Automatic battery sorting system

Abstract

The present disclosure relates to systems, non-transitory computer-readable media, and methods for analyzing and sorting batteries according to battery categories. Specifically, in one or more embodiments, the disclosed system scans a target battery using a plurality of sensors and analyzes signals of the plurality of sensors using a classifier model to determine a battery category of the target battery. Moreover, in some embodiments, the disclosed system indicates the predicted battery category to a battery sorting mechanism for sorting the target battery. Additional mechanisms and related methods for automatic classification and sorting of batteries are disclosed.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of and priority to U.S. Provisional Patent Application No. 63 / 373,380, filed on August 24, 2022, U.S. Patent Application No. 18 / 452,961, filed on August 21, 2023, and U.S. Patent Application No. 18 / 452,974, filed on August 21, 2023. Each of the above applications is incorporated herein by reference in its entirety. Background Art

[0003] As an alternative to fossil fuels and other energy sources, recent years have seen a significant increase in the implementation of various types of batteries. In addition, the recent surge in the popularity of electric vehicles and other electronic devices has led to a significant increase in the demand for battery production, as well as an increase in the need for safe and efficient recycling or disposal of batteries and battery materials.

[0004] Despite advances in battery production and the implementation of batteries in a variety of use areas, existing methods for the bulk recycling and / or disposal of batteries and battery materials face several deficiencies. For example, conventional systems can safely recycle or dispose of batteries of specific chemical compositions, such as lithium-ion batteries. However, certain components of batteries are difficult or impossible to recycle in a safe and efficient manner. Therefore, before recycling or disposal, different categories of batteries must be sorted with a high level of certainty. Unfortunately, conventional methods of sorting and sorting batteries typically require human attention and careful inspection of each battery, resulting in an inefficient and often inaccurate process when handling large quantities of batteries for recycling and / or disposal.

[0005] These, along with additional problems and challenges, exist in conventional battery recycling systems. Summary of the invention

[0006] Embodiments of the present disclosure provide benefits and / or solve one or more of the foregoing or other problems in the art with systems, devices, non-transitory computer-readable media, and methods for automatically classifying and sorting different categories of batteries. For example, the disclosed system utilizes a variety of sensors and sorting mechanisms to intelligently classify and sort batteries according to chemical composition, form factor, and / or other categories.

[0007] In some embodiments, for example, an apparatus for sorting batteries of different configurations includes: a hopper mechanism configured to release batteries individually onto a conveyor; a scanner mechanism having a plurality of sensors configured to determine a predicted battery configuration for each battery. The apparatus also includes a sorting mechanism array configured to deliver the batteries to a plurality of bins according to the predicted battery configuration. In one or more embodiments, additional sensors (such as infrared (IR) cameras) are implemented to detect anomalies (such as high temperatures) in released batteries.

[0008] In addition, in one or more embodiments, the battery classification system receives multiple signals corresponding to the target battery from multiple sensors. For example, the multiple sensors include two or more of an x-ray scanning array, a three-dimensional (3D) scanner, and an RGB camera or an infrared camera. In response, using a classifier model, the system determines a predicted battery category of the target battery from multiple battery categories based on the multiple signals. In one or more embodiments, the system indicates the predicted battery category to the battery sorting mechanism.

[0009] Thus, the disclosed embodiments provide significant advantages over existing solutions, such as increased efficiency achieved by automating battery classification and sorting. Further, the disclosed embodiments present increased safety by eliminating the need for human interaction with potentially volatile or unstable batteries, by accurately predicting battery class to avoid accidents in subsequent handling, and by providing increased measures for detecting battery anomalies throughout the classification and sorting process.

[0010] Additional features and advantages of one or more embodiments of the present disclosure are summarized in the description which follows, and in part will be obvious from the description, or may be learned by practice of such exemplary embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The detailed description provides one or more embodiments with additional specificity and detail through use of the accompanying drawings, as briefly described below.

[0012] Figure 1 Illustrated is a perspective view of an automated battery sorting system according to one or more embodiments.

[0013] Figure 2 Additional perspective views of an automated battery sorting system are illustrated in accordance with one or more embodiments.

[0014] Figure 3 Illustrated is a top view of an automated battery sorting system according to one or more embodiments.

[0015] Figure 4Illustrated is a perspective view of a scanning bay of an automated battery sorting system according to one or more embodiments.

[0016] FIG. 5A to FIG. 5B Additional perspective views of a scanning bay of an automated battery sorting system are illustrated in accordance with one or more embodiments.

[0017] Figure 6 Illustrated is a perspective view of a battery tipper and hopper assembly of an automated battery sorting system according to one or more embodiments.

[0018] Figure 7 Illustrated is a simplified diagram of an environment in which a battery classification system may operate in accordance with one or more embodiments.

[0019] Figure 8 A schematic diagram of a battery classification system is illustrated in accordance with one or more embodiments.

[0020] Fig. 9 A battery sorting system utilizing an RGB camera is illustrated in accordance with one or more embodiments.

[0021] Fig.10 A battery sorting system utilizing a three-dimensional (3D) scanner and an X-ray scanning array is illustrated in accordance with one or more embodiments.

[0022] Fig.11 A flow chart illustrating a series of actions for classifying and sorting batteries according to one or more embodiments is illustrated.

[0023] Fig.12 A block diagram of an example computing device for implementing one or more embodiments of the present disclosure is illustrated. DETAILED DESCRIPTION

[0024] The present disclosure describes one or more embodiments of a battery classification system that intelligently classifies batteries of different configurations based on inputs from different sensors. For example, in some implementations, the battery classification system uses a battery classifier machine learning model to predict the battery class of a target battery based on different sensor signals. In addition, some embodiments include a battery sorting mechanism that is configured to sort batteries according to the battery class predicted by an embodiment of the battery classification system.

[0025] In some embodiments, for example, a battery classification system receives a plurality of signals corresponding to a target battery from a plurality of sensors. In response, the battery classification system determines a predicted battery category of the battery from a plurality of battery categories using a classifier machine learning model based on the plurality of signals. Furthermore, in some embodiments, the battery classification system indicates the predicted battery category to a battery sorting mechanism. Furthermore, in some embodiments, the plurality of sensors include two or more of an x-ray scanning array, a three-dimensional (3D) scanner, an RGB camera, or an infrared camera.

[0026] In addition, one or more embodiments of the apparatus for sorting batteries of different configurations include a hopper mechanism, a scanner mechanism, and a sorting mechanism array. In some embodiments, the hopper mechanism is configured to release batteries from a plurality of batteries individually onto a conveyor. Further, in some embodiments, the scanner mechanism is arranged around the conveyor and includes multiple types of sensors configured to determine a predicted battery configuration for each battery from a plurality of battery configurations. Moreover, in some embodiments, the sorting mechanism array is arranged around the conveyor and configured to transfer the plurality of batteries from the conveyor to a plurality of boxes based on the predicted battery configuration for each battery in the plurality of batteries.

[0027] Thus, the disclosed embodiments provide for automatic classification and sorting of batteries, including multiple categories, such as, but not limited to, chemical composition, form factor, and other configurations. In addition, utilizing various sensors to scan and analyze individual batteries, such as, but not limited to, three-dimensional (3D) scanners, x-ray scanning arrays, RGB cameras, and infrared (IR) cameras, the disclosed system accurately and efficiently classifies and sorts batteries for further processing. In fact, the disclosed system can identify / classify different categories of batteries and sort a large number of batteries accordingly, while improving safety and efficiency. In addition, one or more of the disclosed embodiments utilize a classifier machine learning model to intelligently and efficiently predict battery categories based on data provided by a database of battery categories.

[0028] Accordingly, the disclosed battery classification and sorting system provides many advantages and benefits over conventional systems and methods. For example, by utilizing multiple sensor types and trained classifier models, the battery classification system improves accuracy relative to conventional systems. Specifically, the disclosed embodiments utilize a classifier machine learning model to analyze and compare various attributes provided by sensors with a database of battery categories to intelligently identify battery categories in an automated process.

[0029] Furthermore, by utilizing an automated process for determining battery categories using a trained classifier model, the disclosed system improves efficiency relative to conventional systems. Specifically, the disclosed embodiments automatically and in most implementations intelligently analyze and determine the categories of a large number of batteries without requiring human consideration of each battery. Thus, the disclosed system can accurately and efficiently sort batteries into batches according to their predicted categories.

[0030] In addition, the disclosed system improves the safety of the process for identifying and sorting batteries for recycling and / or disposal. Specifically, the disclosed system reliably predicts battery class to avoid inadvertent handling of batteries due to faulty procedures. For example, batteries with certain chemistries may be impossible or uneconomical to recycle through typical processes, while other types of batteries are readily recycled through known procedures. Furthermore, by utilizing sensors (such as infrared (IR) cameras) to scan batteries for anomalies (e.g., elevated temperatures), the disclosed system can prevent accidents involving volatile materials that are typical of many batteries.

[0031] As demonstrated by the above discussion, the present disclosure utilizes a variety of terms to describe the features and advantages of the disclosed system. Additional details about the meaning of these terms are now provided. For example, as used herein, the term "machine learning model" refers to a model that can be trained and / or tuned to approximate an unknown function based on an input. For example, the term "machine learning model" may include, but is not limited to, a random forest model, a decision tree (e.g., a series of gradient-boosted decision trees (e.g., XGGost algorithm)), a multilayer perceptron, a linear regression, a support vector machine, a deep learning architecture, a deep learning transformer (e.g., a self-attention transformer), or a logistic regression. In other embodiments, the machine learning model includes a neural network, such as a convolutional neural network, a recursive neural network (e.g., LSTM), a graph neural network, a self-attention transformer neural network, a diffusion neural network, or a generated adversary neural network.

[0032] As used herein, the term "x-ray attenuation" refers to the reduction in intensity of x-rays as they pass through matter, which can be caused by absorption or deflection of photons from the x-ray beam. X-ray attenuation can be affected by different factors, such as beam energy and atomic energy of the material absorbing the x-rays. Specifically, the disclosed system takes into account the detected x-ray attenuation when determining the chemical and / or material composition of the target battery, as discussed in more detail below.

[0033] As used herein, the term "object recognition" refers to a computer technique for finding and identifying objects in an image. For example, the term "object recognition" may include, but is not limited to, appearance-based methods, edge matching, segmentation methods, grayscale matching, gradient matching, object detection neural networks, decision trees, or any of a variety of models trained to recognize objects within an image.

[0034] Additional details will now be provided regarding illustrative drawings that depict example embodiments and implementations of the disclosed methods, devices, and systems. For example, Figures 1 to 6 1 illustrates multiple views of an automated battery sorting system 100 according to one or more embodiments. Specifically, Figure 1 illustrates a perspective view of an automated battery sorting system 100, Figure 2 Additional perspective views of the automated battery sorting system 100 are illustrated, and Figure 3 The figure shows a top view of the automated battery sorting system 100. In addition, Figure 4 illustrates a perspective view of the scanning bays 102 and 104 of the automated battery sorting system 100, and FIG. 5A to FIG. 5B Additional perspective views of scanning bays 102 and 104 are illustrated. Figure 6 A perspective view of the tipping mechanism 110 and the hopper 112 of the automated battery sorting system 100 is illustrated.

[0035] like Figures 1 to 3 As shown, the automated battery sorting system 100 includes a tipping mechanism 110 configured to receive a large number of batteries for sorting. In some embodiments, the tipping mechanism 110 includes a large chamber or box 126 into which unsorted batteries can be placed to start the sorting process. In some embodiments, for example, the battery load can be directly transferred from a battery collection container, a barrel, a truck bed, a pallet, etc. to the box 126. In addition, the tipping mechanism 110 can include a device for moving the batteries from the box 126 to the hopper 112, such as but not limited to a movable platform, a slide, a pneumatic lifter and / or a tipper, or a conveyor belt. Therefore, the batteries placed in the box 126 enter the hopper 112 from the box 126 via the tipping mechanism 110, which forces the batteries to pass individually for subsequent scanning. In some embodiments, for example, the hopper includes a vibrating table configured to separate the batteries in the batch by form factor and / or weight. For example, in some embodiments, the hopper 112 removes smaller batteries from the batch as the smaller batteries move away from relatively larger batteries. Additional details and examples regarding the tipping mechanism 110 and the hopper 112 are provided below with respect to Figure 6 Give a description.

[0036] like Figures 1 to 3As further illustrated in , in some embodiments, individual batteries are transferred from the hopper 112 onto an inclined conveyor 114 and dropped from a feeder 116 for individual scanning, sorting, and sorting. Alternatively, in one or more embodiments, the batteries are directed directly from the hopper 112 toward a scanning area of ​​the automated battery sorting system 100. For example, the outlet of the hopper 112 may be positioned directly above or otherwise adjacent to a scanning area or horizontal conveyor of the automated battery sorting system 100. Further, in some embodiments, the automated battery sorting system 100 includes one or more actuators and / or structures configured to align incoming batteries prior to entering a subsequent scanning area. For example, the automated battery sorting system 100 may align individual batteries for individual measurement and individual ejection / diversion into an associated storage bin. Additionally, in some embodiments, one or more alignment mechanisms (e.g., actuators, structures, etc.) can position incoming cells at specific angles relative to x-ray and other scanning sources to increase measurement accuracy. Alignment mechanisms can include, but are not limited to, rails, pneumatic arms, vibrating tables, telescopic supports, gates, etc.

[0037] In the illustrated embodiment, each battery is individually slid or otherwise transferred downward from the feeder 116 toward the first scanning compartment 102 and the second scanning compartment 104. Specifically, the first scanning compartment 102 is equipped with one or more sensors, such as but not limited to one or more RGB cameras, infrared (IR) cameras, or 3D scanners. In addition, the 3D scanner may include but is not limited to one or more of an optical camera, a 3D laser scanner, a computed tomography scanner, a structured light 3D scanner, a LiDAR, or a time-of-flight laser scanner. Therefore, in one or more embodiments, the automated battery sorting system 100 utilizes multiple vision sensors to scan each battery within the first scanning compartment 102. As further described below (e.g., with respect to Figures 7 to 10 ), the automated battery sorting system 100 utilizes a plurality of sensors to determine the properties of each battery based on sensor signals, and predicts the battery category or type of each battery based on the determined properties.

[0038] In addition, the sensor of the first scanning compartment 102 (optionally, the second scanning compartment 104) of the automated battery sorting system 100 includes an x-ray scanning array configured to emit an x-ray beam and detect the x-ray beam as it passes through each battery for further analysis (i.e., to provide additional input to the battery classifier model). In some embodiments, either or both of the first scanning compartment 102 and the second scanning compartment 104 further include an alignment mechanism configured to position each battery in an ideal position and / or orientation for the scanning process. Such an alignment mechanism may include, for example, a series of buffers / rails, a movable actuator, a tumbler, or other devices capable of repositioning small batteries to large batteries. As also mentioned, in response to scanning each battery at the first scanning compartment 102 and the second scanning compartment 104 to determine multiple attributes of each battery, the automated battery sorting system 100 utilizes a classifier model to determine a predicted battery category for each battery.

[0039] In any case, the automated battery sorting system 100 includes at least one scanning compartment and a plurality of sensors. Figures 1 to 3 As shown, in at least one embodiment, the automated battery sorting system 100 includes two scanning compartments 102 and 104, each accommodating a different type of scanner. In one or more implementations, the first scanning compartment 102 includes one or more sensors that detect / measure / capture a first type of attribute of the battery. Similarly, the second scanning compartment 104 includes one or more additional sensors that detect / measure / capture a second type of attribute of the battery that is different from the first type of attribute.

[0040] For example, in one or more implementations, the first scanning compartment 102 includes one or more sensors that detect / measure / capture one or more visible attributes (e.g., first type attributes) of the battery. Specifically, the one or more sensors of the first scanning compartment 102 detect / measure / capture one or more of the size (e.g., volume, height, width) of the battery, the color of the battery, the shape of the battery, the text on the battery, etc. Similarly, in one or more implementations, the second scanning compartment 104 includes one or more sensors that detect / measure / capture one or more invisible attributes (e.g., second type attributes) of the battery. Specifically, the one or more sensors of the second scanning compartment 104 detect / measure / capture one or more of the temperature of the battery, the X-ray attenuation of the battery, the atomic number of one or more materials of the battery, the material composition of the battery, the weight of the battery, etc.

[0041] Also like Figures 1 to 3As shown, the automated battery sorting system 100 includes one or more platforms 124 to provide access to the scanning compartments 102 and 104 and other areas of the system for the screener or operator. In addition, the screener platform 124 shown provides access to the support cabinet 122, which houses various components of the scanning system for the compartments 102 and 104 (such as generators, IO feeds, analog controllers, and / or digital operator consoles).

[0042] In addition, some embodiments of the automated battery sorting system 100 include one or more infrared cameras positioned at one or more locations within the system to monitor the high temperature of the battery, and in some cases, include a safety mechanism. For example, the safety mechanism may include one or more of an alarm, a fire extinguisher, a battery immersion tank, or an emergency power off system (EPO). The safety mechanism is configured to be activated in response to a signal from at least one of a plurality of types of sensors. For example, the automated battery sorting system 100 includes a safety container (e.g., an immersion tank) for disposing of a battery with a high temperature when one or more infrared cameras detects it. The safety container includes a flame retardant material. Alternatively, the automated battery sorting system 100 is configured to add a flame retardant material to the safety container once the battery is added to the safety container. As used herein, the term "safety container" refers to a container for storing batteries and other potentially hazardous materials. In one or more embodiments, the safety container may be made of a durable and fireproof material such as metal. Additionally, the term "fire retardant material" or "fire suppressant" refers to a material used to prevent fires associated with volatile materials such as metals, flammable liquids, or lithium-ion batteries. For example, in some implementations, the fire suppressant may include mineral-based fire suppressants such as vermiculite, perlite, expanded clay, expanded polystyrene (EPS), CellBlockEX, and other fire, heat, and / or smoke suppressant compounds.

[0043] Moreover, in one or more embodiments, various components of the automated battery sorting system 100 are lined with a shock absorbing material (e.g., rubber) to reduce vibrations on the batteries as they pass through the system. Additionally, in some embodiments, the automated battery sorting system 100 includes an antistatic coating on its metal surfaces to reduce the risk of electrical shorting of the batteries.

[0044] As illustrated, after passing through the scanning bays 102 and 104, each battery continues along the conveyor 106 toward a plurality of sorting bins 108. The automated battery sorting system 100 shifts each target battery into one of the sorting bins 108 according to the predicted category for each respective battery. Specifically, the sorting bins 108 include a bin for each of the plurality of battery configurations. In addition, in one or more embodiments, one or more of the sorting bins 108 include a safety bin (e.g.,

[0045] In some embodiments, for example, the automated battery sorting system 100 utilizes a plurality of actuators or gates (ie, diverters) to move the target battery to an associated bin of the sorting bin 108. Figure 2 As shown, for example, some embodiments include a movable actuator 202 that is selectively positioned to push, pull, or guide incoming batteries into a sorting bin 108 according to a predicted category. Alternatively or additionally, embodiments of the automated battery sorting system 100 may include a sorting mechanism that includes a pneumatic actuator, a guide, a winch, a crane, a platform, or an alternative device for diverting the battery into a corresponding bin of the sorting bin 108. Furthermore, in some embodiments, the automated battery sorting system 100 includes one or more sensors configured to verify that the battery has been received by the correct sorting bin 108. For example, the verification sensor may include, but is not limited to, a camera or a pressure sensor integrated with the sorting bin 108.

[0046] Figures 1 to 3 An automated battery sorting system 100 is shown having two rows of sorting boxes 108. Specifically, Figures 1 to 3 A row of sorting boxes on each side of the conveyor 106 is illustrated. In alternative implementations, the automated battery sorting system 100 includes more or less than two rows of sorting boxes. For example, in one or more implementations, the automated battery sorting system 100 includes a single row of sorting boxes 108 on a single side of the conveyor 106. Alternatively, the automated battery sorting system 100 includes three or four rows of sorting boxes 108. For example, the automated battery sorting system 100 includes multiple rows of sorting boxes 108 located at different vertical heights on each side of the conveyor 106.

[0047] As also shown, the automated battery sorting system 100 includes return conveyors 118, 120, and 128 for returning batteries to the hopper 112 (or alternatively to the receiving bin 126 of the tipping mechanism 110) in the event that additional scanning and / or analysis is required. For example, in some cases, the battery classification may be substandard if the battery is not properly scanned for any reason, or if a previously unrecorded battery configuration is encountered. In at least some such cases, the battery is returned to the front of the automated battery sorting system 100 to repeat the scanning and sorting process. In some embodiments, batteries whose battery classification is uncertain or otherwise substandard are sorted into bins in the sorting bin 108 that are designated for alternative processing. In some implementations, unidentified batteries are allowed to pass through the end of the conveyor 106 to fall into one or more bins 130 positioned at the end of the conveyor 106.

[0048] Furthermore, in some embodiments, instead of (or in addition to) the sorting bins 108 located near the conveyor 106, the automated battery sorting system 100 may include an additional conveyor that is proximate to the main conveyor 106 and configured to deliver the batteries to a separate bin or another area for sorting and / or further processing. Thus, in some embodiments, the aforementioned actuators may transfer the batteries to the additional conveyor based on the predicted category for each respective battery.

[0049] like Figure 4 and FIG. 5A to FIG. 5B As shown, the scanning compartments 102 and 104 include one or more sensors for scanning the battery as it passes from the feeder 116 and through the conveyor 106. For example, the scanning compartment 102 includes a scanning bar 402, which includes one or more of a three-dimensional (3D) scanner, an RGB camera, or an infrared (IR) camera. In additional embodiments, the scanning compartment 102 includes additional sensors at additional locations to provide scanning of the battery at multiple angles and / or distances relative to each battery as each battery passes through the scanning compartment 102. For example, in some embodiments, multiple cameras are provided at multiple locations to capture images of each side of each battery. For example, in addition to being located in the scanning bar 402 or as an alternative to being located in the scanning bar 402, the RGB camera is located on the side of the conveyor so as to be able to capture a side view of the battery passing through the scanning compartment 102. Further, in some embodiments, additional sensors are provided to measure the weight of each battery for additional input into predicting the corresponding battery configuration.

[0050] In addition, if Figure 4As shown, the scanning compartment 104 includes an x-ray scanning array consisting of an x-ray emitter 404 and an x-ray detector 406. In one or more embodiments, for example, the x-ray emitter is mounted to the housing of the scanning compartment 104 and is directed to emit an x-ray beam toward the conveyor 106. Accordingly, in one or more embodiments, the x-ray detector 406 is mounted below the conveyor 106 so that when the battery passes through the scanning compartment 104, the x-ray detector 406 detects and measures the x-ray beam emitted from the x-ray emitter 404 and passes through each battery. In addition, in one or more embodiments, the x-ray detector 406 is enclosed in a retractable drawer 412, thereby facilitating adjustment and maintenance of the x-ray detector 406. In additional implementations, the scanning compartment 104 includes more than one x-ray scanning array. For example, the scanning compartment 104 includes two, three, or more x-ray detector arrays, each of which is aligned at a different angle relative to the battery positioned on the conveyor 106. Specifically, in addition to measuring the x-ray beams such as Figure 4 In addition to the x-ray attenuation shown for vertically passing through the battery, the scanning bay 104 also includes an x-ray scanning array that measures the attenuation of x-rays passing horizontally through the battery. In this implementation, the automated battery sorting system 100 includes an x-ray emitter located on one side of the conveyor 106 and an x-ray detector located on the opposite side of the conveyor 106. In any case, in view of the present disclosure, it will be understood that the scanning bays 102, 104 can include sensors on one or more sides of the conveyor 106, vertically above and / or below the conveyor 106, and at one or more acute angles (e.g., 30 degrees, 45 degrees, 60 degrees, 120 degrees, 135 degrees, 150 degrees) relative to the conveyor 106.

[0051] In addition, Figure 4 and FIG. 5A to FIG. 5B As shown, the support cabinet 122 includes a generator 408 for powering the x-ray scanning array and / or other components of the automated battery sorting system 100. In addition, the illustrated support cabinet includes an IO box 410 (i.e., input / output module) and a control assembly 502 (see Figure 5A ). Although the control assembly 502 is shown as an analog controller associated with an x-ray scanning array, embodiments of the automated battery sorting system 100 may also include an operator console capable of controlling sensors of both scanning bays 102 and 104 as well as other aspects of the battery sorting and sorting system of the automated battery sorting system 100.

[0052] like Figure 6As further shown, the automated battery sorting system 100 also includes a tipping bucket mechanism 110 having a receiving box 126 for receiving multiple batches of batteries for classification and sorting. In the case where a batch of batteries is loaded into the receiving box 126 of the tipping bucket mechanism 110, an actuator 606 of the tipping bucket mechanism 110 lifts the batch of batteries into the hopper 112 to disperse the batch of batteries toward a subsequent stage of the automated battery sorting system 100. In some embodiments, for example, the actuator 606 rotates the chamber of the tipping bucket mechanism 110 upward (i.e., dumps) until the batteries are dispersed from an opening 608 near the hopper 112. Additional examples of actuators for moving multiple batches of batteries from the receiving box 126 to the hopper 112 may include, for example, a conveyor belt, a movable platform, a drip trough (e.g., in the case where the receiving box 126 is positioned above the hopper 112), etc. Alternatively, the batteries may be loaded directly into the hopper 112 by hand, a conveyor, a forklift, or other loading mechanism.

[0053] As previously described, the hopper 112 vibrates the batteries to separate them and disperse them onto the inclined conveyor 114, directly into the feed chute 116, or otherwise dispersed toward the scanning area of ​​the automated battery sorting system 100. For example, in some embodiments, the hopper 112 is configured to disperse the batteries from the hopper 112 individually onto the inclined conveyor 114 via the conical outlet 604 (see Figures 1 to 3 ). As shown, the hopper 112 includes a vibrating table 602 configured to separate and direct the batteries toward a conical outlet 604. In some embodiments, the hopper 112 also includes gradually narrowing tracks or sidewalls that further direct the batteries toward the conical outlet 604 as the vibrating table 602 vibrates the batteries apart from each other. In some implementations, the batteries are directed through the conical chamber with or without the use of a vibrating table.

[0054] Also like Figure 6 , the hopper 112 may further include a metering gate 610. In one or more embodiments, for example, the metering gate 610 may include one or more sensors for detecting individual batteries as they pass through the metering gate 610 toward the tapered outlet 604. In some embodiments, a metering gate like the metering gate 610 is positioned in one or more alternative locations of the automated battery sorting system 100 to track the number of batteries passing through multiple stations of the automated battery sorting system 100.

[0055] As previously mentioned, the disclosed embodiments include a battery classification system that is configured to classify a battery based on a plurality of sensor signals (such as those described above with respect to Figures 1 to 6The automated battery sorting system 100 discussed above) intelligently sorts batteries of multiple types or configurations (i.e., multiple categories). For example, Figure 7 A schematic overview of a battery classification system 704 operating in accordance with one or more embodiments is illustrated.

[0056] like Figure 7 As shown, in some embodiments, a computing device 702 includes a battery classification system 704 that communicates with a plurality of sensors 710 and a battery sorting mechanism 720. In alternative embodiments, the battery classification system 704 is included on one or more server devices that communicate with the computing device 702, the plurality of sensors 710, and / or the battery sorting mechanism 720 via a network. In fact, the present disclosure contemplates a variety of configurations of computing devices, server devices, storage devices, sensors, and other components.

[0057] like Figure 7 As shown, the battery classification system 704 includes a classifier model 706 and a database 708, the database 708 containing a plurality of battery categories and their associated attributes (i.e., the chemical composition, form factor, and other attributes of the batteries within each respective category of the plurality of battery categories). Thus, the battery classification system 704 compares the battery category data contained in the database 708 and / or uses the battery category data contained in the database 708 to train the battery classification system 704 to make a battery category prediction based on signals from the plurality of sensors 710. As previously described, in one or more embodiments, the classifier model 706 includes a machine learning model that is trained to classify batteries based on a plurality of signals from the plurality of sensors 710.

[0058] like Figure 7 As shown, in some embodiments, the plurality of sensors 710 include an x-ray scanning array 712 (e.g., as discussed above with respect to the second scanning compartment 104, the x-ray emitter 404 and the x-ray detector 406), a three-dimensional (3D) scanner 714 (e.g., as discussed above with respect to the first scanning compartment 102 and the scanning bar 402), one or more RGB cameras 716 (e.g., as discussed above with respect to the first scanning compartment 102 and the scanning bar 402), and / or an infrared camera 718 (e.g., as discussed above with respect to the first scanning compartment 102 and the scanning bar 402). In fact, embodiments of the battery classification system 704 may include any or all of the plurality of sensors 710 shown, as well as additional sensors for scanning / imaging target batteries. In addition, although Figures 1 to 6Exemplary types, configurations, and locations of sensors for scanning batteries are illustrated, but alternative embodiments include multiple sensors 710 of different types, configurations, and locations. Further, in some embodiments, sensor data is received from an alternative source or indirectly from multiple sensors 710, such as within a data packet received with each target battery.

[0059] Also like Figure 7 As shown, in one or more embodiments, the battery sorting system 704 communicates with a battery sorting mechanism 720, which includes a plurality of actuators 722 (e.g., as described above with respect to the battery sorting system 704) for sorting the batteries into a plurality of sorting bins 724. Figures 1 to 6 For example, in response to predicting the battery category of the target battery, the battery classification system 704 indicates the predicted category to the battery sorting mechanism 720, and in some implementations, indicates the location of the target battery on the conveyor 106. In response, the battery sorting mechanism 720 uses one or more actuators (e.g., actuator 202) of the plurality of actuators 722 to move the target battery to a bin in the sorting bin 724 corresponding to the predicted battery category.

[0060] Additionally, in some embodiments, the battery classification system 704 is configured to detect anomalies within the battery and respond accordingly. Figure 7 As shown, the plurality of sensors 710 include an infrared camera 718 configured to monitor incoming batteries for high temperatures or other thermal anomalies. When a high temperature (or other undesirable thermal condition) above a certain threshold is detected, the battery sorting system 704 indicates the anomaly to the battery sorting mechanism 720, which in turn activates one or more alarms 726, stops the sorting process, and / or ejects the subject battery (e.g., relocates the subject battery to a safe container (e.g., an immersion tank or other enclosure)).

[0061] As described above, the battery classification system 704 can utilize a classifier model to predict the battery category based on signals from multiple sensors. For example, Figure 8 The battery classification system 704 is illustrated for determining a battery category 826 using a classifier machine learning model 802 according to one or more embodiments. Specifically, Figure 8The battery classification system 704 is shown receiving a plurality of signals from the sensor 816 and determining a battery category 826 from a plurality of battery categories stored in the database 814 using a classifier machine learning model 802. For example, the battery category may include various types of batteries, such as, but not limited to, chemical composition, size, intended use, commercial brand, model type / number, etc. The chemical composition of the battery may include, for example, lithium, lithium ion, aluminum ion, magnesium ion, sodium ion, potassium ion, alkaline, nickel, carbon zinc, silver oxide, aluminum air, zinc air, zinc carbon, zinc chloride, zinc ion, lead acid, or any type of battery composition for which the classification is received by the battery classification system 704.

[0062] like Figure 8 As shown, the plurality of sensors 816 include a three-dimensional (3D) scanner, an x-ray scanning array 820, one or more RGB cameras 822, and an infrared (IR) camera 824. The sensors 816 capture / measure signals that are transferred to the battery classification system 704. The battery classification system 704 determines one or more attributes of the battery from the signals. The classifier machine learning model 802 determines the battery category based on the one or more attributes.

[0063] In the illustrated embodiment, for example, the battery classification system 704 receives an image from a 3D scanner 818. Based on the image, the battery classification system 704 determines one or more physical properties of the battery in the image. For example, the battery classification system 704 determines one or more dimensions 804 of the target battery based on the signal (e.g., data) received from the 3D scanner 818. Specifically, the battery classification system 704 determines one or more of the height, width, length, or volume of the target battery. The classifier machine learning model 802 generates a battery category 826 based at least in part on the dimensions 804.

[0064] Alternatively, in one or more embodiments, for example, the battery classification system 704 manipulates the signals received from the 3D scanner 818 to determine or generate properties of the target battery. For example, the battery classification system 704 generates a three-dimensional profile of the target battery. Specifically, the battery classification system 704 uses one or more images or 3D scans to reconstruct the three-dimensional shape or profile of the target battery. In other words, the battery classification system 704 determines the shape of the target battery based on the signals from the sensor 816. The classifier machine learning model 802 generates a battery category 826 based at least in part on the shape or profile of the target battery.

[0065] In addition, the battery classification system 704 determines one or more measurements of x-ray attenuation 806 based on the signal of the target battery received from the x-ray scanning array 820. In one or more embodiments, for example, the x-ray scanning array 820 writes the x-ray data associated with the target battery to a cache memory in one or more data subsets (i.e., chunks). Thus, the battery classification system 704 can access one or more data subsets of the target battery, and in the case where multiple subsets correspond to a single battery, multiple subsets are connected in series for subsequent processing. In some embodiments, the battery classification system 704 receives signals from the x-ray scanning array 820, monitors the temperature of one or more scintillators of the x-ray scanning array 820, and recalibrates the x-ray gain lookup table when necessary based on the static offset and gain rate of each pixel, as the static offset and gain rate of each pixel changes in real time (e.g., in response to temperature). In one or more embodiments, the battery classification system 704 determines the x-ray attenuation 806 using the x-ray gain lookup table to normalize the signal from the detector of the x-ray scanning array 820 to obtain an attenuation measurement adjusted according to the above real-time update. Further, in some embodiments, the battery classification system 704 utilizes a time decay model to reduce the effect of afterglow from the scintillator present in any given signal from the x-ray scanning array 820 .

[0066] In addition, in some embodiments, the battery classification system 704 aligns high and low energy pixels based on the signals of the x-ray scanning array 820 to generate high / low energy pairs for determining what portion of the received pixels falls into a defined area associated with a given chemical composition of the battery (e.g., as indicated in the database 814). Moreover, in one or more embodiments, the battery classification system 704 aligns a segmented 3D scan of the target battery (e.g., received from the 3D scanner 818) with the corresponding low and high energy scans received from the x-ray scanning array 820 to utilize the distribution and ratio of battery height to x-ray attenuation amplitude in determining the chemical composition or type of the target battery. Thus, the battery classification system 704 described herein can store the inferred predictions and / or confidence levels in memory or transmit such results directly to the classifier machine learning model 802 for further analysis, or to the battery sorting mechanism 828 for classifying the target battery based on the inferred predictions and / or confidence levels.

[0067] In one or more embodiments, the battery classification system 704 receives x-ray attenuation values ​​806 from the x-ray scanning array 820. Using these signals, the battery classification system 704 determines the properties of the target battery. For example, the battery classification system 704 determines the battery chemistry or material composition of the battery based on the attenuation values ​​806. The classifier machine learning model 802 generates a battery category 826 based at least in part on the battery chemistry or material composition of the target battery.

[0068] More specifically, the battery classification system 704 receives the x-ray attenuation value 806 in the form of a high energy attenuation value and a low energy attenuation value of the target battery. In one or more embodiments, the battery classification system 704 generates an attenuation energy curve by plotting the high energy attenuation value versus the low energy attenuation value. The battery classification system 704 determines the battery chemistry based on the attenuation energy curve. For example, the battery classification system 704 stores known attenuation energy curves of different battery chemistries in a database 814. The battery classification system 704 maps the generated attenuation energy curve to the known attenuation energy curve to determine the battery chemistry of the target battery.

[0069] In addition, the battery classification system 704 receives one or more images (e.g., signals) of the target battery from the RGB camera 822. The battery classification system 704 generates one or more attributes based on the one or more images. For example, the battery classification system 704 uses character recognition 808 and object recognition 810 to determine additional attributes of the target battery. In one or more embodiments, for example, the RGB camera 822 includes one or more overhead cameras that are mounted / positioned to view each target battery from each side (e.g., top, front, back, side) of each target battery. The battery classification system 704 can use a camera driver to segment the battery from the RGB image using background subtraction techniques, store the segmented image in a cache memory (e.g., a Redis cache including an in-memory key-value memory for storing and retrieving images and data), and associate the coordinates and / or positions of the corresponding battery with the stored image. Thus, the battery classification system 704 can access the segmented image from the memory and perform object and / or character recognition on the associated battery. For example, the battery classification system 704 determines the printed characters or codes on the label of the target battery. The classifier machine learning model 802 generates a battery category 826 based at least in part on the printed characters or encoding of the target battery.

[0070] In one or more embodiments, the battery classification system 704 utilizes a classifier machine learning model 802 to compare words, characters, and other recognized data to different battery types within the database 814. In some embodiments, the classifier machine learning model 802 includes one or more deep learning models configured to perform object recognition to identify the specific brand and / or model of each target battery.

[0071] Additionally, in some embodiments, the battery sorting system 704 receives one or more image frames from the RGB camera 822, segments each image frame to locate one or more individual batteries within the frame, links image frames that include the same battery (e.g., utilizing the shape or outline of the battery), performs character recognition 808 and / or object recognition 810, and associates the resulting object / character output with a battery ID and / or a physical location of the battery (e.g., an xy location on a conveyor belt of a battery sorting device).

[0072] Further, the classifier machine learning model 802 receives a signal corresponding to the temperature 812 and / or thermal profile of the target battery from the infrared camera 824. As previously mentioned, in some embodiments, the battery classification system 704 may activate one or more safety measures in response to detecting that the battery exhibits a high temperature (e.g., upon determining that the temperature of the target battery has reached a threshold temperature) or other thermal anomaly via a signal from the infrared camera 824. For example, in response to detecting a thermal anomaly in the target battery, the battery classification system 704 may activate an alarm and / or a fire extinguisher (or flame retardant), evacuate the target battery to an immersion tank or other disposal, shut down the battery sorting equipment, and / or notify / warn a user of the anomaly.

[0073] Therefore, if Figure 8 As shown, the classifier machine learning model 802 determines the battery category 826 of the target battery based on one or more of the size 804, x-ray attenuation 806, character recognition 808, object recognition 810, or temperature 812 indicated by the signal received from the sensor 816. For example, in some embodiments, the classifier machine learning model 802 aggregates the inferences generated from the signals of each sensor 816 (as described above) to generate the final battery category. For example, in one or more implementations, the classifier machine learning model 802 generates the battery category 826 based on the attributes of the target battery determined according to the signal from the sensor 816 using a decision tree. Alternatively, the classifier machine learning model 802 includes a classification neural network. In this implementation, the classification neural network concatenates the attribute values ​​of the target battery. The classification neural network uses a first set of neural network layers (e.g., encoders) to generate feature vectors from the concatenated attribute values. The classification neural network generates the battery category 826 based on the feature vector using a second set of neural network layers (e.g., decoders).

[0074] Additionally, in one or more embodiments, if the classifier machine learning model 802 cannot determine the battery class within a threshold confidence level, the battery classification system 704 issues an error signal to the sorting mechanism 828. In this case, for example, the sorting mechanism 828 may reroute the target battery to repeat the classification attempt or place the battery into a bin designated for batteries with uncertain class.

[0075] In addition, in some embodiments, the battery classification system 704 indicates the battery category 826 to a sorting mechanism 828 for sorting the target battery. In one or more embodiments, for example, the battery classification system 704 provides the location of the classified battery, such as coordinates associated with a conveyor belt, and the sorting mechanism 828 uses the transmitted location to notify the sorting actuator to move the battery to a corresponding sorting box, or as described above, to a corresponding additional conveyor. In some embodiments, the classifier machine learning model 802 includes a machine learning model trained to predict the battery category, as described herein. Alternatively, some embodiments include a combination of a classical (e.g., mathematical) model and one or more machine learning models.

[0076] As described above, in one or more embodiments, the battery classification system 704 uses images from one or more RGB cameras as input to determine the battery category of the target battery. For example, Fig. 9 A battery classification system 704 is illustrated that utilizes images from one or more RGB cameras 904 to determine a battery type inference 920 for a target battery 902 .

[0077] For example, the one or more RGB cameras 904 provide one or more images 908 including a label on the target battery 902, and the classifier model 912 performs character recognition 914 (e.g., OCR) from the one or more images 908 and compares the recognized characters to a database 918. For example, the characters recognized by the character recognition 914 can be compared to characters in the database 918, or the characters recognized by the character recognition 914 can be used to look up model numbers, serial numbers, and other relevant information that may be useful in determining a battery type inference 920 for the target battery 902.

[0078] Additionally or alternatively, the one or more RGB cameras 904 provide one or more silhouette images 906 of the target battery 902, the one or more silhouette images 906 indicating or including a form factor 910 of the target battery 902. In response, the classifier model 912 utilizes the object detection machine learning model 916 to compare the visual attributes of the target battery 902 with data stored in the database 918 to determine a battery type inference 920 for the target battery 902. Furthermore, the classifier model 912 can be trained to recognize known form factors or other battery attributes based on the form factors and attributes of battery configurations found in the database 918.

[0079] For example, in one or more embodiments, the classifier model 912 includes an object detection machine learning model 916 to detect classified batteries within a digital image. In one or more implementations, the object detection machine learning model 916 includes a deep learning convolutional neural network (CNN). For example, in some embodiments, the object detection machine learning model 916 includes a region-based (R-CNN). Specifically, the object detection machine learning model 916 includes a lower neural network layer and a higher neural network layer. In general, the lower neural network layer forms an encoder together, and the higher neural network layer forms a decoder together. In one or more embodiments, the encoder includes a convolutional layer that encodes the digital image into a feature vector, which is output from the encoder and provided as an input to the decoder. In multiple implementations, the decoder includes a fully connected layer that analyzes the feature vector and outputs a battery category. In one or more embodiments, the object detection machine learning model 916 provides a prediction that the battery in the image is each of a plurality of battery categories. For example, the object detection machine learning model 916 generates a category vector with a prediction (e.g., a number between 0 and 1) indicating that the battery in the image is a predicted percentage (e.g., 1% to 100%) of each of a plurality of battery categories. Thus, if there are 100 different battery classes, the object detection machine learning model 916 may generate a class vector with 100 entries. The classifier model 912 selects the battery type inference 920 as the battery class with the highest prediction percentage.

[0080] As described above, some embodiments of the battery classification system 704 utilize an x-ray scanning array and / or a 3D scanner to determine one or more attributes of a target battery (such as battery chemistry) in order to predict the battery category. Fig.10 A battery classification system 704 is shown that utilizes input from a 3D scanner and an x-ray scanning array 1006 to determine a battery chemistry inference 1018 for a target battery 1002 .

[0081] For example, in the illustrated embodiment, the battery classification system 704 receives or infers dimensions 1008 from the 3D scanner 1004 and receives or infers high energy attenuation 1010 and low energy attenuation 1012 measurements from the x-ray scan array 1006. In response, the battery classification system 704 determines a battery chemistry inference 1018 using an attenuation model 1016 of the classifier model 1014. In one or more embodiments, the x-ray scan array 1006 includes a differential x-ray scan array that includes multiple x-ray scan arrays positioned at different angles or positions relative to the target battery 1002.

[0082] In one or more embodiments, when determining the battery chemistry inference 1018, the battery classification system 704 combines the signals of the target battery 1002 from the x-ray scanning array 1006 and the 3D scanner 1004. When segmenting the target battery 1002, for example, the battery classification system 704 can use the image provided by the 3D scanner 1004 to determine the size 1008 of the target battery 1002, including the area, the centroid, and / or one or more rectangular crops. In one or more embodiments, the battery classification system 704 uses the size 1008 and the signal from the x-ray scanning array 1006 to generate an alignment (e.g., a hash map) of the high energy decay 1010, the low energy decay 1012, and the size 1008 of the target battery 1002. In such an embodiment, the battery classification system 704 may utilize the decay model 1016 of the classifier model 1014 to generate a battery chemistry inference 1018 based on the alignment of the size 1008 , the high energy decay 1010 , and the low energy decay 1012 of the target battery 1002 .

[0083] Furthermore, as discussed above, embodiments of the battery classification system 704 utilize various signals from various sensors to determine the battery class of a target battery. For example, in some embodiments, the battery classification system 704 utilizes Fig. 9 The battery type is inferred by 920 and Fig.10 The battery chemistry inference 1018 of the target battery is used to determine the battery category of the target battery. In fact, embodiments of the battery classification system 704 can use any combination of data from different sensors in conjunction with a classifier model to predict the battery category of the target battery.

[0084] Figures 1 to 10 , corresponding text and examples provide a number of different methods, systems, apparatuses, and non-transitory computer-readable media of the battery classification system 704. In addition to the foregoing, one or more embodiments may also be described according to a flowchart that includes actions for achieving a particular result, such as Fig.11 shown. Fig.11More or fewer actions may be used to perform. Further, these actions may be performed in different orders. In addition, the actions described herein may be repeated or performed in parallel with each other or with different examples of the same or similar actions.

[0085] As mentioned above, Fig.11 1100 is a flow chart illustrating a series of actions 1100 for classifying and sorting batteries according to one or more embodiments. Fig.11 The actions according to one embodiment are illustrated, but alternative embodiments may omit, add, reorder, and / or modify Fig.11 Any action shown in . Fig.11 The actions of may be performed as part of a method. Alternatively, a non-transitory computer readable medium may include instructions that, when executed by one or more processors, cause the one or more processors to perform Fig.11 In some embodiments, the system may perform Fig.11 action.

[0086] As shown in the figure, Fig.11 A series of example actions 1100 for classifying and sorting batteries according to multiple battery categories is illustrated. The series of actions 1100 may include an action 1102 for scanning a target battery using multiple sensors. For example, in some embodiments, action 1102 includes receiving multiple signals corresponding to the target battery from multiple sensors. In one or more embodiments, action 1102 may include receiving multiple signals corresponding to one or more detected attributes of the target battery from multiple sensors, the one or more detected attributes including one or more of the size, battery chemistry, printed characters, or form factor of the target battery. In addition, in some embodiments, the multiple sensors include two or more of an x-ray scanning array, a three-dimensional (3D) scanner, an RGB camera, or an infrared camera.

[0087] In addition, if Fig.11 As shown, the series of actions 1100 may include an action 1104 for determining a battery category using a classifier machine learning model. For example, in some embodiments, action 1104 includes: using a classifier machine learning model to determine a predicted battery category of a target battery from a plurality of battery categories based on a plurality of signals. In one or more embodiments, action 1104 may include: using a classifier machine learning model to determine a predicted battery category of a target battery from a plurality of battery categories based on one or more detected attributes of the target battery.

[0088] Additionally, in some embodiments, action 1104 may include: determining a material property (e.g., battery chemistry) of the target battery based on one or more x-ray attenuation measurements of the target battery received from the x-ray scanning array; and determining a predicted battery category of the target battery based on the material property. For example, in some embodiments, action 1104 may include: receiving x-ray attenuation data of the target battery from an x-ray scanning array of multiple sensors, and determining a predicted battery category of the target battery based on the x-ray attenuation data.

[0089] In one or more embodiments, action 1104 may include: determining the size of the target battery based on the scan data of the target battery received from a three-dimensional (3D) scanner, and determining the predicted battery category of the target battery based on the sizes. Relatedly, in one or more embodiments, action 1104 may include: receiving scan data of the target battery from a three-dimensional (3D) scanner of multiple sensors; determining multiple dimensions of the target battery based on the scan data; and determining the predicted battery category of the target battery based on the multiple dimensions.

[0090] In some embodiments, action 1104 may include: determining multiple printed characters or codes arranged on the target battery based on the image data of the target battery received from the RGB camera; and determining the predicted battery category of the target battery based on the multiple printed characters or codes. Relatedly, in some embodiments, action 1104 may include: receiving one or more label images from the RGB camera in the multiple sensors; using object character recognition (OCR), identifying multiple printed characters or codes from one or more label images; and determining the predicted battery category of the target battery based on the multiple printed characters or codes.

[0091] In addition, in one or more embodiments, action 1104 includes: receiving one or more label images and one or more contour images of the target battery from an RGB camera of the plurality of sensors; determining the printed characters of the target battery according to the one or more label images using optical character recognition; determining the outer shape of the target battery according to the one or more contour images; and determining the predicted battery category of the target battery based on the printed characters and the outer shape using a classifier machine learning model. Moreover, in some embodiments, action 1104 may include: receiving one or more contour images of the target battery from the RGB cameras of the plurality of sensors; determining the outer shape of the target battery based on the one or more contour images of the target battery; and determining the predicted battery category of the target battery based on the outer shape.

[0092] Additionally, in some embodiments, action 1104 may include: determining a size of a target battery based on scan data from a 3D scanner of multiple sensors; determining a battery chemistry of the target battery based on x-ray attenuation data from an x-ray scanning array of multiple sensors; and determining a predicted battery category of the target battery based on size and battery chemistry using a classifier machine learning model.

[0093] Also like Fig.11 As shown, a series of actions 1100 may include an action 1106 for indicating the battery category to the battery sorting mechanism. For example, in some embodiments, action 1106 includes: indicating the predicted battery category to the battery sorting mechanism. Moreover, in some embodiments, action 1106 includes: using the battery sorting mechanism, transferring the target battery to a box corresponding to the predicted battery category of the target battery.

[0094] In addition, in some embodiments, action 1106 may include: indicating the predicted battery category of the target battery and the additional predicted battery category of the additional battery to a battery sorting mechanism, and using the battery sorting mechanism, individually transferring each of the target battery and the additional battery to multiple boxes respectively associated with multiple battery categories.

[0095] In addition, in one or more embodiments, the series of actions 1100 may include: receiving an additional plurality of signals corresponding to the additional battery from the plurality of sensors; and determining, using a classifier machine learning model, an additional predicted battery category of the additional battery from the plurality of battery categories based on the additional plurality of signals. Moreover, in some embodiments, the series of actions 1100 may include: using a battery sorting mechanism, transferring the target battery to a first bin corresponding to the predicted battery category of the target battery, determining the predicted battery category of the additional target battery based on the additional plurality of signals from the plurality of sensors, and using the battery sorting mechanism, transferring the additional target battery to a second bin corresponding to the predicted battery category of the additional target battery.

[0096] Furthermore, in some embodiments, the series of actions 1100 may include: determining a temperature or temperature gradient of the target battery based on a signal received from an infrared camera for the target battery, and activating an alarm in response to determining that the temperature or temperature gradient is above a threshold. Furthermore, in one or more embodiments, the series of actions 1100 may include: receiving an indication of a dangerous anomaly detected in at least one of the additional batteries from an infrared camera in the plurality of sensors, and in response, transferring the at least one battery to an immersion tank containing a flame retardant.

[0097] Embodiments of the present disclosure may include or utilize a special-purpose or general-purpose computer including computer hardware, such as one or more processors and system memory, as discussed in more detail below. Embodiments within the scope of the present disclosure also include physical and other computer-readable media for carrying or storing computer-executable instructions and / or data structures. Specifically, one or more processes described herein may be at least partially implemented as instructions contained in a non-transitory computer-readable medium and executable by one or more computing devices (e.g., any media content access device described herein). In general, a processor (e.g., a microprocessor) receives instructions from a non-transitory computer-readable medium (e.g., a memory) and executes those instructions, thereby performing one or more of the above processes, including one or more of the above processes described herein.

[0098] Computer-readable media can be any available media that can be accessed by a general or special purpose computer system. A computer-readable medium that stores computer-executable instructions is a non-transitory computer-readable storage medium (device). A computer-readable medium that carries computer-executable instructions is a transmission medium. Therefore, by way of example and not limitation, embodiments of the present disclosure may include at least two distinct computer-readable media: a non-transitory computer-readable storage medium (device) and a transmission medium.

[0099] Non-transitory computer-readable storage media (devices) include RAM, ROM, EEPROM, CD-ROM, solid-state drives (“SSD”) (e.g., RAM-based), flash memory, phase change memory (“PCM”), other types of memory, other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired program code devices in the form of computer-executable instructions or data structures and that can be accessed by a general or special purpose computer.

[0100] "Network" is defined as one or more data links that enable electronic data to be transmitted between computer systems and / or modules and / or other electronic devices. When information is transmitted or provided to a computer via a network or another communication connection (hardwired, wireless, or a combination of hardwired or wireless), the computer appropriately regards the connection as a transmission medium. Transmission media may include networks and / or data links that can be used to carry computer executable instructions or desired program code devices in the form of data structures and can be accessed by general or special computers. The above combinations should also be included in the scope of computer-readable media.

[0101] Further, upon reaching multiple computer system components, program code means in the form of computer executable instructions or data structures may be automatically transferred from a transmission medium to a non-transitory computer readable storage medium (means) (or vice versa). For example, computer executable instructions or data structures received over a network or data link may be cached in RAM within a network interface module (e.g., "NIC") and ultimately transferred to computer system RAM and / or a less volatile computer storage medium (means) at the computer system. Thus, it should be understood that a non-transitory computer readable storage medium (means) may be included in a computer system component that also (or even primarily) utilizes a transmission medium.

[0102] Computer executable instructions include, for example, instructions and data, which, when executed by a processor, cause a general-purpose computer, a special-purpose computer, or a special-purpose processing device to perform a certain function or a group of functions. In some embodiments, computer executable instructions are executed by a general-purpose computer to turn a general-purpose computer into a special-purpose computer that implements the elements of the present invention. Computer executable instructions can be, for example, binary, intermediate format instructions such as assembly language, or even source coding. Although the subject matter has been described in a language dedicated to structural features and / or method actions, it is understood that the subject matter defined in the appended claims is not necessarily limited to the above-mentioned features or actions. On the contrary, the described features and actions are disclosed as an example form of implementing the claims.

[0103] Those skilled in the art will recognize that the present disclosure can be implemented in a network computing environment with many types of computer system configurations, including personal computers, desktop computers, laptop computers, message processors, handheld devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, mobile phones, PDAs, tablet computers, pagers, routers, switches, etc. The present disclosure can also be practiced in a distributed system environment, where local and remote computer systems linked by a network (by a hardwired data link, a wireless data link, or a combination of a hardwired and wireless data link) all perform tasks. In a distributed system environment, program modules can be located in both local and remote memory storage devices.

[0104] Embodiments of the present disclosure may also be implemented in a cloud computing environment. As used herein, the term "cloud computing" refers to a model for implementing on-demand network access to a shared pool of configurable computing resources. For example, cloud computing can be used in a market to provide ubiquitous and convenient on-demand access to a shared pool of configurable computing resources. The shared pool of configurable computing resources can be quickly provisioned via virtualization technology and released with low management costs or service provider interactions and expanded accordingly.

[0105] The cloud computing model can consist of a variety of features, such as on-demand self-service, broad network access, resource pools, rapid elasticity, measured services, etc. The cloud computing model can also expose a variety of service models, such as software as a service ("SaaS"), platform as a service ("PaaS"), and infrastructure as a service ("IaaS"). The cloud computing model can also be deployed using different deployment models, such as private cloud, community cloud, public cloud, hybrid cloud, etc. In addition, as used herein, the term "cloud computing environment" refers to an environment that employs cloud computing.

[0106] Fig.12 A block diagram of an example computing device 1200 that can be configured to perform one or more of the above-described processes is shown. It will be appreciated that one or more computing devices such as computing device 1200 may represent the above-described computing devices (e.g., computing device 702). In one or more embodiments, computing device 1200 may be a mobile device (e.g., a mobile phone, a smart phone, a PDA, a tablet computer, a laptop computer, a camera, a tracker, a watch, a wearable device, etc.). In some embodiments, computing device 1200 may be a non-mobile device (e.g., a desktop computer or another type of client device). Further, computing device 1200 may be a server device (e.g., a local server that processes custom TCP or HTTP messages) that includes cloud-based processing and storage capabilities.

[0107] like Fig.12 As shown, computing device 1200 may include one or more processors 1202, memory 1204, storage device 1206, input / output interface 1208 (or "I / O interface 1208"), and communication interface 1210, which may be communicatively coupled via a communication infrastructure (e.g., bus 1212). Fig.12 The computing device 1200 is shown in FIG. Fig.12 The components shown in FIG. 1 are not intended to be limiting. In other embodiments, additional or alternative components may be used. Furthermore, in some embodiments, computing device 1200 includes more than Fig.12 The components shown in FIG. are fewer than those shown in FIG. Fig.12 Components of computing device 1200 shown in .

[0108] In a particular embodiment, the processor 1202 includes hardware for executing instructions, such as those constituting a computer program. As an example and not by way of limitation, to execute instructions, the processor 1202 may retrieve (or fetch) instructions from an internal register, an internal cache, memory 1204, or storage 1206 and decode and execute them.

[0109] The computing device 1200 includes a memory 1204 coupled to the processor 1202. The memory 1204 may be used to store data, metadata, and programs for execution by the processor. The memory 1204 may include one or more of volatile and non-volatile memories, such as random access memory ("RAM"), read-only memory ("ROM"), solid state disk ("SSD"), flash memory, phase change memory ("PCM"), or other types of data storage devices. The memory 1204 may be internal or distributed memory.

[0110] The computing device 1200 includes a storage device 1206, which includes a storage device for storing data or instructions. As an example and not by way of limitation, the storage device 1206 may include the non-volatile storage media described above. The storage device 1206 may include a hard disk drive (HDD), a flash memory, a universal serial bus (USB) drive, or a combination of these or other storage devices.

[0111] In some embodiments, the data storage device within the storage device 1206 may include a remote dictionary server (Redis) data structure that includes an in-memory key-value database for storing and indexing cached data throughout the battery classification and sorting process described herein. In certain embodiments, for example, each sensor array in a plurality of sensor arrays (e.g., RGB cameras, 3D scanners, x-ray scanning arrays) stores images and / or related information with time-related key values ​​representing the xy coordinates of the corresponding image (e.g., "rgb:390:8910"). Thus, when aggregating sensor signals to determine the battery category of a target battery, the battery classification system 704 may associate information from each sensor array pipeline with corresponding physical coordinates. Alternatively or additionally, the battery classification system 704 may store a sorted list of coordinates of the corresponding images so that all images within a given coordinate range can be easily accessed when filtering signals corresponding to the target battery.

[0112] As shown, computing device 1200 includes one or more I / O interfaces 1208 that are provided to allow a user to provide input (such as user strokes) to computing device 1200, receive output from computing device 1200, and otherwise transfer data to computing device 1200. These I / O interfaces 1208 may include a mouse, a keypad or keyboard, a touch screen, a camera, an optical scanner, a network interface, a modem, other known I / O devices, or a combination of these I / O interfaces 1208. A touch screen may be activated with a stylus or a finger.

[0113] The I / O interface 1208 may include one or more devices for presenting output to a user, including, but not limited to, a graphics engine, a display (e.g., a display screen), one or more output drivers (e.g., a display driver), one or more audio speakers, and one or more audio drivers. In some embodiments, the I / O interface 1208 is configured to provide graphical data to the display for presentation to the user. The graphical data may represent one or more graphical user interfaces and / or any other graphical content that may serve a particular implementation.

[0114] The computing device 1200 may further include a communication interface 1210. The communication interface 1210 may include hardware, software, or both. The communication interface 1210 provides one or more interfaces for communication (such as, packet-based communication) between the computing device and one or more other computing devices or one or more networks. As an example and not a limiting method, the communication interface 1210 may include a network interface controller (NIC) or network adapter for communicating with an Ethernet or other wired network, or a wireless NIC (WNIC) or wireless adapter for communicating with a wireless network (such as WI-FI). The computing device 1200 may further include a bus 1212. The bus 1212 may include hardware, software, or both that connect the components of the computing device 1200 to each other.

[0115] In the foregoing description, the present invention has been described with reference to specific exemplary embodiments thereof. Different embodiments and aspects of the present invention are described with reference to the details discussed herein, and the accompanying drawings show different embodiments. The above description and the accompanying drawings are illustrative of the present invention and should not be construed as limiting the present invention. Many specific details are described to provide a thorough understanding of the various embodiments of the present invention.

[0116] The present invention may be embodied in other specific forms without departing from its spirit or essential characteristics. The described embodiments are considered to be illustrative and not restrictive in all respects. For example, the method described herein may be performed with fewer or more steps / actions, or the steps / actions may be performed in different orders. In addition, the steps / actions described herein may be repeated or performed in parallel with each other, or repeated or performed in parallel with different instances of the same or similar steps / actions. Therefore, the scope of the present invention is indicated by the appended claims, rather than by the preceding description. All changes within the meaning and scope of the equivalents of the claims are included within their scope.

Claims

1. A device for sorting batteries, the device comprising: a feed mechanism configured to release cells from the plurality of cells individually onto a conveyor; a scanner mechanism disposed about the conveyor, the scanner mechanism including a plurality of types of sensors configured to capture one or more attributes of each of the plurality of batteries; as well as An array of sorting mechanisms is arranged about the conveyor and is configured to transfer the plurality of batteries from the conveyor to a plurality of bins based on a predicted battery configuration of each of the plurality of batteries, the predicted battery configuration being determined based on the one or more attributes of each battery.

2. The device according to claim 1, wherein: The feed mechanism includes a hopper positioned above the conveyor and configured to direct individual cells of the plurality of cells onto the conveyor.

3. The apparatus of claim 1, further comprising one or more scanning bays configured to receive, via the conveyor, individual batteries of the plurality of batteries to be scanned by the plurality of types of sensors.

4. The device according to claim 1, wherein: The multiple types of sensors include two or more of an x-ray scanning array, a three-dimensional (3D) scanner, an RGB camera, or an infrared camera.

5. The device of claim 1, further comprising one or more processors configured to determine the predicted battery configuration for each battery based on sensor data from one or more sensors of the plurality of types of sensors.

6. The apparatus according to claim 1, further comprising a return mechanism disposed at an end of the conveyor and configured to return unsorted batteries to a starting point of the conveyor.

7. The device of claim 1 further comprising a safety mechanism comprising one or more of an alarm, a fire extinguisher, a battery immersion tank, or an emergency power off system (EPO), the safety mechanism being configured to activate in response to a signal from at least one of the plurality of types of sensors.

8. A device for sorting batteries, the device comprising: one or more scanning bays configured to receive batteries of a plurality of battery configurations; one or more scanner mechanisms associated with the one or more scanning bays, the one or more scanner mechanisms comprising a sensor configured to scan each battery as it passes through the one or more scanning bays; as well as One or more processors configured to determine a predicted battery configuration for each battery based on sensor data from the sensors of the one or more scanner mechanisms.

9. The apparatus of claim 8, further comprising a hopper mechanism configured to release the batteries individually onto a conveyor.

10. The device according to claim 9, wherein: The one or more scanning bays are arranged around the conveyor to receive batteries individually released by the hopper mechanism.

11. The apparatus according to claim 10 further comprises a sorting mechanism array arranged around the conveyor after the one or more scanning compartments, the sorting mechanism array being configured to transfer each battery from the conveyor to a corresponding box among the plurality of boxes based on the predicted battery configuration of each battery.

12. The device according to claim 8, wherein: The one or more scanner mechanisms include an x-ray scanning array configured to determine material characteristics of each battery.

13. The apparatus according to claim 8, wherein: The one or more scanner mechanisms include at least one of a three-dimensional (3D) scanner or an RGB camera configured to determine the dimensions of each battery.

14. The apparatus according to claim 8, wherein: The one or more processors are configured to determine the predicted battery configuration for each battery by processing the sensor data from the sensors of the one or more scanner mechanisms using a machine learning model.

15. A device for sorting batteries of different configurations, the device comprising: a feeding mechanism configured to release batteries from the plurality of batteries individually onto a conveyor; one or more scanner mechanisms configured to scan each battery as it passes on the conveyor; as well as One or more sorting mechanisms disposed about the conveyor and configured to transfer batteries of the plurality of batteries from the conveyor to a plurality of bins based on a predicted battery configuration of each battery determined based on attributes of each battery captured using the one or more scanner mechanisms.

16. The device according to claim 15, wherein: The one or more scanner mechanisms include an x-ray scanning array including an x-ray generator positioned above the conveyor and an x-ray detector positioned below the conveyor.

17. The apparatus of claim 15, the one or more scanner mechanisms comprising two or more of an x-ray scanning array, a three-dimensional (3D) scanner, an RGB camera, or an infrared camera.

18. The apparatus of claim 15, further comprising a return mechanism, the return mechanism being subsequent to the one or more scanner mechanisms, the return mechanism being configured to return a battery for which an accurate predicted battery configuration cannot be determined to the one or more scanner mechanisms.

19. The apparatus according to claim 15, wherein: The one or more sorting mechanisms include one or more actuators configured to push, pull, or guide batteries into corresponding bins of the plurality of bins.

20. The apparatus of claim 15, further comprising a safety mechanism configured to activate in response to detecting a dangerous anomaly within a battery of the plurality of batteries.

21. A computer-implemented method comprising: receiving a plurality of signals corresponding to the target battery from the plurality of sensors; Determining a predicted battery class of the target battery from a plurality of battery classes based on the plurality of signals using a classifier machine learning model; as well as The predicted battery category of the target battery is indicated to a battery sorting mechanism.

22. The computer-implemented method of claim 21, wherein: The plurality of sensors include two or more of an x-ray scanning array, a three-dimensional (3D) scanner, an RGB camera, or an infrared camera.

23. The computer-implemented method of claim 21: Also included is determining a material property of the target cell based on one or more x-ray attenuation measurements of the target cell received from an x-ray scanning array; in, Determining, using a classifier machine learning model, a predicted battery class of the target battery from a plurality of battery classes based on the plurality of signals includes determining the predicted battery class of the target battery based at least in part on the material properties.

24. The computer-implemented method of claim 23, further comprising: capturing a first set of x-ray attenuation measurements using a first x-ray scanning array oriented in a first orientation relative to the target cell; as well as capturing a second set of x-ray attenuation measurements using a second x-ray scanning array oriented in a second orientation relative to the target cell; Wherein determining the material property of the target battery based on one or more x-ray attenuation measurements includes: determining the material property based on the first set of x-ray attenuation measurements and the second set of x-ray attenuation measurements.

25. The computer-implemented method of claim 21: Also included is determining a size of the target battery based on scan data of the target battery received from a three-dimensional (3D) scanner; in, Determining, using a classifier machine learning model, a predicted battery class of the target battery from a plurality of battery classes based on the plurality of signals includes determining the predicted battery class of the target battery based at least in part on the size.

26. The computer-implemented method of claim 21: Also included is determining a plurality of printed characters or codes disposed on the target battery based on image data of the target battery received from one or more RGB cameras; in, Determining a predicted battery class of the target battery from a plurality of battery classes based on the plurality of signals using a classifier machine learning model includes determining the predicted battery class of the target battery based at least in part on the plurality of printed characters or codes.

27. The computer-implemented method of claim 21, further comprising: determining a temperature or a temperature gradient of the target battery according to a signal of the target battery received from an infrared camera; as well as In response to determining that the temperature or the temperature gradient is above a threshold, an alarm is activated or the target battery is transferred to a safety box.

28. The computer-implemented method of claim 21, wherein: Determining a predicted battery class of the target battery from a plurality of battery classes based on the plurality of signals using a classifier machine learning model includes determining the predicted battery class from one or more images of the target battery using an object detection neural network.

29. A system comprising: one or more memory devices, the one or more memory devices comprising a classifier machine learning model and a plurality of battery classes; as well as one or more processors configured to cause the system to perform operations comprising: determining one or more attributes of a target battery based on a plurality of signals from a plurality of sensors, the one or more attributes comprising one or more of a size, a battery chemistry, printed characters, or a form factor of the target battery; as well as A predicted battery class of the target battery is determined from a plurality of battery classes based on the determined one or more attributes of the target battery using the classifier machine learning model.

30. The system of claim 29, the operations further comprising: indicating the predicted battery category of the target battery to a battery sorting mechanism; as well as The target battery is transferred to a box corresponding to the predicted battery category of the target battery using the battery sorting mechanism.

31. The system of claim 29, wherein: The operations also include receiving one or more label images and one or more contour images of the target battery from an RGB camera of the plurality of sensors; Determining the one or more attributes of the target battery includes: Determining the printed characters of the target battery from the one or more label images using optical character recognition; and determining a form factor of the target battery from the one or more contour images; and The predicted battery category of the target battery is determined based on the printed characters and the form factor using the classifier machine learning model.

32. The system of claim 29, wherein: Determining the one or more attributes of the target battery includes: determining one or more dimensions of the target battery based on scan data from the 3D scanner of the plurality of sensors; and determining the battery chemistry of the target battery based on x-ray attenuation data from an x-ray scanning array of the plurality of sensors; and The predicted battery class of the target battery is determined based on the one or more dimensions and the battery chemistry using the classifier machine learning model.

33. The system of claim 29, the operations further comprising: receiving additional signals corresponding to additional batteries from the plurality of sensors; as well as An additional predicted battery class for the additional battery is determined from the plurality of battery classes based on the additional signal using the classifier machine learning model.

34. The system of claim 33, the operations further comprising: indicating the predicted battery category of the target battery and the additional predicted battery category of the additional battery to a battery sorting mechanism; as well as Using the battery sorting mechanism, each of the target battery and the additional battery is individually transferred to a plurality of bins respectively associated with the plurality of battery categories.

35. The system of claim 33, the operations further comprising: determining a dangerous anomaly within at least one of the additional batteries based on signals from the infrared camera of the plurality of sensors; as well as In response, the at least one cell is transferred to an immersion tank containing a flame retardant.

36. A non-transitory computer readable medium storing executable instructions, which when executed by at least one processor causes the at least one processor to perform operations comprising: receiving a plurality of signals corresponding to the target battery from the plurality of sensors; Determine, using a classifier machine learning model, a predicted battery class of the target battery from a plurality of battery classes based on the plurality of signals; as well as The predicted battery category of the target battery is indicated to a battery sorting mechanism.

37. The non-transitory computer readable medium of claim 36, the operations further comprising: capturing one or more tag images using an RGB camera of the plurality of sensors; as well as Using object character recognition (OCR), a plurality of printed characters or codes are recognized from the one or more label images.

38. The non-transitory computer readable medium of claim 37, the operations further comprising: capturing one or more contour images of the target battery using one or more RGB cameras of the plurality of sensors; as well as The physical specifications of the target battery are determined based on the one or more contour images of the target battery.

39. The non-transitory computer readable medium of claim 38, the operations further comprising capturing x-ray attenuation data of the target cell using an x-ray scanning array of the plurality of sensors.

40. The non-transitory computer readable medium of claim 39, wherein: Utilizing a classifier machine learning model to determine a predicted battery category of the target battery based on the multiple signals includes utilizing a decision tree to determine the predicted battery category based on the x-ray attenuation data of the target battery, the form factor, and the multiple printed characters or codes from the one or more label images.

Images