Smart shelf brackets

Abstract

A smart-bracket system (100) includes shelf brackets (110), weight sensors (202), and one or more signal processors (122, 230). The shelf brackets are configured to support a chain of two or more shelves (104, 106). The weight sensors are coupled to at least one of the shelf brackets, each weight sensor being configured to sense a load on a respective shelf bracket and to output a signal indicative of the sensed load. A given weight sensor among the weight sensors is configured to output a signal indicative of the load on first and second adjacent shelves. The one or more signal processors are configured to identify, based on one or more signals output by one or more of the weight sensors, including the given weight sensor, an interaction in which one or more items are picked from or placed on one or more of the shelves.

Description

[0001] SMART SHELF BRACKETS

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] This application claims the benefit of U.S. Provisional Patent Application 63 / 672,895, filed July 18, 2024, whose disclosure is incorporated herein by reference.

[0004] FIELD OF THE DISCLOSURE

[0005] The present disclosure relates generally to inventory management and more particularly to weight sensor arrangements useful for inventory management.

[0006] BACKGROUND OF THE DISCLOSURE

[0007] Various types of weight sensor arrangements useful for inventory management are known in the art.

[0008] SUMMARY OF THE DISCLOSURE

[0009] An embodiment that is described herein provides a smart-bracket system including shelf brackets, weight sensors, and one or more signal processors. The shelf brackets are configured to support a chain of two or more shelves. The weight sensors are coupled to at least one of the shelf brackets, each weight sensor being configured to sense a load on a respective shelf bracket and to output a signal indicative of the sensed load. A given weight sensor among the weight sensors is configured to output a signal indicative of the load on first and second adjacent shelves. The one or more signal processors are configured to identify, based on one or more signals output by one or more of the weight sensors, including the given weight sensor, an interaction in which one or more items are picked from or placed on one or more of the shelves.

[0010] In some embodiments, a given shelf bracket is a common bracket configured to support both the first shelf and the second shelf, and the given weight sensor is coupled to the common bracket so as to sense the load on both the first shelf and the second shelf. In other embodiments, the given weight sensor is coupled to a shelf bracket that supports the first shelf, and the first and second shelves are mechanically coupled to one another, thereby causing the given weight sensor to sense the load on both the first shelf and the second shelf.

[0011] In some embodiments, the one or more signal processors are configured to estimate a location of the interaction along the chain of shelves based on the signals. In an example embodiment, the one or more signal processors are configured to identify a subset of the shelf brackets whose weight sensors are active due to the interaction, and to estimate the location of the interaction based on the signals output by the weight sensors of the shelf brackets in the subset. In an embodiment, the one or more signal processors are configured to activate and add to the subset at least one shelf bracket that is an immediate neighbor of a shelf bracket in the subset. In some embodiments, the one or more signal processors are configured to estimate the location of the interaction by comparing the signals to a response model that specifies a response of a weight sensor as a function of distance from the location of the interaction. In an embodiment, based on the response model, the one or more signal processors are configured to distinguish between a single interaction and multiple simultaneous interactions. In an example embodiment, upon identifying multiple simultaneous interactions, the one or more signal processors are configured to estimate, based on the response model, respective locations of the multiple simultaneous interactions. In an embodiment, by comparing the signals to the response model, the one or more signal processors are configured to detect a fault and to initiate a responsive action.

[0012] There is additionally provided, in accordance with an embodiment that is described herein, a method including, in a system that includes shelf brackets that support a chain of two or more shelves, operating a plurality of weight sensors coupled to at least one of the shelf brackets, so that each weight sensor senses a load on a respective shelf bracket and outputs a signal indicative of the sensed load. A given weight sensor among the weight sensors outputs a signal indicative of the load on first and second adjacent shelves. An interaction, in which one or more items are picked from or placed on one or more of the shelves, is identified based on one or more signals output by one or more of the weight sensors, including the given weight sensor.

[0013] The present invention will be more fully understood from the following detailed description of the embodiments thereof, taken together with the drawings in which:

[0014] BRIEF DESCRIPTION OF DRAWINGS

[0015] The present disclosure will be understood and appreciated from the following detailed description, taken in conjunction with the drawings in which:

[0016] Fig. 1A is a simplified perspective view of a weight sensing system, constructed and operative in accordance with an embodiment of the present invention;

[0017] Fig. IB is a simplified front view of the system of Fig. 1A, also including external communication links;

[0018] Figs. 2A - 2F are simplified respective exploded, assembled, rear, top, side and underside views of a smart bracket suitable for use in a system of the type shown in Figs. 1A and IB, constructed and operative in accordance with an embodiment of the present invention;

[0019] Figs. 3 A - 3E are simplified respective exploded, assembled, side, rear and top views of a smart bracket suitable for use in a system of the type shown in Figs. 1 A and IB, constructed and operative in accordance with another embodiment of the present invention; Figs. 4A - 4E are simplified respective exploded, assembled, side, top and rear views of a smart bracket suitable for use in a system of the type shown in Figs. 1 A and IB, constructed and operative in accordance with a further embodiment of the present invention;

[0020] Figs. 5 and 6 are highly simplified representations of various interactions occurring on shelves forming part of a system constructed and operative in accordance with an embodiment of the present invention;

[0021] Fig. 7 is a simplified flow chart illustrating a method in accordance with an embodiment of the present invention;

[0022] Fig. 8 is a simplified flow chart illustrating a method in accordance with a further embodiment of the present invention;

[0023] Fig. 9 is a simplified illustration of another embodiment of the present invention; and

[0024] Figs. 10A and 10B are simplified respective assembled and rear views of a smart bracket suitable for use in a system of the type shown in Figs. 1 A and IB, constructed and operative in accordance with an embodiment of the present invention.

[0025] DETAILED DESCRIPTION OF EMBODIMENTS

[0026] SYSTEM DESCRIPTION

[0027] Reference is now made to Fig. 1A, which is a simplified perspective view of a weight sensing system, constructed and operative in accordance with an embodiment of the present invention; and to Fig. IB, which is a simplified front view of the system of Fig. 1A, also including external communication links.

[0028] In the weight sensing system of the present disclosure, at least one weight sensing assembly is mounted on at least one bracket, in a shelving system, between the bracket and a shelf supported thereby. In the following description, a shelf bracket having a weight sensing assembly mounted thereon, between the shelf bracket and a shelf supported thereby, may be termed a ‘smart bracket’ . The smart brackets of the present disclosure may be operative to sense variations in weight of the shelves supported thereby, which variations may arise from interactions with items on the shelves. A signal processor (implemented in hardware and / or software), in some embodiments included in the smart brackets, may be operative to ascertain a weight change, position, timing or nature of an interaction, or a combination thereof.

[0029] In accordance with an embodiment of the present invention, an array of shelves may include a plurality of weight sensor assemblies mounted on a plurality of brackets. In some embodiments, although not necessarily, at least some of the weight sensing assemblies of the plurality of weight sensor assemblies are mounted on two adjacent brackets, straddled across both brackets. Each weight sensor assembly, when so mounted, is operative to sense variations in weight at least of the two shelves respectively supported by the two brackets on which the weight sensor assembly is mounted. This arrangement has been found to be particularly advantageous in terms of signal analysis and ease of implementation, as is detailed henceforth.

[0030] In alternative embodiments, a weight sensor may sense the load on two adjacent shelves without necessarily using a common shelf bracket. Consider, for example, two adjacent shelves that are mechanically coupled to one another (e.g., via friction between their ends, or via some mechanical connection or coupling mechanism between the ends of the shelves). In such a case, a weight sensor that is coupled to a bracket supporting one of the shelves will sense the load on the other shelf, as well.

[0031] The weight sensing assembly of the present disclosure is configured to be fit between a conventional shelf bracket and shelf, without requiring special adaptation of the conventional shelf or bracket thereto. The smart bracket of the present disclosure is thus highly versatile and readily retrofittable to existing shelving installations.

[0032] While it is appreciated that the system and method of the present disclosure may be employed in any suitable shelving system, the system and method of the present disclosure are particularly suitable for use in automated retail stores. Accordingly, much of the description which follows relates to the use of the smart brackets of the present disclosure in the context of automated retail stores.

[0033] Referring now to Figs. 1 A and IB, it is seen that a system 100 of the present disclosure may be implemented in a facility, such as a retail store, including a plurality of shelves 102. Here, by way of example, shelves 102 are shown to include a first shelf 104 and a second shelf 106, although it is appreciated that shelves 102 may include only one shelf or more than two shelves, which may be arranged in any suitable arrangement. For example, in a typical retail store, shelves 102 may be mounted on a vertical support element such as a gondola 108, as shown in Fig. 9.

[0034] Each of shelves 102 is supported, for example at each end thereof, by a shelf bracket 110. Shelf bracket 110 may be a conventional shelf bracket, but for the mounting of a weight sensor assembly 112 thereon. As seen most clearly in Fig. IB, one weight sensor assembly 112 is mounted on each at least one shelf bracket, in some embodiments on two adjacent shelf brackets. Weight sensor assembly 112 is positioned between the shelf bracket(s) 110 and the corresponding shelf 102 supported thereby. Shelf bracket 110 in combination with weight sensor assembly 112 mounted thereon may be termed a ‘smart bracket’ 114.

[0035] Weight sensor assembly 112 is operative to sense a load on the shelf or shelves 102 supported by the shelf bracket(s) 110 upon which the weight sensor assembly is mounted. As shoppers remove or replace items on shelves 102, the weight of shelves 102 as measured by weight sensor assemblies 112, varies accordingly. Weight sensor assembly 112 is operative to sense the changes in weight in real-time and to ascertain, based on the sensed data, whether one or more items have been picked from or placed on shelf or shelves 102. Weight sensor assembly 112, and particularly a signal processor included therein, may be operative to ascertain a change in weight due to an interaction, a two-dimensional position of an interaction, a time of an interaction, a nature of an interaction such as pick or place, or a combination thereof.

[0036] In some embodiments of the present invention, smart bracket 114 may include a ‘self- contained’ signal processor and may be capable of performing edge processing of the signals sensed thereby. Smart bracket 114 may provide an output 120 indicative of the change in weight, nature, time or location of an interaction, or a combination thereof to an external server 122, as shown in Fig. IB. External server 122 may be a component of an inventory management system, of which system 100 forms a part. Smart bracket 114 may be in wireless communication with external server 122. In some embodiments, further processing of output 120 may be carried out at external server 122.

[0037] In accordance with an embodiment of the present invention, a weight sensor assembly 112 may be mounted on (across) two shelf brackets 110, such that the weight sensor assembly 112 straddles both brackets 110. This arrangement is particularly efficacious in the case of two adjacent brackets 110 of adjacent shelves 102. Rather than affixing one weight sensor assembly 112 per each bracket 110, it has been found to be particularly advantageous to affix one weight sensor assembly 112 across two adjacent brackets 110.

[0038] This arrangement is illustrated in Fig. IB, wherein weight sensor assembly 112B is shown to be mounted across both of adjacent brackets HOB and 110C, which brackets HOB and 110C respectively support adjacent shelves 104 and 106. Shelf 104 is thus supported at a first end thereof by bracket 110A having weight sensor assembly 112A mounted thereon and at a second end thereof by bracket 110B having weight sensor assembly 112B mounted thereon, which weight sensor assembly 112B is also mounted on bracket 110C. Similarly, shelf 106 is supported at a first end thereof by bracket HOC having weight sensor assembly 112B mounted thereon, which weight sensor assembly 112B is also mounted on bracket 110B, and at a second end thereof by bracket 110D having weight sensor assembly 112C mounted thereon. It is appreciated that row of shelves 102 may include more than two adjacent shelves, each shelf being supported by a smart bracket at either end thereof, which smart bracket may be shared with another adjacent shelf. As may be appreciated from consideration of Fig. IB, such an arrangement results in shelves 104 and 106 being interconnected, since weight sensor assembly 112B straddles two adjacent brackets HOB and HOC, creating a junction between adjacent shelves 104 and 106. Novel signal processing methods are employed, in accordance with embodiments of the present invention, in order to handle signals arising from interactions on such interconnected shelves 102, as is further detailed henceforth in reference to Figs. 5 and 6.

[0039] The mounting of weight sensor assembly 112 on a pair of two adjacent brackets 110, rather than on individual brackets, serves to reduce installation costs, since fewer weight sensor assemblies 112 per total number of brackets 110 are required. Furthermore, in cases of tight space constraints, the mounting of one weight sensor assembly 112 across two brackets 110 may be easier to achieve than the mounting of two weight sensor assemblies 112.

[0040] Additionally, should weight sensor assemblies 112 be installed individually per each one of brackets 110, there may be cross-talk between weight sensor assemblies 112, despite the weight sensor assemblies 112 for different ones of shelves 102 being ostensibly separated. Rather than attempting to identify and mitigate such cross-talk, in the present disclosure weight sensor assemblies 112 are deliberately shared between adjacent ones of shelves 102. As a result, shelves 102 become interconnected by way of the weight sensor assemblies 112 linking therebetween. Appropriate signal processing methods, for associating measured signals with those shelves upon which interactions giving rise to those signals have occurred, are employed, as is detailed henceforth.

[0041] It is appreciated, however, that in some embodiments of the present invention, a weight sensor assembly 112 may be mounted on only one bracket 110. This may be the case, for example, for brackets 110 at the end of a row of shelves 102. Such an arrangement is illustrated in Fig. IB, for brackets 110A and 110D at the end of row of shelves 102. This may additionally or alternatively be the case for other possible implementations, in which it may be desirable to mount weight sensor assemblies 112 individually on corresponding ones of brackets 110.

[0042] EXAMPLE MECHANICAL CONFIGURATIONS

[0043] Reference is now made to Figs. 2A - 2F, which are simplified respective exploded, assembled, rear, top, side and underside views of a smart bracket suitable for use in a system of the type shown in Figs. 1 A and IB, constructed and operative in accordance with an embodiment of the present invention.

[0044] As seen in Figs. 2A - 2F, smart bracket 114 may be formed by a combination of a single weight sensor assembly 112 mounted on two adjacent brackets 110B and 110C. Although not shown in Figs. 2A - 2F, smart bracket 114 may alternatively comprise a single weight sensor assembly 112 mounted on a single corresponding bracket 110. In typical in-store installations, a row of shelves 102 may include both of these arrangements, with smart bracket 114 at the end of a row of shelves being mounted on only one bracket 110, since no second bracket is available, and smart brackets 114 in the middle of a row of shelves, being mounted on two adjacent brackets 110.

[0045] Weight sensor assembly 112 includes at least one weight sensor 202, e.g., two or three weight sensors. Here, by way of example, smart bracket 114 is shown to include two weight sensors 202, although more or less are also possible depending on design constraints and operational requirements. Sensors 202 are spatially distributed along an upper edge of brackets 110. Sensors 202 may be embodied as load cells, although it is appreciated that any suitable type of weight sensor may be employed in the present solution.

[0046] Weight sensor assembly 112 may further include a bracket adapter 204 mounted on at least one bracket 110 and configured to hold the at least one weight sensor 202. Bracket adapter 204 is adapted to directly interface with (connect to) one or two of brackets 110. Here, by way of example, bracket adapter 204 is seen to include a first bracket adapter portion 204A, holding (housing) one weight sensor 202 and a second bracket adapter portion 204B, holding (housing) another weight sensor 202. It is appreciated, however, that other alternative configurations of bracket adapter 204 are possible, depending on design requirements, including, by way of example only, one continuous bracket adapter 204 holding a plurality of weight sensors 202, and a bracket adapter 204 having more than two bracket adapter portions.

[0047] Bracket adapter 204 includes a bracket mounting feature 206, for facilitating mounting of bracket adapter 204 upon at least one bracket 110. Here, by way of example, each of first and second bracket adapter portions 204 A and 204B, may be formed with two adjacent parallel bracket insertion grooves 208 and 210, e.g., extending longitudinally along bracket adapters 204A and 204B. Bracket HOB slots into bracket insertion groove 208 and bracket HOC slots into bracket insertion groove 210. It is appreciated, however, that bracket insertion grooves 208 and 210 are just one possible exemplification of bracket mounting feature 206. Other bracket mounting features 206 are also possible.

[0048] For convenience of manufacturing, bracket mounting feature 206 may be adapted to facilitate mounting of bracket adapter 204 upon two brackets, such as brackets 110B and 110C. In the case that bracket adapter 204 is mounted on only one bracket, part of bracket mounting feature 206 may be unused but bracket adapter 204 is nonetheless mountable on only one bracket 110. This design option may be more convenient from a manufacturing viewpoint, since bracket adapter 204 is compatible with mounting on one bracket or on two brackets, without requiring modification thereto depending on the particular use case. It is understood, however, that bracket adapter 204 may alternatively include dedicated features for one-bracket mounting and two- bracket mounting respectively, or different bracket adapters may be provided for each case.

[0049] Weight sensor assembly 112 further includes a shelf adapter 220 configured to support at least one shelf thereupon. Shelf adapter 220 is in contact with at least one weight sensor 202, either directly or indirectly. Shelf adapter has a first shelf support arm 222 and a second shelf support arm 224, adapted to respectively support a first and second shelf, such as shelves 104 and 106 (Fig. 1). It is understood that in a conventional shelving system, in the absence of weight sensor assembly 112, the first and second shelves would respectively be directly supported by brackets HOB and 110C. Shelf adapter 220 interfaces with the shelves that would otherwise be directly supported by brackets HOB and 110C.

[0050] In cases in which weight sensor assembly 112 is mounted upon only one bracket 110, and supports only one shelf 102, it is understood that only one shelf support arm 222, 224 of shelf adapter 220 may be required.

[0051] In some embodiments of shelf adapter 220, first and second shelf support arms 222 and 224 may be formed as a U-shaped unitary structure, as seen, for example, in a smart bracket 314 having U-shaped unitary shelf adapter 320, shown in Figs. 3 A - 3E.

[0052] Weight sensor assembly 112 additionally includes electronic circuitry 230, also referred to as an electronic kit, electronically coupled to weight sensors 202. Electronic circuitry 230 provides electricity to weight sensors 202, for example from the electrical grid. Electronic circuitry 230 also includes signal processing circuitry (also referred to as a signal processor, implemented in software and / or hardware), for receiving and processing data received from weight sensors 202. Signal processing circuitry of electronic circuitry 230 is operative to receive signals generated by weight sensors 202 and to process the signals in order to ascertain, based on said signals, whether one or more items have been picked from or placed on said at least one shelf. Weight sensors 202 sense changes in weight in real time and transmit the sensed data, in real time, to electronic circuitry 230. Signal processing circuitry in electronic circuitry 230 may be embodied as a dedicated electronic circuit board, employing signal processing methods to detect interactions of shoppers with items on shelve 102, based on the signals generated by weight sensors 202.

[0053] Electronic circuitry 230 may further include transmission circuitry, for sending a message to external server 122, including details of the detected interaction. Weight sensors 202 are electrically connected to electronic circuitry 230. For example, a cabled connection 232 may be provided between weight sensors 202 and electronic circuitry 230.

[0054] Weight sensor assembly 112 may be adapted to fit upon various different types of brackets, such as smart brackets 414 shown in Figs. 4A - 4E. The smart bracket of the present disclosure may be easily customized to fit a range of various different types of brackets, such as brackets for fridge, freezer, bakery shelves or others. Figs. 2A - 4E show various exemplary smart brackets 114, 314, 414 Other designs are also possible.

[0055] As explained hereinabove, in some embodiments of the present invention, a weight sensor assembly 112 may be mounted on (across) two adjacent brackets 110. An array of shelves may include a plurality of smart brackets 114 mounted in this way, at junctions between adjacent shelves, resulting in an interconnected ‘chain’ of shelves. Changes in weight of a given shelf, forming part of the chain of shelves, may thus be sensed not only by the smart brackets directly supporting the given shelf but also by other smart brackets further along the chain of shelves, due to the shelves being interconnected. Weight signals measured in such an arrangement require specialized signal processing methods, in order to accurately identify the origin of the interaction or interactions causing the changes in weight. Two exemplary methods useful for analyzing such signals are detailed hereinbelow. Such methods may be implemented locally, for example by a signal processor (implemented in hardware and / or software) within the smart brackets themselves; remotely, for example by signal processing functionality (implemented in hardware and / or software) at server 122; or by a combination of both.

[0056] EXAMPLE SIGNAL PROCESSING METHODS

[0057] This section describes several signal processing methods for analyzing the signals produced by weight sensors in order to identify interactions accurately and reliably. The description below refers to the methods as being carried out by "a signal processor". In various embodiments, any of the disclosed methods can be carried out by signal processing circuitry (in hardware and / or software) embodies in the smart brackets (e.g., 114, 314, 414) and / or in external server 122.

[0058] 1. Dynamic chain method

[0059] Turning now to Fig. 5, a highly schematic representation of a row of four shelves, SI - S4 supported by five smart brackets Bl - B5 is shown in a top panel 500. Shelf SI is supported at one end thereof by smart bracket Bl and at the other end thereof by smart bracket B2; shelf S2 is supported at one end thereof by smart bracket B2 and at the other end thereof by smart bracket B3; shelf S3 is supported at one end thereof by smart bracket B3 and at the other end thereof by smart bracket B4; and shelf S4 is supported at one end thereof by smart bracket B4 and at the other end thereof by smart bracket B5. Smart bracket B2 is thus common to shelves SI and S2; smart bracket B3 common to shelves S2 and S3 and smart bracket B4 common to shelves S3 and S4. Smart brackets Bl - B5 include brackets 502. Shelves SI - S4 would be directly supported by conventional brackets 502, if not for the mounting of weight sensor assemblies thereon, thereby converting brackets 502 into smart brackets.

[0060] As appreciated from consideration of the row of shelves in top panel 500, shelves SI - S4 are interconnected by virtue of the shared smart brackets therebetween, linking adjacent shelves. Shelves SI - S4 may be referred to as a chain of shelves.

[0061] In some embodiments, the signal processor considers shelves SI - S4 as a single extended shelf, and hence to use data from all weight sensors of all smart brackets Bl - B5 in order to calculate parameters of a sensed interaction. Although feasible, such an approach may result in two sources of error. Firstly, the location of an interaction along the chain of shelves may be incorrectly identified, due to unknown fixture inaccuracies and noisy signals obtained from smart brackets distant from the origin of an interaction. Secondly, data from all of smart brackets Bl - B5 may not be available, due to communication constraints.

[0062] In an alternative embodiment, for a given interaction, the signal processor identifies a partial subset of (one or more of) the smart brackets as "active" brackets. The other smart brackets are considered "inactive" for the given interaction. In this embodiment, the signal processor considers only the signals obtained from the active smart brackets rather than the entirety of smart brackets along the chain of shelves. In this approach, for a given interaction, the system may ascertain which smart brackets in the chain of smart brackets were active during the given interaction. The system may then create a sub-chain including only those smart brackets identified as active brackets. Subsequent signal processing for the given interaction may be based only on signals measured by those smart brackets belonging to the sub-chain of smart brackets.

[0063] The signal processor may identify a smart bracket as being active or inactive based on an extent of fluctuations in the signal measured by its weight sensor(s). For example, a smart bracket may be identified as active when fluctuations in signal exceed a predetermined or learned threshold and as inactive when fluctuations in signal are less than a predetermined or learned threshold.

[0064] By way of example, an interaction on shelf S4 is represented by an arrow in a middle panel 504 of Fig. 5. An interaction on shelf S4 may cause smart brackets B3 - B5 to be active, whilst smart brackets B 1 and B2 are inactive. Here, inactive brackets are shown in broken lines. A subchain consisting only of smart brackets B3 - B5 may be created for the interaction on shelf S4 and only these smart brackets taken into consideration for subsequent signal processing and analysis. Further by way of example, as seen in a bottom panel 506 of Fig. 5, an interaction on shelf S2 may cause smart brackets Bl - B4 to be active and hence only these smart brackets to be subsequently included in a sub-chain taken into account for further analysis. Such further analysis may include center of mass and moment of equilibrium calculations, for example.

[0065] The delineation between active and inactive smart brackets serves to improve the accuracy with which interactions may be detected and characterized, notwithstanding the interconnected arrangement of shelves and smart brackets.

[0066] In an embodiment, after classifying the smart brackets (e.g., Bl - B4) into active and inactive brackets, the signal processor also wakes up and activates any inactive bracket that is an immediate neighbor of an active bracket. As part of activating a previously inactive weight sensor, the processor causes the sensor to transmit a signal indicative of the sensed load. Referring to panel 504, for example, the signal processor may wake up bracket B2, which is an immediate (direct, adjacent) neighbor of bracket B3, even though B2 was not initially identified as active. For subsequent processing, such neighbor brackets are also considered active and added to the chain of shelves.

[0067] 2. Sensor response model method

[0068] Turning now to Fig. 6, top panel 600 shows two interactions, each represented by an arrow. The two interactions occur simultaneously on shelves SI and S3. The dynamic chain method, described hereinabove with reference to Fig. 5, may be used to ascertain active brackets relevant to these two simultaneous interactions. For example, based on the dynamic chain method, smart brackets B 1 - B4 may be identified as active.

[0069] Should signals measured by smart brackets Bl - B4 be analyzed using simple center of mass or moment of equilibrium calculations, the analysis may falsely yield a single interaction on S2, as shown in a second panel 602 of Fig. 6.

[0070] This erroneous result may be avoided by the use of a novel sensor response model method to analyze signals. The sensor response model method is based on the understanding that a given interaction at a given location along the chain of shelves will result in a particular pre-known response of a given sensor, for each smart bracket. The variation of sensor response as a function of the sensor's distance from the interaction location is referred to herein as the sensor's response profile (the terms "response profile" and "response model" are used interchangeably herein). Thus, for example, the response profile is typically maximal for an interaction location that coincides with the sensor's location. As the interaction location moves away from the sensor's location, the sensor's response becomes weaker and the response profile therefore decreases.

[0071] An example of variation of the response profile with interaction location, for smart brackets Bl and B3, is shown in a third panel 604 of Fig. 6. Variation of the sensor response with interaction location for smart bracket Bl is shown by a dashed line 606. Variation of the sensor response with interaction location for bracket B3 is shown by a solid line 608. Response profiles of the sensors may be derived from an analytical model, an advanced simulation, a machine learning model based on historical sensor data or a combination of these methods or others. Response profiles for an interaction in which an item is placed on a shelf may have the same shape as response profiles for an interaction in which an item is picked from a shelf, but with opposite sign. For example, an interaction in which an item is placed on a shelf may create a positive response profile, whereas an interaction in which that item is picked from the shelf may create a negative response profile.

[0072] Since each interaction creates a given response profile per smart bracket, the response profiles of sensors for two interactions on two different shelves will differ from the response profiles for those sensors for a single interaction on one shelf. In some embodiments, the signal processor may compare measured sensor responses to the expected response profiles in order to (i) distinguish between a single interaction and multiple simultaneous interactions, and / or (ii) in the case of multiple simultaneous interactions, identify the number and locations of the interactions. In the example of Fig. 6, the sensor response model method may be used to identify active brackets and create sub-chains for each of the interactions on shelves SI and S3, and thus analyze the interaction with greater accuracy.

[0073] For a given interaction, variation of the response profile across the brackets is illustrated in a bottom panel 610 in Fig. 6. An interaction occurring on shelf SI has a first modelled sensor response profile across brackets Bl - B4, as indicated by a first dashed line 612; an interaction occurring on shelf S2 has a modelled sensor response profile indicated by a second dotted line 614 and an interaction occurring on shelf S3 has a modelled sensor response profile indicated by a third solid line 616. It is appreciated from consideration of panel 610, that a given interaction at a given location creates a unique modeled sensor response profile.

[0074] When using the sensor response model method, the signal processor may employ likelihood methods to find the location of an interaction, based on the distribution of weights between shelves. The location of an interaction may be expressed as the location of the shelf upon which the interaction occurred and the location within the shelf at which the interaction occurred. The location within the shelf may be found based on modeled distribution of weight between shelves, depending on where along a given shelf an item is located. The likelihood method may be used to identify the interaction most likely to give rise to the measured signal distribution. This may also be useful in analyzing weight loss of items due to friction with the surroundings.

[0075] An additional use of the sensor response mode relates to detection and reaction to faults. In some embodiments, the signal processor may detect various types of faults by comparing the signals provided by the weight sensors to the sensor response model. Faults that can be detected in this manner include, for example, a faulty weight sensor, a mechanical or electrical fault in a shelf bracket that prevents reliable weight sensing, or an item stuck on a shelf in a manner that prevents reliable weight sensing. Upon detecting a fault, the signal processor may initiate various responsive actions, e.g., issuing an alert to a technician or other staff to resolve the fault. As another example, upon detecting a fault affecting the weight sensing, the signal processor may revert to (or give more weight to) other sensors, e.g., video cameras, used for identifying interactions.

[0076] EXAMPLE METHOD FLOWS

[0077] Reference is now made to Fig. 7, which is a simplified flow chart, illustrating a method in accordance with an embodiment of the present invention.

[0078] As seen in Fig. 7, a method 700 may include mounting at least one weight sensor on at least one shelf bracket, between the at least one shelf bracket and at least one shelf (702). In some embodiments of the method, the at least one weight sensor may be mounted on two adjacent shelf brackets, straddling both shelf brackets. The method may include sensing, by the at least one weight sensor, a load on the at least one shelf (704). For example, the at least one weight sensor may sense a load on one shelf supported thereby or on both shelves directly supported thereby. The at least one weight sensor may additionally sense a load on other shelves, more distant from the at least one weight sensor and not directly supported thereby, due to adjacent shelves being interconnected by shared sensors.

[0079] The method may include communicating the sensed signals to a signal processor (706). The signal processor may be a component mounted on the shelf bracket, in communication with the at least one weight sensor, and / or a component that is part of central server 122. The method may further include the signal processor processing the signal to ascertain the occurrence or nature of an interaction that has occurred, such as weight change associated with the interaction, location, time, nature of the interaction or a combination thereof (708). This information may be useful for inventory management by an inventory management system which may employ method 700.

[0080] Reference is now made to Fig. 8, which is a simplified flow chart, illustrating a method in accordance with an embodiment of the present invention.

[0081] As seen in Fig. 8, a method 800 may include mounting at least one first weight sensor on at least one first shelf bracket, between the at least one shelf bracket and at least one shelf (802). In some embodiments of the method, the at least one first weight sensor may be mounted on two adjacent shelf brackets, straddling both shelf brackets.

[0082] The method may further include mounting at least one second weight sensor on at least one second shelf bracket, between the at least one second shelf bracket and the at least one shelf (804). In some embodiments of the method, the at least one second weight sensor may be mounted on two adjacent shelf brackets, straddling both shelf brackets.

[0083] According to method 800, the shelf is supported by first and second shelf brackets, each having at least one weight sensor mounted thereon. In a conventional shelving system, the first and second shelf brackets would directly support the shelf. In accordance with an embodiment of the present invention, the first and second shelf brackets are converted to smart brackets by way of installation of the first and second weight sensors respectively thereon. For example, the first and second shelf brackets may typically be located at either end of a shelf.

[0084] The method may include sensing, by the at least one first and second weight sensors, a load on the at least one shelf (806). For example, the first and second weight sensors may each sense a load on the shelf supported thereby or on both shelves directly supported thereby. The first and second weight sensors may additionally sense a load on other shelves, more distant from the first and second weight sensors and not directly supported thereby, due to adjacent shelves being interconnected by shared sensors.

[0085] The method may include outputting signals, by the at least one first and second weight sensors, indicative of the sensed load (808). By way of example, the first and second weight sensors may respectively output signals to first and second signal processors in respective communication therewith. The signal processor may be a component mounted on the conventional shelf bracket, in communication with the at least one weight sensor. The method may further include the signal processors processing the signal to ascertain the occurrence or nature of an interaction that has occurred, such as weight change associated with the interaction, location, time, nature of the interaction or a combination thereof (810). This information may be useful for inventory management by an inventory management system which may employ method 800. ADDITIONAL SMART BRACKET CONFIGURAION

[0086] Figs. 10A and 10B are simplified respective assembled and rear views of a smart bracket 1000 suitable for use in a system of the type shown in Figs. 1 A and IB, constructed and operative in accordance with another embodiment of the present invention. In the present example, smart bracket 1000 comprises the following elements:

[0087] - Shelf support arms 1002 and 1004, configured to support respective adjacent shelves,

[0088] - Weight sensors 1006 and 1008, configured to sense, via shelf support arms 1002 and

[0089] 1004, weight / force exerted on the shelves.

[0090] - Bracket adapters 1010 and 1012, configured to enclose and provide mechanical support for weight sensors 1006 and 1008, respectively. Bracket adapter 1012 also serves as a housing for the electronic circuitry of smart bracket 1000.

[0091] - A support arm 1014, configured to provide mechanical support for bracket adapters

[0092] 1010 and 1012, and also to run cables between bracket adapters 1010 and 1012.

[0093] The various mechanical and electrical smart bracket configurations described herein are example configurations that are chosen purely for the sake of conceptual clarity. In alternative embodiments, any other suitable configuration can be used. In various embodiments, the electronic circuitry of the disclosed smart brackets may be implemented using suitable software, using suitable hardware such as one or more Application-Specific Integrated Circuits (ASIC) or Field-Programmable Gate Arrays (FPGA), or using a combination of hardware and software.

[0094] Although the embodiments described herein mainly address smart shelf brackets for use in retail stores, the methods and systems described herein can also be used in other applications.

[0095] It will thus be appreciated that the embodiments described above are cited by way of example, and that the present invention is not limited to what has been particularly shown and described hereinabove. Rather, the scope of the present invention includes both combinations and sub-combinations of the various features described hereinabove, as well as variations and modifications thereof which would occur to persons skilled in the art upon reading the foregoing description and which are not disclosed in the prior art. Documents incorporated by reference in the present patent application are to be considered an integral part of the application except that to the extent any terms are defined in these incorporated documents in a manner that conflicts with the definitions made explicitly or implicitly in the present specification, only the definitions in the present specification should be considered.

Claims

CLAIMS1. A smart-bracket system, comprising: shelf brackets configured to support a chain of two or more shelves; weight sensors coupled to at least one of the shelf brackets, each weight sensor configured to sense a load on a respective shelf bracket and to output a signal indicative of the sensed load, wherein a given weight sensor among the weight sensors is configured to output a signal indicative of the load on first and second adjacent shelves; and one or more signal processors, configured to identify, based on one or more signals output by one or more of the weight sensors, including the given weight sensor, an interaction in which one or more items are picked from or placed on one or more of the shelves.

2. The system according to claim 1, wherein a given shelf bracket is a common bracket configured to support both the first shelf and the second shelf, and wherein the given weight sensor is coupled to the common bracket so as to sense the load on both the first shelf and the second shelf.

3. The system according to claim 1, wherein the given weight sensor is coupled to a shelf bracket that supports the first shelf, and wherein the first and second shelves are mechanically coupled to one another, thereby causing the given weight sensor to sense the load on both the first shelf and the second shelf.

4. The system according to any of claims 1-3, wherein the one or more signal processors are configured to estimate a location of the interaction along the chain of shelves based on the signals.

5. The system according to claim 4, wherein the one or more signal processors are configured to identify a subset of the shelf brackets whose weight sensors are active due to the interaction, and to estimate the location of the interaction based on the signals output by the weight sensors of the shelf brackets in the subset.

6. The system according to claim 5, wherein the one or more signal processors are configured to activate and add to the subset at least one shelf bracket that is an immediate neighbor of a shelf bracket in the subset.

7. The system according to any of claims 1-3, wherein the one or more signal processors are configured to estimate the location of the interaction by comparing the signals to a response model that specifies a response of a weight sensor as a function of distance from the location of the interaction.

8. The system according to claim 7, wherein, based on the response model, the one or more signal processors are configured to distinguish between a single interaction and multiple simultaneous interactions.

9. The system according to claim 8, wherein, upon identifying multiple simultaneous interactions, the one or more signal processors are configured to estimate, based on the response model, respective locations of the multiple simultaneous interactions.

10. The system according to claim 7, wherein, by comparing the signals to the response model, the one or more signal processors are configured to detect a fault and to initiate a responsive action.

11. A method, comprising: in a system that includes shelf brackets that support a chain of two or more shelves, operating a plurality of weight sensors coupled to at least one of the shelf brackets, so that each weight sensor senses a load on a respective shelf bracket and outputs a signal indicative of the sensed load, wherein a given weight sensor among the weight sensors outputs a signal indicative of the load on first and second adjacent shelves; and identifying, based on one or more signals output by one or more of the weight sensors, including the given weight sensor, an interaction in which one or more items are picked from or placed on one or more of the shelves.

12. The method according to claim 11, wherein a given shelf bracket is a common bracket configured to support both the first shelf and the second shelf, and wherein the given weight sensor is coupled to the common bracket so as to sense the load on both the first shelf and the second shelf.

13. The method according to claim 11, wherein the given weight sensor is coupled to a shelf bracket that supports the first shelf, and wherein the first and second shelves are mechanicallycoupled to one another, thereby causing the given weight sensor to sense the load on both the first shelf and the second shelf.

14. The method according to any of claims 11-13, wherein identifying the interaction comprises estimating a location of the interaction along the chain of shelves based on the signals.

15. The method according to claim 14, wherein identifying the interaction comprises identifying a subset of the shelf brackets whose weight sensors are active due to the interaction, and estimating the location of the interaction based on the signals output by the weight sensors of the shelf brackets in the subset.

16. The method according to claim 15, wherein identifying the interaction comprises activating and adding to the subset at least one shelf bracket that is an immediate neighbor of a shelf bracket in the subset.

17. The method according to any of claims 11-13, wherein identifying the interaction comprises estimating the location of the interaction by comparing the signals to a response model that specifies a response of a weight sensor as a function of distance from the location of the interaction.

18. The method according to claim 17, wherein identifying the interaction comprises distinguishing between a single interaction and multiple simultaneous interactions based on the response model.

19. The method according to claim 18, wherein, upon identifying multiple simultaneous interactions, the one or more signal processors are configured to estimate, based on the response model, respective locations of the multiple simultaneous interactions.

20. The method according to claim 17, wherein, by comparing the signals to the response model, the one or more signal processors are configured to detect a fault and to initiate a responsive action.

Images