image of the portion of the installation site comprises acquiring the image with a camera mounted to the drone, and determining the position of the RFID tag comprises using visual inertial odometry to determine the position of the drone.

18. The method of claim 15, wherein determining the position of the RFID tag in the coordinate system of the installation site comprises registering the image of the portion of the installation site surrounding the RFID tag to the first representation of the installation site.

19. The method of claim 15, further comprising: acquiring an image of a quick response (QR) code or bar code associated with the RFID tag, wherein acquiring the image of the QR code or bar code triggers interrogation of the RFID tag by the RFID tag reader.

20. A system for interrogating RFID tags comprising RFID tag readers installed according to the method of claim 1.

21. A system for locating a radio-frequency identification (RFID) tag in an environment, the system comprising: an RFID tag reader, fixedly installed in the environment, to receive a reply from the RFID tag; and an interrogator controller, operably coupled to the RFID tag reader, to estimate an angle of arrival at the RFID tag reader of the reply in a coordinate system centered on the RFID tag reader, to translate the angle of arrival from the coordinate system centered on the RFID tag reader to a coordinate system of the environment based on registration of a position and / or an orientation of the RFID tag reader with respect to the coordinate system of the environment, and to estimate a location of the RFID tag in the coordinate system of the environment based on the angle of arrival in the coordinate system of the environment.

22. The system of claim 21, wherein the RFID tag reader is shifted and / or rotated with respect to a desired position and / or orientation of the RFID tag reader.

23. A method of locating a radio-frequency identification (RFID) tag in an environment, the method comprising: receiving, by an RFID tag reader installed at a fixed position in the environment, a reply from the RFID tag; determining an angle of arrival at the RFID tag reader of the reply in a coordinate system centered on the RFID tag reader; translating the angle of arrival from the coordinate system centered on the RFID tag reader to a coordinate system of the environment based on registration of a position and / or an orientation of the RFID tag reader with respect to the coordinate system of the environment; and estimating a location of the RFID tag in the coordinate system of the environment based on the angle of arrival in the coordinate system of the environment.

24. The method of claim 23, wherein the fixed position of the RFID tag reader is different than a desired position of the RFID tag reader.

25. A method of deploying a radio-frequency identification (RFID) tag reader, the method comprising:acquiring an image of a portion of an environment surrounding an RFID tag; determining a position of an RFID tag in the environment based on the image; receiving, by the RFID tag reader, a reply from an RFID tag; determining, based on the reply, a channel estimate representing a communications channel between the RFID tag reader and the RFID tag; and associating the channel estimate with the position of the RFID tag.

26. The method of claim 25, wherein acquiring the image of the environment comprises photographing the RFID tag and the portion of the environment surrounding the RFID tag.

27. The method of claim 25, wherein the RFID tag is mounted to a drone, acquiring the image of the environment comprises acquiring the image with a camera mounted to the drone, and determining the position of the RFID tag comprises using visual inertial odometry to determine a position of the drone.

28. The method of claim 25, further comprising: acquiring an image of a quick response (QR) code or bar code associated with the RFID tag, wherein acquiring the image of the QR code or bar code triggers interrogation of the RFID tag by the RFID tag reader.

Images