A system and method for detecting anomalies related to dispensers using machine learning and machine vision.

Abstract

A system configured to determine the presence of an anomaly related to the dispensing system components, fluid supply unit and / or tool of a dispenser includes: a fluid supply unit for containing fluid; a tool for supporting parts; one or more sensors configured to detect the characteristics of the dispensing system components, fluid supply unit and / or tool; a controller configured to activate one or more sensors to acquire initial data related to the dispensing system components, fluid supply unit and / or tool at a first time point; process the initial data in a trained neural network to determine the presence of an anomaly related to the dispensing system components, fluid supply unit and / or tool; log the presence of the anomaly; generate an alert indicating the presence of the anomaly and transmit it to a human-machine interface; and to activate one or more sensors to acquire subsequent data related to the dispensing system components, fluid supply unit and / or tool at a second time point after the first time point; process the subsequent data to determine whether the anomaly has been corrected.

Description

Technical Field

[0001] [Cross - Reference to Related Applications] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 520,557, filed Aug. 18, 2023, and U.S. Provisional Patent Application No. 63 / 520,565, filed Aug. 18, 2023, each of which is hereby incorporated by reference in its entirety as if fully set forth herein for all purposes.

[0002] The present disclosure relates to systems and methods for determining the existence of anomalies (e.g., low fluid level height in a fluid supply, high fluid accumulation height on a tool) associated with dispenser ejection system components, and specifically to systems and methods for capturing data related to ejection system components and processing it in a trained neural network to determine the existence thereof.

Background Art

[0003] Ejection system components and their characteristics (e.g., fluid level height in a fluid supply, fluid accumulation on a tool) are important to monitor in an automated ejection system so that an operator and / or an automated process can determine the existence of anomalies associated with the ejection system components in order to maintain productivity, prevent yield problems in production, and / or efficiently utilize materials, and / or predict when corrective actions should be taken to correct anomalies.

[0004] Certain methods for monitoring ejection system components, specifically fluid syringes, are known but have various drawbacks, such as requiring consistent operator supervision or calibration and / or being limited in certain aspects (e.g., accuracy, range). For example, the fluid level height in a fluid syringe is typically monitored using one or more of the following methods, each of which has one or more drawbacks.

[0005] [Table 1] [Overview of the project] [Problems that the invention aims to solve]

[0006] Therefore, there is a need for a system and method to determine the presence of abnormalities related to the dispenser's dispensing system components (e.g., low fluid level in the fluid supply section, high fluid accumulation level on the tool), specifically by taking in data related to the dispensing system components and processing it in a trained neural network. Such a system and method is particularly useful and desirable in the processing / packaging of large quantities of electronic equipment where process monitoring and predictive maintenance are important. [Means for solving the problem]

[0007] In one example, a system is provided configured to determine the presence of an anomaly related to the fluid supply section of a dispenser. The system includes a fluid supply section for containing fluid. The system further includes a tool for supporting parts. The system further includes one or more sensors configured to detect characteristics of dispensing system components. The system further includes a controller. The controller is configured to activate one or more sensors to capture initial data related to the fluid supply section at a first time point. The controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to the fluid supply section. The controller is further configured to log the presence of the anomaly. The controller is further configured to generate an alert indicating the presence of the anomaly and transmit it to a human-machine interface, the alert providing the user with location information about the anomaly and / or instructions for correcting the anomaly. The controller is further configured to activate one or more sensors to capture subsequent data related to the fluid supply section at a second time point after the first time point. The controller is further configured to process the subsequent data to determine whether the anomaly has been corrected.

[0008] In certain examples, one or more sensors may include cameras. In such examples, the controller may be configured to activate the cameras to capture an initial image of the fluid supply at a first time point, and to activate the cameras to capture subsequent images of the fluid supply at a second time point. In some examples, the anomaly may be the fluid surface height within the fluid supply. In such examples, a trained neural network may be trained on at least one image of the fluid supply having a fluid surface height above a predetermined threshold. In such examples, the controller may be configured to use the trained neural network to categorize the fluid surface height within the fluid supply as high, medium, or low. In the same or other examples, the controller may be configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. In the same or other examples, an alert may indicate that the fluid surface height within the fluid supply is below a predetermined threshold. In the same or other examples, an alert may provide the user with instructions to replenish the fluid supply so that the fluid surface height is above a predetermined threshold. In the same or other examples, the alert may provide the user with instructions to replace the fluid supply unit with a second fluid supply unit having a fluid surface height above a predetermined threshold. In several examples, the controller may be configured to determine the predicted fluid height in the fluid supply unit based on the amount of fluid discharged by the dispenser, to determine the detected fluid height in the fluid supply unit based on an initial image of the fluid supply unit, and to identify excursions by comparing the predicted fluid height in the fluid supply unit with the detected fluid height in the fluid supply unit. In several examples, the trained neural network may be trained on data related to multiple fluid supply units. In such examples, the controller may use the trained neural network to determine the presence of anomalies related to the fluid supply unit based on the classification of the initial data.

[0009] In several examples, the system may further include a discharge head. In such examples, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. In further examples, if the controller determines that an anomaly is present, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head to cease. In further examples, if the controller determines that the anomaly has been corrected, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head to resume.

[0010] In another example, a system is provided configured to determine the presence of an anomaly related to a dispenser tool. The system includes a fluid supply unit for containing fluid. The system further includes a tool for supporting a part. The system further includes one or more sensors configured to detect the characteristics of the dispensing system components. The system further includes a controller. The controller is configured to activate one or more sensors to capture initial data related to the tool at a first time point. The controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to the tool. The controller is further configured to log the presence of the anomaly. The controller is further configured to generate and transmit an alert indicating the presence of an anomaly to a human-machine interface, the alert providing the user with location information about the anomaly and / or instructions for correcting the anomaly. The controller is further configured to activate one or more sensors to capture subsequent data related to the tool at a second time point after the first time point. The controller is further configured to process the subsequent data to determine whether the anomaly has been corrected.

[0011] In certain examples, one or more sensors may include cameras. In such examples, the controller may be configured to activate the cameras to capture an initial image of the tool at a first time point, and to activate the cameras to capture a subsequent image of the tool at a second time point. In several examples, the anomaly may be the fluid accumulation height on the tool. In such examples, a trained neural network may be trained on at least one image of the tool having a fluid accumulation height below a predetermined threshold. In such examples, the controller may be configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. In the same or other examples, an alert may indicate that the fluid accumulation height is greater than a predetermined threshold. In the same or other examples, an alert may provide the user with instructions to clean the tool so that the fluid accumulation height is below a predetermined threshold. In several examples, the controller may be configured to determine a predicted fluid accumulation height on the tool based on the number of parts removed from the tool, determine a detected fluid accumulation height on the tool based on an initial image of the tool, and identify deviations by comparing the predicted fluid accumulation height on the tool with the detected fluid accumulation height on the tool. In several examples, the trained neural network may be trained on data associated with multiple tools. In such examples, the controller may use the trained neural network to determine the presence of tool-related anomalies based on the classification of initial data. In several examples, the controller may be configured to repeatedly capture and process initial and subsequent data each time a part is removed from the tool.

[0012] In several examples, the system may further include a discharge head. In such examples, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. In further examples, if the controller determines that an anomaly is present, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head to cease. In further examples, if the controller determines that the anomaly has been corrected, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head to resume.

[0013] In a further example, a system is provided configured to determine the presence of an anomaly related to the fluid supply unit and / or tool of a dispenser. The system includes a fluid supply unit for containing fluid. The system further includes a tool for supporting parts. The system further includes one or more sensors configured to detect characteristics of dispensing system components. The system further includes a controller. The controller is configured to activate one or more sensors to capture initial data related to at least one of the fluid supply unit and the tool at a first time point. The controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to at least one of the fluid supply unit and the tool. The controller is further configured to log the presence of the anomaly. The controller is further configured to generate and transmit an alert indicating the presence of an anomaly to a human-machine interface, the alert providing the user with location information about the anomaly and / or instructions for correcting the anomaly. The controller is further configured to activate one or more sensors to capture subsequent data related to at least one of the fluid supply unit and the tool at a second time point after the first time point. The controller is further configured to process subsequent data and determine whether the anomaly has been corrected.

[0014] In several examples, the system may further include a discharge head. In such examples, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. In further examples, if the controller determines that an anomaly is present, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head to cease. In further examples, if the controller determines that the anomaly has been corrected, the controller may be configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head to resume.

[0015] In another example, a system is provided configured to determine the presence of an anomaly related to a dispensing system component of a dispenser. The system includes a fluid supply unit for containing fluid. The system further includes a tool for supporting the component. The system further includes one or more sensors configured to detect the characteristics of the dispensing system component. The system further includes a controller. The controller is configured to activate one or more sensors to capture initial data related to the dispensing system component at a first time point. The controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to the dispensing system component. The controller is further configured to log the presence of the anomaly. The controller is further configured to generate and transmit an alert indicating the presence of the anomaly to a human-machine interface, the alert providing the user with location information about the anomaly and / or instructions for correcting the anomaly. The controller is further configured to activate one or more sensors to capture subsequent data related to the dispensing system component at a second time point after the first time point. The controller is further configured to process the subsequent data to determine whether the anomaly has been corrected.

[0016] In several examples, the controller may use a trained neural network to determine the presence of anomalies based on training with data related to multiple discharge system components.

[0017] In several examples, one or more sensors may include a camera. In such examples, the controller may be configured to activate the camera to capture an initial image of the dispenser's dispensing system components at a first time point, and to activate the camera to capture a subsequent image of the dispenser's dispensing system components at a second time point.

[0018] In several examples, the anomaly may be the fluid surface height within the fluid supply unit. In such examples, the trained neural network may be trained on at least one image of the fluid supply unit having a fluid surface height above a predetermined threshold. In such examples, the controller may be configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image.

[0019] In several examples, the anomaly may be the fluid accumulation height on the tool. In such examples, the trained neural network may be trained on at least one image of a tool having a fluid accumulation height below a predetermined threshold. In such examples, the controller may be configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image.

[0020] In several examples, the anomaly may be the presence of a foreign object in the dispenser's working area. In such examples, the trained neural network may be trained on at least one image of the dispenser's working area that is free of foreign objects. In such examples, the controller may be configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image.

[0021] In several examples, the anomaly may be at least one of a full purge cup and a dirty cleaning strip in the dispenser's service station. In such examples, the trained neural network may be trained on at least one image of at least one of a non-full purge cup and a clean cleaning strip in the dispenser's service station. In such examples, the controller may be configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial images.

[0022] In several cases, the anomaly may be the misloading of a part on the tool. In such cases, the trained neural network may be trained on at least one image of a part properly placed on the tool. In such cases, the controller may be configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image.

[0023] In several cases, the anomaly may be the misposition of the bubble on the dispenser's spirit level. In such cases, the trained neural network may be trained on at least one image of a properly positioned bubble on the dispenser's spirit level. In such cases, the controller may be configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

[0024] In multiple examples, the anomaly may be a user without permission to the dispenser. In such examples, the trained neural network may be trained based on at least one of face recognition of an authorized user, biometric information, and a unique badge. In such examples, the controller may be configured to utilize the trained neural network to determine the presence of an anomaly based on the classification of the initial image.

[0025] In multiple examples, the anomaly may be a characteristic of a dispensing system component selected from the group consisting of a leak related to a dispensing system component, damage or wear related to a dispensing system component, and a jam or disconnection related to a dispensing system component. In such examples, the trained neural network may be trained based on at least one image of a dispensing system component without a leak, damage or wear, and / or a jam or disconnection. In such examples, the controller may be configured to utilize the trained neural network to determine the presence of an anomaly based on the classification of the initial image.

[0026] In multiple examples, the anomaly may be the temperature and / or LED status of a dispensing system component. In such examples, the trained neural network may be trained based on at least one image of the temperature and / or LED status of a dispensing system component having a temperature below a predetermined threshold and / or an operable LED status. In such examples, the controller may be configured to utilize the trained neural network to determine the presence of an anomaly based on the classification of the temperature and / or LED status of the dispensing system component. In the same or other examples, one or more sensors may include an infrared camera configured to determine the temperature and / or LED status of a dispensing system component.

[0027] In multiple examples, the system may further include a dispensing head. In such examples, the controller may be configured to send a signal to the dispensing head to cause the dispensing head to dispense fluid from the dispensing head before activating one or more sensors to capture initial data. In a further example, if the controller determines the presence of an anomaly, the controller may be configured to send a signal to the dispensing head to cause the dispensing head to cease dispensing fluid from the dispensing head. In a further example, if the controller determines that the anomaly has been corrected, the controller may be configured to send a signal to the dispensing head to cause the dispensing head to resume dispensing fluid from the dispensing head.

[0028] The following description of examples for illustration can be better understood when read in conjunction with the accompanying drawings. It is understood that the possible examples of the disclosed systems and methods are not limited to those depicted.

Brief Description of the Drawings

[0029] [Figure 1] FIG. showing a system configured to determine the presence of an anomaly related to the dispensing system components of a dispenser according to one example. [Figure 2] FIG. showing a method for determining the presence of an anomaly related to the dispensing system components of a dispenser according to one example. [Figure 3A] FIG. showing a fluid supply having a fluid surface height highly categorized using a trained neural network according to one example. [Figure 3B] FIG. showing a fluid supply having a fluid surface height moderately categorized using a trained neural network according to one example. [Figure 3C] FIG. showing a fluid supply having a fluid surface height lowly categorized using a trained neural network according to one example. [Figure 4A] FIG. showing a fluid supply containing fluid. [Figure 4B] Figure 4A shows the masking of the fluid color in the fluid supply section. [Figure 4C] Figure 4B shows the extraction of fluid color from masking. [Figure 5A] This figure shows a tool without fluid buildup, as detected using a trained neural network, as an example. [Figure 5B] This figure shows another tool without fluid accumulation, as detected using a trained neural network, as an example. [Figure 5C] This figure shows a tool with multiple fluid accumulation examples detected using a trained neural network, as an example. [Figure 5D] This figure shows another tool with a single fluid accumulation example detected using a trained neural network, as an example. [Figure 6] This figure illustrates the process for training and / or retraining a model whenever a new discharge system component and / or potential anomaly is introduced, using one example. [Modes for carrying out the invention]

[0030] In the following detailed description, references will be made to the accompanying drawings that form part of this specification. In these drawings, unless otherwise indicated in the context, similar symbols will identify similar components. The illustrative examples provided in the detailed description and drawings are for illustrative purposes only and are not intended to limit. Other examples may be used and other modifications may be made without departing from the spirit or scope of the subject matter presented herein. It will be readily apparent that the aspects of this disclosure as generally described herein and illustrated in the drawings may be arranged, substituted, combined and designed in a wide variety of different configurations. Each of these different configurations is explicitly assumed and forms part of this disclosure.

[0031] Conventional methods for monitoring and / or inspecting dispensers for abnormalities have been sufficient for their intended purposes. However, there is a need for systems and methods to determine the presence of abnormalities related to the dispenser's dispensing system components (e.g., low fluid level in the fluid supply section, high fluid accumulation level on the tool), specifically by taking in data related to the dispensing system components and processing it in a trained neural network. These systems and methods are particularly useful and desirable in the processing / packaging of large quantities of electronic equipment where process monitoring and predictive maintenance are important.

[0032] Referring first to Figure 1, an exemplary system 10 is shown. As shown, system 10 may include a dispenser 100, such as a spray dispenser. The dispenser 100 may include a dispensing head or nozzle 102. The dispenser 100 may further include a movement control system, such as a robot 104, for moving the dispenser (e.g., the dispenser's dispensing head or nozzle 102) with three or more degrees of freedom. In certain examples, the dispenser 100 may be mounted on the robot 104, and in several examples, the robot 104 may be configured to move the dispenser (e.g., the dispenser's dispensing head or nozzle 102) in a certain pattern relative to certain components, for example, across the surface of a circuit board having components. In a non-limiting example, the robot 104 may be configured to move the dispenser 100 from the dispenser to a suitable dispensing location for dispensing material onto a corresponding area or component on the circuit board, and then move the dispenser to the next area or component designated to receive the material. As a further non-limiting example, underfill material may be dispensed by the dispenser 100 in a series of events to define a short line segment, using a robot 104 configured to move the dispenser 100 between consecutive areas or components of a circuit board. In some examples, the robot 104 may include multiple dispensers.

[0033] The dispenser 100 may further include various other components or stations typically found in an injection dispensing system, such as a fluid supply unit 110 for containing fluid (e.g., syringe, cartridge, reservoir), a tool 120 for supporting parts (e.g., a wafer chuck, vacuum fastener, loading / unloading robot or conveyor or similar transfer element), a dispensing head or nozzle, a robot for operating the dispenser (e.g., the dispenser's dispensing head or nozzle), a work area, a service station (e.g., including a purge cup, cleaning strip, etc.), a conveyor, a pump, associated hoses and electronic equipment.

[0034] System 10 may include a controller 140 configured to perform the operations described herein (e.g., activating sensors(s), processing captured data, determining the presence of anomalies, logging anomalies, generating and sending alerts regarding anomalies). While this disclosure describes a controller 140 that performs the operations described herein, it should be understood that the term “controller” as used herein may include one or more controllers, microprocessors, etc., that work together to perform the operations described herein.

[0035] System 10 may further include one or more sensors 130. In the example shown in Figure 1, the sensor 130 is in the form of a single camera, but other examples may include one, two, three, four, five, or more sensors, which are cameras or other sensors, and are not limited to this. Each embodiment of sensor 130 is generally configured to detect characteristics of the dispenser's dispensing system components, such as the fluid supply unit 110, the tool 120, and / or any other components of System 10. In examples where sensor 130 is one or more cameras, the camera(s) may generally be configured to capture images of desired dispensing system components of the dispenser, specifically dispenser 100.

[0036] In certain examples, the sensor(s) 130 may be mounted on and / or adjacent to the dispenser 100, and the dispenser 100 may include mounting components, mounting structures, mounting fasteners, etc. In some examples, the sensor(s) 130 may be configured to rotate around the dispenser 100 to capture images of one or more dispensing system components from multiple sides thereof. In this regard, the system 10 may include components, controllers, motors, etc. for performing the rotation of the sensor(s) relative to components, controllers, motors, etc., or the rotation of the sensor(s) relative to components, controllers, motors, etc. As will be understood by those skilled in the art, in addition to and / or instead of capturing one or more still images of the dispensing system components, the sensor(s) 130 may be configured to capture one or more videos of the dispensing system components and / or other data or characteristics associated with the dispensing system components (e.g., temperature, LED status).

[0037] Furthermore, system 10 may include a neural network 160. The neural network 160 may be implemented as a trained neural network. In several examples, the neural network 160 may be trained by system 10. In several examples, the neural network 160 may be trained by the controller 140 of system 10. The neural network 160 may implement artificial neurons in an artificial neural network or a simulated neural network. The neural network 160 may implement artificial neurons in an interconnected group of natural or artificial neurons that may use a mathematical model or computational model based on a connectionism approach of computation for information processing within system 10. In most cases, the neural network 160 may implement an adaptive system that changes its structure based on external or internal information received within and / or from system 10. In several examples, sensors 130 may be directly connected to the neural network 160. In several examples, the controller 140 may trigger or poll the neural network 160 to communicate information between the controller 140 and the neural network 160 (for example, data acquired by sensors 130 and / or decisions processed by the neural network 160), as described herein. The controller 140 is generally configured to use the neural network 160 to perform various operations described herein, and the neural network 160 may be part of the controller 140 in some particular examples, or separate from the controller 140 in other examples.

[0038] In several embodiments, the system 10, neural network 160, controller 140, sensor(s) 130, etc., may be configured to inspect the tool mounting area to detect spilled fluid or other foreign matter on the tool and prevent damage to the parts.

[0039] In several embodiments, the system 10, neural network 160, controller 140, sensors (which may be more than one) 130, etc., may be configured to monitor the conveyor system for abnormalities such as worn belts, broken belts, jammed parts, etc.

[0040] In several embodiments, the system 10, neural network 160, controller 140, sensors (which may be more than one) 130, etc., may be configured to monitor the payload on the robot for any abnormalities such as loose parts, detached wires and hoses, bent needles, etc.

[0041] In several embodiments, the system 10, neural network 160, controller 140, sensors (or more) 130, etc., may be configured to monitor for foreign objects in the work area (e.g., hands / body parts, or components placed in the wrong place or position, such as loose wires). The equipment should be interlocked to prevent this, but it is (unfortunately) common practice for users to circumvent this interlock mechanism.

[0042] In several embodiments, the system 10, neural network 160, controller 140, sensors (which may be more than one) 130, etc., may be configured to monitor the service station for a full purge cup, a dirty cleaning strip, or other abnormalities.

[0043] In several embodiments, the system 10, neural network 160, controller 140, sensors(s) 130, etc., may be configured to monitor customer parts to ensure that the customer parts are properly placed and locked in the correct position before the dispensing process.

[0044] In several embodiments, the system 10, neural network 160, controller 140, sensors (which may be more than one) 130, etc., may be configured to monitor the temperatures of various components of the equipment using infrared cameras to ensure that these components are within the correct / expected temperature range. This may include fluid temperature, valve temperature, tool temperature, bearing temperature, etc.

[0045] In several embodiments, the system 10, neural network 160, controller 140, sensors (which may be more than one) 130, etc., may be configured to use facial recognition, a camera that uses biometric information, or a camera that reads a user badge, in order to ensure that the user has the authority to use the device.

[0046] In several embodiments, the system 10, neural network 160, controller 140, sensors (which may be more than one) 130, etc., may be configured to monitor the bulk material feed system for leaks or abnormalities using cameras mounted on the outside of the machine.

[0047] In several embodiments, the system 10, neural network 160, controller 140, sensor(s) 130, etc., may be configured to verify whether all items have been returned to their proper places after cleaning / maintenance.

[0048] In several embodiments, the system 10, neural network 160, controller 140, sensor(s) 130, etc., may be configured to monitor temperature and / or LED status using cameras in the electronic equipment area.

[0049] In several embodiments, the system 10, neural network 160, controller 140, sensor(s) 130, etc., may be configured to inspect for nozzle contamination.

[0050] In several embodiments, the system 10, neural network 160, controller 140, sensor(s) 130, etc., may be configured to monitor a bubble on a spirit level to confirm that it is level.

[0051] In several embodiments, the system 10, neural network 160, controller 140, sensor(s) 130, etc., may be configured to take photographs to ensure that everything is back to normal before restarting when the machine is open and a safety interlock is triggered.

[0052] Referring now to Figure 2, a method for determining the presence of an anomaly related to the dispensing system components of a dispenser is shown according to one example. Specifically, the anomaly detection method 200 may be performed in a system 10, such as the system 10 in Figure 1. The anomaly detection method 200 may be performed, for example, in the system 10 in Figure 1, using a controller 140 and / or a neural network 160. The following description of the anomaly detection method 200 will be given in relation to system 10 and its components, but it will be understood that the anomaly detection method 200 and its techniques are equally applicable to other related systems and components.

[0053] As shown in Figure 2, the anomaly detection method 200 may be initiated in step 202. The anomaly detection method 200 (and step 202) may be triggered. In several examples, the anomaly detection method 200 may be executed automatically at predetermined intervals in response to a command from the controller 140, in response to user input, and / or in succession. In some examples, whenever a specific action is taken with respect to a discharge system component, the anomaly detection method 200 may be activated so that the sensor(s) 130 are activated to capture or recapture data related to the discharge system component. In a non-limiting example, whenever a new fluid supply unit 110 is installed, the anomaly detection method 200 may be activated, and the sensor(s) 130 may be activated by the controller 140 as described herein to capture data related to the fluid supply unit 110. In a further non-limiting example, the anomaly detection method 200 may be activated each time a part is removed from the tool 120, and sensors 130 may be activated by the controller 140 as described herein to capture data related to the tool 120. In several examples, user input may be provided to a human-machine interface (HMI), such as one or more buttons or a touchscreen. Furthermore, user input may be provided to the controller 140. Furthermore, the HMI may be configured to provide outputs to the user, which may include visual outputs, audible outputs, tactile outputs, sensory outputs, etc. In a non-limiting example, the HMI may be implemented remotely from the system 10 (e.g., via a network).

[0054] In step 202, initial data related to the dispensing system components is acquired at a first time point. As described herein, the initial data may be acquired by a sensor(s) 130, which may be activated by a controller 140. In several examples, the initial data may be, for example, an initial image of the dispensing system components of the dispenser at a first time point. In an indefinite example, the initial data may be an initial image of the fluid supply unit 110 at a first time point for determining the fluid level height in the fluid supply unit 110, etc. In a further indefinite example, the initial data may be an initial image of the tool 120 at a first time point for determining the fluid accumulation height in the tool 120, etc. The initial data may generally be any data or characteristics related to the selected dispensing system components (e.g., temperature or LED status of the dispensing system components). Step 202 may generally be repeated each time the selected dispensing system components are replaced, cleaned, etc., or whenever it is desirable to determine or re-determine the presence of any abnormality related to the dispensing system components.

[0055] In step 204, the acquired data relating to the initial dispensing system components is processed. As described herein, the acquired initial data may be processed by the controller 140 and / or the neural network 160. In several examples, the acquired initial data may be processed (e.g., by the controller 140 and / or the neural network 160) to generate an initial value, which may generally be based on an aspect or parameter relating to the dispensing system components. In an unspecified example, this value may be based on the fluid surface height in the fluid supply unit 110. In a further unspecified example, this value may be based on the presence and / or amount of fluid accumulation in the tool 120. This value may generally be based on the operating status of the dispensing system components or any other characteristic of the dispensing system components.

[0056] Furthermore, various values, data, images, etc., may be used by the system 10 and / or controller 140 to train the neural network 160 to determine the presence and / or absence of anomalies related to the ejection system components. Once the neural network 160 is trained, it may be implemented by the system 10 and / or controller 140 as a trained neural network.

[0057] As shown in Figures 3A to 3C, a predetermined portion (subset) of an image depicting the ejection system components may be captured, and this predetermined portion may be input into a neural network for training, as described herein, and then processed to generate initial values ​​and / or to determine the presence of anomalies associated with the ejection system components. As described herein, in some examples, these values ​​may be based on the pixel intensity of the initial image. In such examples, the initial image may be processed (e.g., by the controller 140) by classifying the initial image (e.g., according to a weighted statistic), such as by predicting the statistical likelihood that the initial image belongs to a particular category or class (e.g., whether the statistical likelihood is greater than a threshold probability). In some examples, the neural network 160 may analyze one or more pixels of the captured initial image to determine the pixel intensity, etc., which may be used to generate a trained neural network. In several examples, the neural network 160 may be trained on arbitrary captured data (e.g., images), such as a standard image (referring to the fluid supply unit 110D in Figure 4A) and / or a modified image (referring to the fluid supply unit 110E in Figure 4B and the fluid supply unit 110F in Figure 4C). In several examples, the captured data may be modified, such as by changing the form of an image to enhance contrast. In several examples, the neural network 160 may be trained using one or more images of the discharge system components as described herein, and then process captured data related to the same or another discharge system component to determine the presence of possible anomalies related to the discharge system component from which the data was captured. As a non-limiting example, the neural network 60 may process the captured data presented to (e.g., input to) its neural network by generally performing a large number of simple calculations on the captured data, as described herein.The coefficients of the formulas used to perform such calculations are determined by the process of training the neural network. For example, during the training of neural network 160, the coefficients of the formulas used to perform such calculations may be randomly modified and optimized until the neural network model demonstrates success in predicting which class new input data should be assigned to (e.g., by comparing the predictions with annotated classes). Furthermore, as desired, as described herein, the training dataset may be stored or otherwise retained (e.g., as a historical dataset) and / or modified for retraining the neural network (e.g., to include new classes or supplementary examples of existing classes when data related to new ejection system components is taken in). It should be understood that the above example for processing initial images is merely one method of processing initial images, and other examples are not limited in this way.

[0058] In several examples, the trained neural network may include data (e.g., images) related to multiple dispensing system components, including dispensing system components of other dispensers. For example, data from all dispensing system components of a particular type may be taken in and input to train the neural network. In the same or other examples, data from the states (e.g., lighting, appearance) of multiple different dispensing system components may be taken in and input to train the neural network, which has been shown to improve accuracy. In addition to or instead of the above, each time a new dispensing system component of this type is added to the production line, the controller may take in and process initial data (and, in several examples, additional subsequent data) as described herein, update the neural network dataset with the newly taken in and processed data, retrain the neural network model, and deploy the retrained neural network model to other dispensers having the same type of dispensing system component. In several examples, the neural network may be trained on both historical data (e.g., test or training data from past or other ejection system components) and new data (e.g., test or training data from new ejection system components). In a non-limiting example, whenever new ejection system components and / or potential anomalies are introduced (602), as shown in process 600 of Figure 6, data associated with them may be taken up and / or processed (e.g., according to an anomaly detection method 200) as described herein (604). The newly taken up and processed data 606 may be annotated and / or categorized in 607 by classifying the data (e.g., images) using supervised learning. The newly annotated and / or categorized data may then be randomly separated in 608 into a new training and test dataset 614 and a new validation dataset 616.The training and test dataset 614 may subsequently be input to a model (e.g., a neural network) and used to train or retrain the model (610). As illustrated, in addition to the newly acquired and processed data 606 and / or randomly separated training and test datasets 614, historical training and / or test data 612 (e.g., from past or other output system components) may be input to the model and used to train or retrain the model (610). In several examples, the historical training and / or test data 612 may already be annotated and / or classified before being input to the model and used to train or retrain the model (610). The model may then be validated in 618 using the validation dataset 616. As illustrated, in addition to the randomly separated validation dataset 616, the historical training and / or test data 612 (e.g., from past or other output system components) and / or the output from the retrained model network are input to the validation in 618 and used for this validation. Subsequently, in 620, it is determined whether the accuracy and / or precision is sufficient (e.g., within an acceptable range and / or above a confidence threshold). If the accuracy and / or precision is determined to be insufficient in 620, process 600 loops back to 604 to recapture and / or reprocess additional data. Conversely, if the accuracy and / or precision is determined to be sufficient in 620, the new model is used in 622. In several examples, the new model may be placed in other dispensers or systems having similar or the same type of dispensing system components, thereby improving the accuracy and / or precision in determining the presence of anomalies related to dispensing system components across multiple dispensers or systems. While illustrated examples of process 600 have been described herein, other examples are not limited thereto, and other processes for training the neural networks or models of this disclosure may be used.

[0059] As a non-limiting example, Figure 3A shows an captured image of the fluid supply unit 110A having a fluid surface height categorized as "high" using a trained neural network. In the example shown in Figure 3A, the fluid surface height is above a predetermined fluid surface height threshold, and therefore there are no identifiable anomalies regarding the fluid surface height in the fluid supply unit 110A. Using the anomaly detection method 200 of Figure 2, the controller 140 may, as a result, determine in step 206 that no anomalies exist. Accordingly, in a particular example, the controller 140 may be configured to start or continue the dispensing operation in the system 10. For comparison, Figure 3B shows an captured image of the fluid supply unit 110B after a certain number of dispensing cycles. As seen in this figure, the captured image of the fluid supply unit 110B has a fluid surface height categorized as "moderate" using a trained neural network. Depending on the predetermined fluid surface height threshold or range described herein, such heights may generally be large enough to avoid a decrease in dispensing efficiency or accuracy. Using the anomaly detection method 200 in Figure 2, the controller 140 may, as a result, determine in step 206 that no anomalies exist. Accordingly, in a particular example, the controller 140 may be configured to start or continue the dispensing operation in the system 10. On the other hand, Figure 3C shows an captured image of the fluid supply unit 110C after many more dispensing cycles. As can be seen in this figure, the captured image of the fluid supply unit 110C has a fluid level that has been categorized as "low" using a trained neural network. As can be understood, the fluid level may be insufficient to continue the dispensing operation, thereby potentially degrading the quality of the dispensing to an unacceptable level. Based on the processing in Figure 3C, the controller 140 may, upon processing the image in Figure 3C, determine the presence of an anomaly (i.e., an unacceptably low fluid level in the fluid supply unit).Accordingly, in certain examples, the controller 140 may be configured to cause the system 10 to stop dispensing operations, log abnormalities, and / or generate and transmit alerts indicating the presence of abnormalities, as described herein. In several examples, the trained neural network may include data (e.g., images) related to multiple fluid supply units, including fluid supply units of other dispensers. For example, data from all fluid supply units of a particular type may be taken in and input to the neural network to train it. In the same or other examples, data from the states (e.g., lighting, appearance) of multiple different fluid supply units may be taken in and input to the neural network to train it, which has been shown to improve accuracy. In addition to or instead of the above, each time a new fluid supply unit of this type is added to the production line, the controller may take in and process initial data (and, in several examples, additional subsequent data) as described herein, update the neural network dataset with the newly taken in and processed data, and place the updated dataset in other dispensers having the same type of fluid supply unit.

[0060] In some examples, an initial image of a fluid supply unit 110D containing a colored fluid is captured, as shown in Figure 4A. A color mask of the fluid may be created, as shown in Figure 4B, and the colors may be extracted, as shown in Figure 4C. In such examples, the extracted colors and / or masks may be classified by a trained neural network. This neural network may be trained on one or more images of the fluid supply unit (e.g., with fluid at full height) to determine the fluid surface height and thereby determine whether there is an anomaly associated with the fluid supply unit (e.g., an unacceptably low fluid surface height). In other examples, the trained neural network may be used to classify and / or categorize the fluid surface heights within the fluid supply unit based on the classification of the initial image by a neural network trained on images of other fluid supply units (e.g., using object detection and / or classification).

[0061] As a further non-limiting example, Figures 5A and 5B show captured images of tools 120A and 120B, respectively. Each tool does not have a fluid accumulation height that can be identified using a trained neural network. In the examples shown in Figures 5A and 5B, the fluid accumulation height is below a predetermined fluid accumulation threshold, so tools 120A and 120B do not have any identifiable anomalies with respect to fluid accumulation height. Using the anomaly detection method 200 of Figure 2, the controller 140 (e.g., utilizing the neural network 160) may, as a result, determine in step 206 that no anomalies exist. Accordingly, in certain examples, the controller 140 may be configured to cause the system 10 to start or continue the dispensing operation. For comparison, Figures 5C and 5D show captured images of tools 120C and 120D, respectively, after a certain number of dispensing cycles. As can be seen in these figures, the captured images of tools 120C and 120D each have fluid accumulation heights that can be identified using a trained neural network, as described herein, with Figure 5C showing tool 120C having multiple fluid accumulation examples detected using a trained neural network, and Figure 5D showing tool 120D having a single fluid accumulation example detected using a trained neural network. As can be understood, fluid accumulation can be quite sufficient to potentially impair the quality of dispensing to an unacceptable level and / or damage subsequent parts. Based on the processing of Figures 5C and 5D, the controller 140 may use the neural network 160 to determine the presence of an anomaly (i.e., an unacceptably high fluid accumulation height on the tool) when processing the images of Figures 5C and 5D. Accordingly, in certain examples, the controller 140 may be configured to cause the system 10 to stop dispensing, log the anomaly, and / or generate and transmit an alert indicating the presence of the anomaly, as described herein.In several examples, the controller 140 may be configured to send an alert to the HMI (for example, via a network) indicating the presence of an anomaly, to display an image of the anomaly on the HMI (for example, remotely from system 10), and to annotate the image displayed on the HMI, indicating the location of the anomaly, a description of the anomaly, instructions for correcting the anomaly, etc.

[0062] In several examples, the trained neural network may be trained on data (e.g., images) related to multiple tools, including tools on other dispensers. For example, data from all tools of a particular type may be taken in and used to train the neural network. In the same or other examples, data from the conditions of multiple different tools (e.g., lighting, appearance) may be taken in and used to train the neural network, which has been shown to improve accuracy. In addition to or instead of the above, each time a new tool of this type is added to the production line, the controller may take in and process initial data (and, in several examples, additional subsequent data) as described herein, update the neural network dataset with the newly taken in and processed data, retrain the neural network, and deploy the retrained neural network model to other dispensers having the same type of tool.

[0063] In step 206, the captured initial image (and, in examples where an initial image is generated, the initial value) is used to determine the presence of possible anomalies. In a particular example, the presence of an anomaly related to the ejection system component may be determined by the statistical likelihood that the image belongs to the category of anomaly, and it is determined whether or not this statistical likelihood exceeds an acceptable value, thereby indicating the presence or absence of an anomaly related to the ejection system component. If it is determined that no anomalies exist (for example, by the controller 140 using the neural network 160), the controller 140 may then determine that the ejection system component is clean, or, for any other reason, that it is fully operational for the selected ejection operation, and the controller 140 may start or continue ejecting to the system until it becomes necessary or desirable to re-determine the presence of possible anomalies related to the ejection system component. Conversely, if it is determined that an anomaly exists (for example, by the controller 140 using the neural network 160 determining the statistical likelihood that the initial data taken in belongs to an anomaly category), the controller 140 may, as described herein, stop the discharge operation, log the anomaly, and / or generate and transmit an alert indicating the presence of the anomaly.

[0064] In step 206, the controller 140 (e.g., utilizing a neural network 160) may, in a particular example, perform any of the functions described herein (e.g., by classifying and / or categorizing the acquired data and predicting anomalies using the neural network 160) and / or be trained to perform such functions to determine the presence of anomalies related to the discharge system components, such as by analyzing time-offset or time-lapse images, videos, or other data from multiple sources. If the controller 140 (e.g., utilizing a neural network 160) determines, for example, that the statistical likelihood of an anomaly is below a certain value, the controller 140 may determine that there are no anomalies related to the discharge system components and may start or continue discharging the system 10. In such a situation, the controller 140 may still log the inspection (e.g., log that initial data has been acquired and processed and / or that it has been determined that no anomalies exist). Conversely, if the controller 140 (for example, using a neural network 160) determines that the statistical likelihood of an anomaly exceeds a certain value, the controller 140 may determine that an anomaly related to the discharge system component exists, and in response to such determination, the controller 140 may log the existence of the anomaly related to the discharge system component and / or discontinue the discharge operation in step 208.

[0065] In step 210, the controller 140 may generally generate an alert and send it (for example, via a network) to an HMI that may be implemented remotely from the system 10. The alert may generally indicate (for example, to the user of the HMI) the presence of an anomaly related to a discharge system component.

[0066] In addition, alerts may generally provide the user with location information regarding anomalies related to components of the discharge system. As a non-limiting example, if the controller 140 (e.g., utilizing a neural network 160) determines the presence of an anomaly related to the fluid supply unit 110, the alert may indicate that the fluid supply unit 110 has encountered an anomaly and / or include specific details about the anomaly (e.g., the fluid level in the fluid supply unit is unacceptably low). As a further non-limiting example, if the controller 140 (e.g., utilizing a neural network 160) determines the presence of an anomaly related to the tool 120, the alert may indicate that the tool 120 has encountered an anomaly and / or include specific details about the anomaly (e.g., the fluid accumulation height on the tool is unacceptably high and / or a component is misloaded on the tool).

[0067] In addition to or instead of the above, the alert may provide the user with instructions to correct an anomaly. As a non-limiting example, if the controller determines the presence of an anomaly related to the fluid supply unit 110, the alert may provide the user with instructions to replenish or replace the fluid supply unit 110 (for example, so that the fluid level in the fluid supply unit is at an acceptable height). As a further non-limiting example, if the controller 140 (for example, utilizing a neural network 160) determines the presence of an anomaly related to the tool 120, the alert may provide the user with instructions to clean the tool 120 and / or reposition the part on the tool 120 (for example, so that the fluid accumulation height on the tool is reduced to an acceptable level and / or so that the part is properly placed on the tool). The alert may include any visual, auditory, tactile, or sensory indicators as desired to suit a particular application. As a non-limiting example, the controller 140 may be configured to indicate the presence of an abnormality related to a dispensing system component by (for example, by switching from green to red) an LED associated with that component.

[0068] In some cases, the controller 140 may be configured to automatically initiate corrective actions aimed at correcting abnormalities (for example, automatic replenishment of the fluid supply unit when the fluid level in the fluid supply unit falls below a predetermined fluid level threshold, or automatic cleaning of the tool when the fluid accumulation height on the tool exceeds a predetermined threshold).

[0069] After initiating corrective action and / or generating and transmitting alerts that provide the user with location information about the anomaly and / or instructions to correct the anomaly, it is generally desirable to determine whether the anomaly has been corrected (i.e., whether the values ​​of the characteristics of the dispensing system component are within an acceptable range). Therefore, in step 212, subsequent data of the dispensing system component is acquired. Similar to the initial data, the subsequent data may be one or more images, videos, or other data or characteristics (e.g., temperature, LED status) related to the dispensing system component. The subsequent data may generally be acquired using the same or a different sensor 130 as the one used to acquire the initial data. In some examples, for the purpose of more accurate classification and / or categorization, it may be desirable to acquire the subsequent data (e.g., one or more subsequent images) using the same sensor (e.g., a camera) used to acquire the initial data (e.g., one or more initial images) and / or from the same angle to the dispensing system component. Subsequent data is generally captured at a second time point after the first time point (for example, after some corrective action aimed at correcting an anomaly has been taken by the user or controller 140, etc.).

[0070] Subsequently, in step 214, subsequent data related to the captured discharge system components is processed. As described herein, the captured subsequent data may be processed by the controller 140. In several examples, the captured subsequent data may be processed (e.g., by the controller 140) to generate a subsequent value. This value may generally be based on an aspect or parameter related to the discharge system component. In an unspecified example, this value may be based on the fluid surface height in the fluid supply unit 110. In a further unspecified example, this value may be based on the presence and / or amount of fluid accumulation on the tool 120. This value may generally be based on the operating status of the discharge system component or any other characteristic of the discharge system component.

[0071] Subsequent data may generally be processed using any of the techniques described herein for processing initial data and generating initial values ​​(and, in examples where subsequent values ​​are generated, subsequent values ​​may be generated). The subsequent data may be processed to determine whether the anomaly has been corrected. In some examples, the generated second value may be used to determine whether the corrective action was successful (i.e., whether the subsequent value is within an acceptable range). If the controller 140 (e.g., using a neural network 160) determines in step 214 that the anomaly has been corrected, the controller 140 may generally cause the system 10 to start, restart, or continue discharging. Conversely, if the controller 140 (e.g., using a neural network 160) determines in step 214 that the anomaly has not been corrected, the controller 140 may generally repeat step 210 (generating and sending an alert indicating that the anomaly still exists) and / or cause the system 10 to stop discharging. Generally, the controller 140 may be configured to generate and send alerts and / or halt dispensing with sufficient time to allow the system and / or user to take corrective action to avoid production delays and prevent yield losses.

[0072] In some examples, in step 214, the generated subsequent values ​​may be used to determine the presence of an anomaly related to the discharge system components (including the continued presence of any of the initially identified anomalies or the presence of another anomaly). In a non-limiting example, the subsequent data may be acquired at a second time point after the first time point in which the initial data is acquired, and the generated subsequent values ​​may be compared with the generated initial values ​​to determine whether this value (e.g., pixel intensity) has changed over time. A change over time may indicate an anomaly as described herein.

[0073] The controller 140 (for example, utilizing a neural network 160) may generally be able to determine the presence of an anomaly related to one or more dispensing system components of the dispenser 100 by monitoring such dispensing system components (for example, remotely). These anomalies may take any of the non-limiting forms described herein.

[0074] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the morphology of the fluid surface height within the fluid supply unit 110 (e.g., an unacceptably low fluid surface height), based on classification of an initial image of the fluid supply unit 110 by a neural network trained on at least one image of the fluid supply unit having a fluid surface height above a predetermined threshold (e.g., an acceptablely high fluid surface height). In such examples, the controller 140 (e.g., using a neural network 160) may be configured to use a trained neural network to categorize the fluid surface height within the fluid supply unit 110 by categorizing the fluid surface height as being below or above a predetermined fluid surface height threshold, and / or categorizing the fluid surface height as high, medium, or low. The controller 140 is generally configured to log anomalies and generate and transmit alerts indicating the presence of anomalies to the HMI, as described herein. In some such examples, the alert may indicate that the fluid level in the fluid supply is below a predetermined threshold, and / or provide the user with instructions to replenish the fluid supply so that the fluid level is equal to or above the predetermined threshold, and / or to replace the fluid supply with a second fluid supply having a fluid level equal to or above the predetermined threshold. In some such examples, the controller 140 may be configured to determine the predicted fluid level in the fluid supply based on the amount of fluid discharged by the dispenser, etc. The predicted fluid level may then be compared with the detected fluid level in the fluid supply (based on, for example, an initial image of the fluid supply and its processing) to identify possible deviations (e.g., a discrepancy between two values ​​that suggests a problem related to the fluid supply or dispenser).

[0075] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the form of fluid accumulation height on the tool 120 (e.g., an unacceptably high fluid accumulation height), based on classification of an initial image of the tool 120 by a neural network trained on at least one image of the tool having a fluid accumulation height below a predetermined threshold (e.g., an acceptablely low fluid accumulation height). In such examples, the controller 140 (e.g., using a neural network 160) may be configured to use a trained neural network to categorize the fluid accumulation height on the tool 120 by categorizing the fluid accumulation height as being below or above a predetermined fluid accumulation threshold, and / or categorizing the fluid accumulation height as high, medium, or low. The controller 140 is generally configured to log anomalies and generate and transmit alerts indicating the presence of anomalies to the HMI, as described herein. In some such examples, the alert may indicate that the fluid accumulation height on the tool is greater than a predetermined threshold, and / or provide the user with instructions to clean the tool so that the fluid accumulation height is below the predetermined threshold. In some such examples, the controller 140 may be configured to determine the predicted fluid accumulation height on the tool based on the number of parts removed from the tool, etc. The predicted fluid accumulation height may then be compared with the detected fluid accumulation height on the tool (based on, for example, an initial image of the tool and its processing) to identify possible deviations (e.g., a discrepancy between two values ​​that suggests a problem related to the tool or dispenser).

[0076] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the form of a foreign object in the dispenser's work area, based on the classification of an initial image of the dispenser's work area by a neural network trained on at least one image of the dispenser's work area that is completely free of foreign objects. The controller 140 is generally configured to log the anomaly and generate and transmit an alert indicating the presence of the anomaly to the HMI, as described herein. In some such examples, the alert may indicate that a foreign object (e.g., a body part or a component placed in the wrong place or position, such as a loose wire) has been detected in the work area, and / or may provide the user with instructions to clean the work area so that no foreign objects are present in the work area.

[0077] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the form of a full purge cup and / or a dirty cleaning strip at the dispenser's service station, based on the classification of an initial image of the service station by a neural network trained on at least one image of at least one of a full purge cup and / or a dirty cleaning strip at the dispenser's service station. The controller 140 is generally configured to log the anomaly and generate and transmit an alert indicating the presence of the anomaly to the HMI, as described herein. In some such examples, the alert may indicate that a full purge cup and / or a dirty cleaning strip has been detected at the service station, and / or provide the user with instructions to empty or replace the purge cup and / or to clean or replace the cleaning strip.

[0078] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the form of misloading of a part on a tool, based on the classification of an initial image of a part on a tool by a neural network trained on at least one image of a part properly loaded on a tool. The controller 140 is generally configured to log the anomaly and generate and transmit an alert indicating the presence of the anomaly to the HMI, as described herein. In some such examples, the alert may indicate that a part is misloaded on the tool and / or provide the user with instructions for properly loading the part on the tool.

[0079] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the form of a misposition of the bubble on the dispenser's spirit level, based on the classification of an initial image of the bubble on the spirit level by a neural network trained on at least one image of a properly positioned bubble on the spirit level of the dispenser. The controller 140 is generally configured to log the anomaly and generate and transmit an alert indicating the presence of the anomaly to the HMI, as described herein. In some such examples, the alert may indicate that the bubble on the spirit level is in the wrong position on the tool, and / or provide the user with instructions for properly positioning the bubble on the spirit level.

[0080] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the form of an unauthorized user of the dispenser, based on, for example, the classification of an initial image of the user by a neural network trained on at least one of authorized user facial recognition, biometric information, and a unique badge. The controller 140 is generally configured to log the anomaly and generate and transmit an alert indicating the presence of the anomaly to the HMI, as described herein. In some such examples, the alert may indicate that the user is not an authorized user of the dispenser and / or provide the user with instructions to identify an authorized user of the dispenser.

[0081] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the form of a characteristic of a dispensing system component, such as a leak, damage, or wear, and / or a blockage or interruption, based on the classification of an initial image of a dispensing system component by a neural network trained on at least one image of a dispensing system component that is free from leaks, damage, or wear, and / or blockages or interruptions. The controller 140 is generally configured to log the anomaly and to generate and transmit an alert indicating the presence of the anomaly to the HMI, as described herein. In some such examples, the alert may indicate that the user is not an authorized user of the dispenser and / or may provide the user with instructions to identify an authorized user of the dispenser.

[0082] In several examples, the controller 140 may be able to use a trained neural network to determine the presence of an anomaly in the form of temperature and / or LED status of a discharge system component, such as classifying the temperature and / or LED status of a discharge system component by a neural network trained on at least one image of the temperature and / or LED status of a discharge system component having a temperature below a predetermined threshold and / or operational LED status. The controller 140 is generally configured to log the anomaly and generate and transmit an alert indicating the presence of the anomaly to the HMI, as described herein. In some such examples, the alert may indicate that the temperature of the discharge system component is above a predetermined temperature threshold and / or that the LED status is an operational LED status, and / or provide instructions to the user to correct the anomaly. In such examples, the sensor(s) may include an infrared camera configured to determine the temperature and / or LED status of the discharge system component.

[0083] As can be understood, various techniques may be employed to process captured data related to the ejection system components, to generate one or more values ​​based on such captured data, and / or to determine the presence of anomalies related to the ejection system components using such generated one or more values. For example, the techniques described herein may utilize or incorporate machine learning and / or artificial intelligence. Machine learning may include the use of any of the various machine learning tools that employ machine learning algorithms, including neural networks such as deep neural networks (DNNs). Other examples of machine learning tools include, but are not limited to, XGBoost, convolutional machine learning tools (CNNs), support vector machines (SVMs), multiple linear regression, random forests, AdaBoost, artificial machine learning tools (ANNs): "traditional" machine learning tools, decision trees (DTs), naive Bayes, K-nearest neighbors (KNNs), hidden Markov models (HMMs), cybernetics and brain simulation, symbolic, cognitive simulation, logic-based, anti-logic, knowledge-based, sub-symbolic, concrete intelligence, computational intelligence, and soft computing. Machine learning may utilize feature vector processes, classification processes, grouping processes, regression processes, analysis processes, matching processes, training processes, diagnostic processes, etc. In one example, one or more images of output system components may be input to a closed-loop control machine learning tool for determining the classification of one or more of those images, and values ​​may be generated based on such classifications. In another example, a quality class may be assigned to one or more images of the dispensing system components, and a machine learning tool for closed-loop control may be trained to associate the quality class with the images, by assigning a quality class to each of one or more images according to a quality classification system.Depending on the quality class assigned to the image, the trained machine learning tool may control the system 10 to start, continue, or stop dispensing in response to a determination of the presence of an anomaly related to the dispensing system components of the dispenser, as described herein.

[0084] An example of system 10 may include a discharge head or nozzle 102, a fluid supply unit 110, a tool 120, a sensor(s) 130, a controller 140, and / or various drivers, control modules, etc., for controlling the operation of any other components of system 10. An example of system 10 may also include communication channels for communication of control signals, data, images, etc., between the discharge head or nozzle 102, the fluid supply unit 110, the tool 120, a sensor(s) 130, a controller 140, and / or any other components of system 10. These communication channels may be any type of wired or wireless electronic communication network, such as a wired / wireless local area network (LAN), wired / wireless personal area network (PAN), wired / wireless home area network (HAN), wired / wireless wide area network (WAN), campus network, metropolitan network, enterprise private network, virtual private network (VPN), internetwork, backbone network (BBN), global area network (GAN), internet, intranet, extranet, overlay network, etc.

[0085] The system 10 and / or controller 140 may be implemented in any type of computing device, such as a desktop computer, personal computer, laptop / mobile computer, personal data assistant (PDA), mobile phone, tablet computer, or cloud computing device, which has wired / wireless communication capabilities via a communication channel. In several examples, the controller 140 may be implemented as an inspection controller, cleaning controller, camera operation controller, ejection system controller, etc.

[0086] Furthermore, as illustrated by various examples of this disclosure, the anomaly detection method 200 may be intended for operation using a dedicated hardware implementation. This dedicated hardware includes, but is not limited to, PCs, PDAs, semiconductors, application-specific integrated circuits (ASICs), programmable logic arrays, cloud computing devices, and other hardware devices constructed to implement the methods described herein.

[0087] It should also be noted that the controller 140 and / or the anomaly detection method 200 may optionally utilize a software implementation that may be stored on a tangible storage medium. This tangible storage medium may be a magnetic medium such as a disk or tape; a magneto-optical or optical medium such as a disk; or a solid medium such as a memory card or other package containing one or more read-only (non-volatile) memories, random-access memories, or other rewritable (volatile) memories. Digital attachments to emails or other self-contained information archives or sets of archives are considered equivalent distribution media to tangible storage media. Accordingly, this disclosure is deemed to include the tangible storage media or distribution media listed herein on which the software implementations herein are stored, and also to include equivalents and successor media recognized in the art.

[0088] In addition, various examples of System 10 and / or Controller 140 may be implemented in non-general-purpose computer implementations. Furthermore, the various examples of the disclosure described herein improve the functionality of System 10, as is evident from the disclosure. Moreover, the various examples of the disclosure include computer hardware specifically programmed to solve the complex problems addressed by the disclosure. Thus, the various examples of the disclosure improve the functionality of the system overall in their specific implementations to perform the processes described by the disclosure and defined by the claims.

[0089] The sensors (which may include multiple sensors) may include cameras. The cameras may include charge-coupled devices (CCDs), CMOS image sensors, back-illuminated CMOS sensors, etc. Images captured by the cameras may be converted to various formats, including JPEG file format, RAW feature format such as Android® (operating system) 5.0 Lollipop, and stored. System 10 may include multiple cameras. Other sensor types, such as temperature sensors, accelerometers, pressure sensors, or motion sensors, may be used alone or in combination with cameras, including combinations of these sensors. Any of these sensors may be used alone, in combination with each other, or in combination with one or more cameras.

[0090] It should be noted that the explanations and descriptions of the examples shown in the figures are for illustrative purposes only and should not be construed as limiting the disclosure. Those skilled in the art will understand that this disclosure is intended to provide a variety of examples. In addition, it should be understood that the concepts described above with respect to the above examples may be used alone or in combination with any of the other examples described above. It should also be understood that the various alternative examples described above with respect to one illustrated example may apply to all examples described herein unless otherwise specified.

[0091] The following are some non-limiting examples of aspects of the present disclosure. One example is: Example 1. A system configured to determine the presence of an anomaly related to a fluid supply unit of a dispenser, comprising a fluid supply unit for containing fluid, a tool for supporting a component, one or more sensors configured to detect characteristics of the fluid supply unit, and a controller configured to activate one or more sensors to capture initial data related to the fluid supply unit at a first time point, wherein the controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to the fluid supply unit, the controller is further configured to log the presence of the anomaly, the controller is further configured to generate and transmit an alert indicating the presence of the anomaly to a human-machine interface, the alert providing the user with location information regarding the anomaly and / or instructions for correcting the anomaly, the controller is further configured to activate one or more sensors to capture subsequent data related to the fluid supply unit at a second time point after the first time point, the controller is further configured to process the subsequent data to determine whether the anomaly has been corrected.

[0092] The above examples may further include any one of the following examples or a combination of two or more of the following examples: Example 2. A system according to any example of this specification, wherein one or more sensors include cameras, and the controller is configured to activate the cameras to capture an initial image of the fluid supply at a first time point, and to activate the cameras to capture a subsequent image of the fluid supply at a second time point. Example 3. A system according to any example of this specification, wherein the anomaly is the fluid surface height in the fluid supply. Example 4. A system according to any example of this specification, wherein the controller is configured to use a trained neural network to categorize the fluid surface height in the fluid supply as high, medium, or low. Example 5. A system according to any example of this specification, wherein the trained neural network is trained on at least one image of the fluid supply having a fluid surface height above a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 6. A system according to any example of this specification, wherein the alert indicates that the fluid surface height in the fluid supply is below a predetermined threshold. Example 7. A system according to any example herein, wherein an alert provides the user with instructions to replenish a fluid supply unit so that the fluid surface height is above a predetermined threshold. Example 8. A system according to any example herein, wherein an alert provides the user with instructions to replace a fluid supply unit with a second fluid supply unit having a fluid surface height above a predetermined threshold. Example 9. A system according to any example herein, wherein the controller is configured to determine the predicted height of the fluid in a fluid supply unit based on the amount of fluid discharged by a dispenser, to determine the detected height of the fluid in the fluid supply unit based on an initial image of the fluid supply unit, and to identify deviations by comparing the predicted height of the fluid in the fluid supply unit with the detected height of the fluid in the fluid supply unit. Example 10. A system according to any example herein, wherein a trained neural network is trained on data related to multiple fluid supply units, and the controller uses the trained neural network to determine the presence of anomalies related to the fluid supply units based on the classification of the initial data.Example 11. The system according to any example herein, further comprising a discharge head, wherein the controller is further configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. Example 12. The system according to any example herein, further configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head if the controller determines that an anomaly is present. Example 13. The system according to any example herein, further configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head to resume if the controller determines that the anomaly has been corrected.

[0093] One example is as follows: Example 14. A system configured to determine the presence of an anomaly related to a tool in a dispenser comprises a fluid supply unit for containing fluid, a tool for supporting a part, one or more sensors configured to detect the characteristics of the tool, and a controller configured to activate one or more sensors to capture initial data related to the tool at a first time point, wherein the controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to the tool, the controller is further configured to log the presence of the anomaly, the controller is further configured to generate and transmit an alert indicating the presence of the anomaly to a human-machine interface, the alert provides the user with location information regarding the anomaly and / or instructions for correcting the anomaly, the controller is further configured to activate one or more sensors to capture subsequent data related to the tool at a second time point after the first time point, and the controller is further configured to process the subsequent data to determine whether the anomaly has been corrected.

[0094] The above examples may further include any one of the following examples or a combination of two or more of the following examples. Example 15. A system according to any example of this specification, wherein one or more sensors include cameras, and the controller is configured to activate the cameras to capture an initial image of the tool at a first time point, and to activate the cameras to capture a subsequent image of the tool at a second time point. Example 16. A system according to any example of this specification, wherein the anomaly is the fluid accumulation height on the tool. Example 17. A system according to any example of this specification, wherein a trained neural network is trained on at least one image of the tool having a fluid accumulation height below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 18. A system according to any example of this specification, wherein an alert indicates that the fluid accumulation height is greater than a predetermined threshold. Example 19. A system according to any example of this specification, wherein an alert provides the user with instructions to clean the tool so that the fluid accumulation height is below a predetermined threshold. Example 20. A system according to any example herein, wherein the controller is configured to determine a predicted height of fluid accumulation on the tool based on the number of parts removed from the tool, to determine a detected height of fluid accumulation on the tool based on an initial image of the tool, and to identify deviations by comparing the predicted height of fluid accumulation on the tool with the detected height of fluid accumulation on the tool. Example 21. A system according to any example herein, wherein a trained neural network is trained on data associated with multiple tools, and the controller uses the trained neural network to determine the presence of anomalies associated with the tools based on the classification of the initial data into data. Example 22. A system according to any example herein, wherein the controller is configured to repeat the acquisition and processing of initial and subsequent data each time a part is removed from the tool. Example 23. A system according to any example herein, further comprising a discharge head, wherein the controller is further configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data.Example 24. A system according to any example herein, further configured to send a signal to the discharge head to stop discharging fluid from the discharge head if the controller determines that an abnormality is present. Example 25. A system according to any example herein, further configured to send a signal to the discharge head to resume discharging fluid from the discharge head if the controller determines that the abnormality has been corrected.

[0095] One example is as follows: Example 26. A system configured to determine the presence of an anomaly related to a fluid supply unit and / or tool of a dispenser comprises a fluid supply unit for containing fluid, a tool for supporting a component, one or more sensors configured to detect characteristics of at least one of the fluid supply unit and the tool, and a controller configured to activate one or more sensors to capture initial data related to at least one of the fluid supply unit and the tool at a first time point, wherein the controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to at least one of the fluid supply unit and the tool, the controller is further configured to log the presence of the anomaly, the controller is further configured to generate and transmit an alert indicating the presence of the anomaly to a human-machine interface, the alert providing the user with location information regarding the anomaly and / or instructions for correcting the anomaly, the controller is further configured to activate one or more sensors to capture subsequent data related to at least one of the fluid supply unit and the tool at a second time point after the first time point, the controller is further configured to process the subsequent data to determine whether the anomaly has been corrected.

[0096] The above examples may further include any one of the following examples or a combination of two or more of the following examples: Example 27. A system according to any example herein, further comprising a discharge head, wherein the controller is further configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. Example 28. A system according to any example herein, wherein if the controller determines the presence of an anomaly, the controller is further configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head. Example 29. A system according to any example herein, wherein if the controller determines that the anomaly has been corrected, the controller is further configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head to resume.

[0097] One example is as follows: Example 30. A process for determining the presence of an anomaly related to the fluid supply unit of a dispenser includes configuring one or more sensors to detect the characteristics of the fluid supply unit; configuring a controller to activate one or more sensors to capture initial data related to the fluid supply unit at a first time point; configuring the controller to process the initial data in a trained neural network to determine the presence of an anomaly related to the fluid supply unit; configuring the controller to log the presence of the anomaly; and configuring the controller to generate an alert indicating the presence of the anomaly and transmit it to a human-machine interface, wherein the alert provides the user with location information regarding the anomaly and / or instructions for correcting the anomaly; configuring the controller to activate one or more sensors to capture subsequent data related to the fluid supply unit at a second time point after the first time point; and configuring the controller to process the subsequent data to determine whether the anomaly has been corrected.

[0098] The above examples may further include any one of the following examples or a combination of two or more of the following examples. Example 31. A process according to any example of this specification, wherein one or more sensors include cameras, and the controller is configured to activate the cameras to capture an initial image of the fluid supply at a first time point, and to activate the cameras to capture a subsequent image of the fluid supply at a second time point. Example 32. A process according to any example of this specification, wherein the anomaly is the fluid surface height in the fluid supply. Example 33. A process according to any example of this specification, wherein the controller is configured to use a trained neural network to categorize the fluid surface height in the fluid supply as high, medium, or low. Example 34. A process according to any example of this specification, wherein the trained neural network is trained on at least one image of the fluid supply having a fluid surface height above a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 35. A process according to any example of this specification, wherein the alert indicates that the fluid surface height in the fluid supply is below a predetermined threshold. Example 36. The process described in any example herein, wherein the alert provides the user with instructions to replenish the fluid supply unit so that the fluid surface height is above a predetermined threshold. Example 37. The process described in any example herein, wherein the alert provides the user with instructions to replace the fluid supply unit with a second fluid supply unit having a fluid surface height above a predetermined threshold. Example 38. The process described in any example herein, wherein the controller is configured to determine the predicted height of the fluid in the fluid supply unit based on the amount of fluid discharged by the dispenser, determine the detected height of the fluid in the fluid supply unit based on an initial image of the fluid supply unit, and identify deviations by comparing the predicted height of the fluid in the fluid supply unit with the detected height of the fluid in the fluid supply unit. Example 39. The process described in any example herein, wherein a trained neural network is trained on data related to multiple fluid supply units, and the controller uses the trained neural network to determine the presence of anomalies related to the fluid supply units based on the classification of the initial data.Example 40. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. Example 41. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to cause fluid to be discharged from the discharge head if the controller determines that an anomaly is present. Example 42. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to cause fluid to be discharged from the discharge head to resume when the controller determines that the anomaly has been corrected.

[0099] One example is as follows: Example 43. A process for determining the presence of an anomaly related to a tool in a dispenser comprises configuring one or more sensors to detect the characteristics of the tool; configuring a controller to activate one or more sensors to capture initial data related to the tool at a first time point; configuring the controller to process the initial data in a trained neural network to determine the presence of an anomaly related to the tool; configuring the controller to log the presence of the anomaly; configuring the controller to generate an alert indicating the presence of an anomaly and transmit it to a human-machine interface, wherein the alert provides the user with location information about the anomaly and / or instructions for correcting the anomaly; configuring the controller to activate one or more sensors to capture subsequent data related to the tool at a second time point after the first time point; and configuring the controller to process the subsequent data to determine whether the anomaly has been corrected.

[0100] The above examples may further include any one of the following examples or a combination of two or more of the following examples. Example 44. A process according to any example of this specification, wherein one or more sensors include cameras, and the controller is configured to activate the cameras to capture an initial image of the tool at a first time point, and to activate the cameras to capture a subsequent image of the tool at a second time point. Example 45. A process according to any example of this specification, wherein the anomaly is a fluid accumulation height on the tool. Example 46. A process according to any example of this specification, wherein a trained neural network is trained on at least one image of the tool having a fluid accumulation height below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 47. A process according to any example of this specification, wherein an alert indicates that the fluid accumulation height is greater than a predetermined threshold. Example 48. A process according to any example of this specification, wherein an alert provides the user with instructions to clean the tool so that the fluid accumulation height is below a predetermined threshold. Example 49. The process according to any example herein, wherein the controller is configured to determine a predicted height of fluid accumulation on the tool based on the number of parts removed from the tool, to determine a detected height of fluid accumulation on the tool based on an initial image of the tool, and to identify deviations by comparing the predicted height of fluid accumulation on the tool with the detected height of fluid accumulation on the tool. Example 50. The process according to any example herein, wherein a trained neural network is trained on data associated with multiple tools, and the controller uses the trained neural network to determine the presence of anomalies associated with the tools based on the classification of the initial data. Example 51. The process according to any example herein, wherein the controller is configured to repeat the acquisition and processing of initial and subsequent data each time a part is removed from the tool. Example 52. The process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data.Example 53. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to stop discharging fluid from the discharge head if the controller determines that an abnormality is present. Example 54. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to resume discharging fluid from the discharge head if the controller determines that an abnormality has been corrected.

[0101] One example is as follows: Example 55. A process for determining the presence of an anomaly related to the fluid supply unit and / or tool of a dispenser includes configuring one or more sensors to detect characteristics of at least one of the fluid supply unit and / or tool; configuring a controller to activate one or more sensors to capture initial data related to at least one of the fluid supply unit and / or tool at a first time point; configuring the controller to process the initial data in a trained neural network to determine the presence of an anomaly related to at least one of the fluid supply unit and / or tool; configuring the controller to log the presence of the anomaly; and configuring the controller to generate an alert indicating the presence of an anomaly and transmit it to a human-machine interface, wherein the alert provides the user with location information regarding the anomaly and / or provides the user with instructions for correcting the anomaly; configuring the controller to activate one or more sensors to capture subsequent data related to at least one of the fluid supply unit and / or tool at a second time point after the first time point; and configuring the controller to process the subsequent data to determine whether the anomaly has been corrected.

[0102] The above examples may further include any one of the following examples or a combination of two or more of the following examples: Example 56. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. Example 57. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to cause fluid to be discharged from the discharge head if the controller determines that an anomaly is present. Example 58. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to cause fluid to be discharged from the discharge head to resume when the controller determines that an anomaly has been corrected.

[0103] One example is as follows: Example 59. A system configured to determine the presence of an anomaly related to a dispensing system component of a dispenser comprises a fluid supply unit for containing fluid, a tool for supporting a component, one or more sensors configured to detect characteristics of a dispensing system component, and a controller configured to activate one or more sensors to capture initial data related to the dispensing system component at a first time point, wherein the controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to the dispensing system component, the controller is further configured to log the presence of the anomaly, the controller is further configured to generate an alert indicating the presence of an anomaly and transmit it to a human-machine interface, the alert provides the user with location information regarding the anomaly and / or instructions for correcting the anomaly, the controller is further configured to activate one or more sensors to capture subsequent data related to the dispensing system component at a second time point after the first time point, and the controller is further configured to process the subsequent data to determine whether the anomaly has been corrected.

[0104] The above examples may further include any one of the following examples or a combination of two or more of the following examples. Example 60. A system according to any example of this specification, wherein the controller uses a trained neural network to determine the presence of an anomaly based on training with data related to a plurality of dispensing system components. Example 61. A system according to any example of this specification, wherein one or more sensors include cameras, and the controller is configured to activate the cameras to capture initial images of the dispenser's dispensing system components at a first time point, and to activate the cameras to capture subsequent images of the dispenser's dispensing system components at a second time point. Example 62. A system according to any example of this specification, wherein the anomaly is a fluid surface height in a fluid supply unit, and the trained neural network is trained on at least one image of the fluid supply unit having a fluid surface height greater than or equal to a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial images. Example 63. The system according to any example herein, wherein the anomaly is the fluid accumulation height on a tool, and a trained neural network is trained on at least one image of a tool having a fluid accumulation height below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image. Example 64. The system according to any example herein, wherein the anomaly is the presence of foreign matter in the working area of ​​a dispenser. Example 65. The system according to any example herein, wherein a trained neural network is trained on at least one image of the working area of ​​a dispenser that is completely free of foreign matter, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image. Example 66. The system according to any example herein, wherein the anomaly is at least one of a full purge cup and a soiled cleaning strip in the service station of a dispenser.Example 67. The system according to any example herein, wherein a trained neural network is trained on at least one image of at least one of a purge cup and a cleaning strip that is not soiled on the service station of a dispenser, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 68. The system according to any example herein, wherein the anomaly is the misloading of a part on a tool. Example 69. The system according to any example herein, wherein a trained neural network is trained on at least one image of a part properly placed on a tool, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 70. The system according to any example herein, wherein the anomaly is the mispositioning of a bubble on the spirit level of a dispenser. Example 71. The system according to any example herein, wherein a trained neural network is trained on at least one image of a properly positioned bubble on the spirit level of a dispenser, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 72. The system described in any example herein, wherein the anomaly is an unauthorized user of the dispenser. Example 73. The system described in any example herein, wherein a trained neural network is trained on at least one of the following: facial recognition of an authorized user, biometric information, and a unique badge, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of an initial image. Example 74. The system described in any example herein, wherein the anomaly is a characteristic of a dispensing system component selected from the group consisting of leakage related to a dispensing system component, damage or wear related to a dispensing system component, and blockage or interruption related to a dispensing system component.Example 75. A system according to any example herein, wherein a trained neural network is trained on at least one image of a discharge system component that is free from leaks, damage or wear and / or blockages or interruptions, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 76. A system according to any example herein, wherein the anomaly is the temperature and / or LED status of a discharge system component. Example 77. A system according to any example herein, wherein a trained neural network is trained on at least one image of the temperature and / or LED status of a discharge system component having a temperature below a predetermined threshold and / or an operable LED status, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the temperature and / or LED status of a discharge system component. Example 78. A system according to any example herein, wherein one or more sensors include an infrared camera configured to determine the temperature and / or LED status of a discharge system component. Example 79. A system according to any example herein, further comprising a discharge head, wherein the controller is further configured to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. Example 80. A system according to any example herein, further configured to send a signal to the discharge head to stop discharging fluid from the discharge head if the controller determines that an anomaly is present. Example 81. A system according to any example herein, further configured to send a signal to the discharge head to resume discharging fluid from the discharge head if the controller determines that the anomaly has been corrected.

[0105] One example is as follows: Example 82. A process for determining the presence of an anomaly related to a dispenser dispensing system component comprises configuring one or more sensors to detect the characteristics of the dispensing system component; configuring a controller to activate one or more sensors to capture initial data related to the dispensing system component at a first time point; configuring the controller to process the initial data in a trained neural network to determine the presence of an anomaly related to the dispensing system component; configuring the controller to log the presence of the anomaly; and configuring the controller to generate an alert indicating the presence of an anomaly and transmit it to a human-machine interface, wherein the alert provides the user with location information regarding the anomaly and / or instructions for correcting the anomaly; configuring the controller to activate one or more sensors to capture subsequent data related to the dispensing system component at a second time point after the first time point; and configuring the controller to process the subsequent data to determine whether the anomaly has been corrected.

[0106] The above examples may further include any one of the following examples or a combination of two or more of the following examples. Example 83. A process according to any example of this specification, wherein the controller uses a trained neural network to determine the presence of an anomaly based on training with data related to a plurality of dispensing system components. Example 84. A process according to any example of this specification, wherein one or more sensors include cameras, and the controller is configured to activate the cameras to capture initial images of the dispenser's dispensing system components at a first time point, and to activate the cameras to capture subsequent images of the dispenser's dispensing system components at a second time point. Example 85. A process according to any example of this specification, wherein the anomaly is a fluid surface height in a fluid supply unit, and the trained neural network is trained on at least one image of the fluid supply unit having a fluid surface height greater than or equal to a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial images. Example 86. A process according to any example herein, wherein the anomaly is the fluid accumulation height on a tool, and a trained neural network is trained on at least one image of the tool having a fluid accumulation height below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image. Example 87. A process according to any example herein, wherein the anomaly is the presence of foreign matter in the working area of ​​a dispenser. Example 88. A process according to any example herein, wherein a trained neural network is trained on at least one image of the working area of ​​a dispenser that is completely free of foreign matter, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image. Example 89. A process according to any example herein, wherein the anomaly is at least one of a full purge cup and a soiled cleaning strip in the service station of a dispenser.Example 90. The process described in any example herein, wherein a trained neural network is trained on at least one image of at least one of a purge cup and a cleaning strip that is not soiled in the service station of a dispenser, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 91. The process described in any example herein, wherein the anomaly is the misloading of a part on a tool. Example 92. The process described in any example herein, wherein a trained neural network is trained on at least one image of a part properly placed on a tool, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 93. The process described in any example herein, wherein the anomaly is the mispositioning of a bubble on the spirit level of a dispenser. Example 94. The process described in any example herein, wherein a trained neural network is trained on at least one image of a properly positioned bubble on the spirit level of a dispenser, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 95. The process described in any example herein, wherein the anomaly is an unauthorized user of the dispenser. Example 96. The process described in any example herein, wherein a trained neural network is trained on at least one of the following: facial recognition of an authorized user, biometric information, and a unique badge, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of an initial image. Example 97. The process described in any example herein, wherein the anomaly is a characteristic of a dispensing system component selected from the group consisting of leakage related to a dispensing system component, damage or wear related to a dispensing system component, and blockage or interruption related to a dispensing system component.Example 98. A process according to any example herein, wherein a trained neural network is trained on at least one image of a discharge system component that is free from leaks, damage or wear, and / or blockages or interruptions, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial image. Example 99. A process according to any example herein, wherein the anomaly is the temperature and / or LED status of a discharge system component. Example 100. A process according to any example herein, wherein a trained neural network is trained on at least one image of the temperature and / or LED status of a discharge system component having a temperature below a predetermined threshold and / or an operable LED status, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the temperature and / or LED status of a discharge system component. Example 101. A process according to any example herein, wherein one or more sensors include an infrared camera configured to determine the temperature and / or LED status of a discharge system component. Example 102. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire initial data. Example 103. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to stop discharging fluid from the discharge head if the controller determines that an abnormality is present. Example 104. A process according to any example herein, further comprising configuring the controller to send a signal to the discharge head to resume discharging fluid from the discharge head if the controller determines that an abnormality has been corrected.

[0107] In particular, conditional language used herein, such as "can," "could," "might," "may," and "eg (for example)," is generally intended to convey that a particular example includes a particular feature, element, and / or step, while other examples do not, unless otherwise specifically stated or understood in the context in which they are used. Therefore, such conditional language is not generally intended to mean that a feature, element, and / or step is absolutely necessary for one or more examples, or that one or more examples necessarily include these features, elements, and / or steps. Terms such as "comprising," "including," and "having" are synonymous, used in an open-ended, inclusive manner, and do not exclude additional elements, features, actions, or behaviors.

[0108] While certain specific examples have been provided, these examples are presented for illustrative purposes only and are not intended to limit the scope of the inventions disclosed herein. Therefore, nothing in the above description is intended to imply that any particular characteristic, feature, step, module, or block is necessary or essential. Indeed, the novel methods and articles described herein can be embodied in various other forms, and furthermore, various omissions, substitutions, and modifications of the forms of the methods and articles described herein can be made without departing from the spirit of the inventions disclosed herein. The appended claims and equivalents are intended to encompass forms or modifications that fall within some of the scope and spirit of the inventions disclosed herein.

[0109] It should be understood that the steps of the exemplary methods described herein do not necessarily have to be performed in the order described, and the order of steps in such methods should be understood to be merely illustrative. Similarly, additional steps may be included in such methods, and certain steps may be omitted or combined in a manner that is not inconsistent with the various examples of the invention.

[0110] The elements in the attached method claims are described in a specific order using corresponding labels, but unless the description in the claims otherwise implies a specific order for carrying out some or all of these elements, these elements are not necessarily intended to be limited to being carried out in that specific order.

[0111] It will be understood that any singular reference ("a" or "one") in this specification to describe a feature such as a component or step does not preclude the presence of additional features or multiple features. For example, a reference to a device having or defining "one" of a certain feature does not preclude the device having or defining two or more of that feature, insofar as the device has or defines at least one of that feature. Similarly, a reference in this specification to "one of" multiple features does not preclude the invention from including two or more, or even all, of those features. For example, a reference to a device having or defining "one of a X and Y" does not preclude the device having both X and Y.

Claims

1. A system configured to determine the presence of an abnormality related to the fluid supply section of a dispenser, wherein the system A fluid supply unit for containing the fluid, Tools for supporting the parts, One or more sensors configured to detect the characteristics of the fluid supply unit, A controller configured to activate one or more sensors to acquire initial data related to the fluid supply unit at a first time point, Equipped with, The controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to the fluid supply unit. The controller is further configured to log the presence of the abnormality, The controller is further configured to generate an alert indicating the presence of the anomaly and to transmit it to a human-machine interface, the alert providing the user with location information regarding the anomaly and / or instructions for correcting the anomaly. The controller is further configured to activate one or more sensors to acquire subsequent data related to the fluid supply unit at a second time point after the first time point, The system further comprises a controller configured to process the subsequent data and determine whether the anomaly has been corrected.

2. The one or more sensors include a camera, The aforementioned controller, The camera is activated to capture an initial image of the fluid supply unit at the first time point, The camera is activated to capture a subsequent image of the fluid supply unit at the second time point, The system according to claim 1, configured to perform the following:

3. The system according to claim 2, wherein the abnormality is the fluid level in the fluid supply unit.

4. The aforementioned controller, Using the trained neural network, the fluid surface height within the fluid supply unit is categorized as high, medium, or low. The system according to claim 3, configured as described above.

5. The system according to claim 3, wherein the trained neural network is trained on at least one image of a fluid supply unit having a fluid surface height above a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

6. The system according to claim 3, wherein the alert indicates that the fluid surface height in the fluid supply unit is below a predetermined threshold.

7. The system according to claim 6, wherein the alert provides the user with instructions to replenish the fluid supply unit so that the fluid surface height is equal to or greater than a predetermined threshold.

8. The system according to claim 7, wherein the alert provides the user with instructions to replace the fluid supply unit with a second fluid supply unit having a fluid surface height equal to or greater than a predetermined threshold.

9. The aforementioned controller, Based on the amount of fluid discharged by the dispenser, the predicted height of the fluid in the fluid supply unit is determined, Based on the initial image of the fluid supply unit, the detection height of the fluid within the fluid supply unit is determined. The deviation is identified by comparing the predicted height of the fluid in the fluid supply unit with the detected height of the fluid in the fluid supply unit. The system according to claim 2, configured to perform the following:

10. The system according to claim 1, wherein the trained neural network is trained on data related to a plurality of fluid supply units, and the controller uses the trained neural network to determine the presence of the abnormality related to the fluid supply unit based on the classification of the initial data.

11. It also has a discharge head, The system according to claim 1, wherein the controller is further configured to transmit a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire the initial data.

12. The system according to claim 11, wherein, if the controller determines the presence of the abnormality, the controller is further configured to transmit a signal to the discharge head to cause the discharge head to stop discharging the fluid from the discharge head.

13. The system according to claim 12, wherein, if the controller determines that the abnormality has been corrected, the controller is further configured to send a signal to the discharge head to cause the discharge of the fluid from the discharge head to resume.

14. A system configured to determine the presence of an abnormality related to the tools of a dispenser, wherein the system A fluid supply unit for containing the fluid, Tools for supporting the parts, One or more sensors configured to detect the characteristics of the tool, A controller configured to activate one or more sensors to acquire initial data related to the tool at a first point in time, Equipped with, The controller is further configured to process the initial data in a trained neural network to determine the presence of anomalies related to the tool. The controller is further configured to log the presence of the abnormality, The controller is further configured to generate and transmit an alert indicating the presence of the anomaly to a human-machine interface, the alert providing the user with location information relating to the anomaly and / or instructions for correcting the anomaly. The controller is further configured to activate one or more sensors to capture subsequent data related to the tool at a second time point following a first time point. The system further comprises a controller configured to process the subsequent data and determine whether the anomaly has been corrected.

15. The one or more sensors include a camera, The aforementioned controller, The camera is activated to capture an initial image of the tool at a first point in time, The camera is activated to capture a subsequent image of the tool at the second time point, The system according to claim 14, configured to perform the following:

16. The system according to claim 15, wherein the abnormality is the fluid accumulation height on the tool.

17. The system according to claim 16, wherein the trained neural network is trained on at least one image of a tool having a fluid accumulation height below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

18. The system according to claim 17, wherein the alert indicates that the fluid accumulation height is greater than the predetermined threshold.

19. The system according to claim 18, wherein the alert provides the user with instructions to clean the tool so that the fluid accumulation height falls below a predetermined threshold.

20. The aforementioned controller, Based on the number of parts removed from the tool, the predicted height of fluid accumulation on the tool is determined, Based on the initial image of the tool, the detection height of fluid accumulation on the tool is determined, The deviation is identified by comparing the predicted height of the fluid accumulation on the tool with the detected height of the fluid accumulation on the tool. The system according to claim 15, configured to perform the following:

21. The system according to claim 14, wherein the trained neural network is trained on data related to a plurality of tools, and the controller uses the trained neural network to determine the presence of the anomaly related to the tools based on the classification of the initial data into data.

22. The system according to claim 14, wherein the controller is configured to repeatedly capture and process the initial data and the subsequent data each time a part is removed from the tool.

23. It also has a discharge head, The system according to claim 14, wherein the controller is further configured to transmit a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire the initial data.

24. The system according to claim 23, wherein, if the controller determines the presence of the abnormality, the controller is further configured to transmit a signal to the discharge head to cause the discharge head to stop discharging the fluid from the discharge head.

25. The system according to claim 24, wherein, if the controller determines that the abnormality has been corrected, the controller is further configured to send a signal to the discharge head to cause the discharge of the fluid from the discharge head to resume.

26. A system configured to determine the presence of an abnormality related to the fluid supply section and / or tools of a dispenser, wherein the system A fluid supply unit for containing the fluid, Tools for supporting the parts, One or more sensors configured to detect the characteristics of at least one of the fluid supply unit and the tool, A controller configured to activate one or more sensors to acquire initial data related to at least one of the fluid supply unit and the tool at a first time point, Equipped with, The controller is further configured to process the initial data in a trained neural network to determine the presence of an anomaly related to at least one of the fluid supply unit and the tool. The controller is further configured to log the presence of the abnormality, The controller is further configured to generate and transmit an alert indicating the presence of the anomaly to a human-machine interface, the alert providing the user with location information relating to the anomaly and / or instructions for correcting the anomaly. The controller is further configured to activate one or more sensors to capture subsequent data related to at least one of the fluid supply unit and the tool at a second time point after the first time point, The system further comprises a controller configured to process the subsequent data and determine whether the anomaly has been corrected.

27. It also has a discharge head, The system according to claim 26, wherein the controller is further configured to transmit a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire the initial data.

28. The system according to claim 27, wherein, if the controller determines the presence of the abnormality, the controller is further configured to transmit a signal to the discharge head to cause the discharge head to stop discharging the fluid from the discharge head.

29. The system according to claim 28, wherein, if the controller determines that the abnormality has been corrected, the controller is further configured to send a signal to the discharge head to cause the discharge of the fluid from the discharge head to resume.

30. A process for determining the presence of an abnormality related to the fluid supply section of a dispenser, One or more sensors are configured to detect the characteristics of the fluid supply unit, The controller is configured to activate one or more of the aforementioned sensors to acquire initial data related to the fluid supply unit at a first time point in time, The controller is configured to process the initial data in a trained neural network to determine the presence of an anomaly related to the fluid supply unit, The controller is configured to log the existence of the aforementioned abnormality, Configuring the controller to generate an alert indicating the presence of the anomaly and to transmit it to a human-machine interface, wherein the alert provides the user with location information relating to the anomaly and / or instructions for correcting the anomaly. The controller is configured to activate one or more of the sensors to acquire subsequent data related to the fluid supply unit at a second time point after the first time point, The controller is configured to process the subsequent data and determine whether the anomaly has been corrected. A process that includes this.

31. The one or more sensors include a camera, The aforementioned controller, The camera is activated to capture an initial image of the fluid supply unit at the first time point, The camera is activated to capture a subsequent image of the fluid supply unit at the second time point, The process according to claim 30, configured to perform the following:

32. The process according to claim 31, wherein the abnormality is the fluid level in the fluid supply unit.

33. The aforementioned controller, Using the trained neural network, the fluid surface height within the fluid supply unit is categorized as high, medium, or low. The process according to claim 32, configured as follows.

34. The process according to claim 32, wherein the trained neural network is trained on at least one image of a fluid supply unit having a fluid surface height above a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

35. The process according to claim 32, wherein the alert indicates that the fluid surface height in the fluid supply unit is less than the predetermined threshold.

36. The process according to claim 35, wherein the alert provides the user with instructions to replenish the fluid supply unit so that the fluid surface height is equal to or greater than a predetermined threshold.

37. The process according to claim 36, wherein the alert provides the user with instructions to replace the fluid supply unit with a second fluid supply unit having a fluid surface height equal to or greater than a predetermined threshold.

38. The aforementioned controller, Based on the amount of fluid discharged by the dispenser, the predicted height of the fluid in the fluid supply unit is determined, Based on the initial image of the fluid supply unit, the detection height of the fluid within the fluid supply unit is determined. The deviation is identified by comparing the predicted height of the fluid in the fluid supply unit with the detected height of the fluid in the fluid supply unit. The process according to claim 31, configured to perform the following:

39. The process according to claim 30, wherein the trained neural network is trained on data related to a plurality of fluid supply units, and the controller uses the trained neural network to determine the presence of the anomaly related to the fluid supply unit based on the classification of the initial data.

40. The process according to claim 30, further comprising configuring the controller to transmit a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire the initial data.

41. The process according to claim 40, further comprising configuring the controller to transmit a signal to the discharge head to stop discharging the fluid from the discharge head when the controller determines that the presence of the abnormality is present.

42. The process according to claim 41, further comprising configuring the controller to send a signal to the discharge head to restart the discharge of the fluid from the discharge head when the controller determines that the abnormality has been corrected.

43. A process for determining the presence of an abnormality related to the dispenser tool, One or more sensors are configured to detect the characteristics of the tool, The controller is configured to activate one or more of the aforementioned sensors to acquire initial data related to the tool at a first point in time, The controller is configured to process the initial data in a trained neural network to determine the presence of anomalies related to the tool, The controller is configured to log the existence of the aforementioned abnormality, Configuring the controller to generate an alert indicating the presence of the anomaly and to transmit it to a human-machine interface, wherein the alert provides the user with location information relating to the anomaly and / or instructions for correcting the anomaly. The controller is configured to activate one or more of the sensors to capture subsequent data related to the tool at a second time point after the first time point, The controller is configured to process the subsequent data and determine whether the anomaly has been corrected. A process that includes this.

44. The one or more sensors include a camera, The aforementioned controller, The camera is activated to capture an initial image of the tool at a first point in time, The camera is activated to capture a subsequent image of the tool at the second time point, The process according to claim 43, configured to perform the following:

45. The process according to claim 44, wherein the abnormality is the fluid accumulation height on the tool.

46. The process according to claim 45, wherein the trained neural network is trained on at least one image of a tool having a fluid accumulation height below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

47. The process according to claim 46, wherein the alert indicates that the fluid accumulation height is greater than the predetermined threshold.

48. The process according to claim 47, wherein the alert provides the user with instructions to clean the tool so that the fluid accumulation height falls below a predetermined threshold.

49. The aforementioned controller, Based on the number of parts removed from the tool, the predicted height of fluid accumulation on the tool is determined, Based on the initial image of the tool, the detection height of fluid accumulation on the tool is determined, The deviation is identified by comparing the predicted height of the fluid accumulation on the tool with the detected height of the fluid accumulation on the tool. The process according to claim 44, configured to perform the following:

50. The process according to claim 43, wherein the trained neural network is trained on data related to a plurality of tools, and the controller uses the trained neural network to determine the presence of the anomaly related to the tools based on the classification of the initial data.

51. The process according to claim 43, wherein the controller is configured to repeatedly capture and process the initial data and the subsequent data each time a part is removed from the tool.

52. The process according to claim 43, further comprising configuring the controller to transmit a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire the initial data.

53. The process according to claim 52, further comprising configuring the controller to transmit a signal to the discharge head to stop discharging the fluid from the discharge head when the controller determines that the presence of the abnormality is present.

54. The process according to claim 53, further comprising configuring the controller to send a signal to the discharge head to restart the discharge of the fluid from the discharge head when the controller determines that the abnormality has been corrected.

55. A process for determining the presence of an abnormality related to the fluid supply section and / or tools of a dispenser, One or more sensors are configured to detect the characteristics of at least one of the fluid supply unit and the tool. The controller is configured to activate one or more of the aforementioned sensors to acquire initial data related to at least one of the fluid supply unit and the tool at a first time point in time. The controller is configured to process the initial data in a trained neural network to determine the presence of an anomaly related to at least one of the fluid supply unit and the tool, The controller is configured to log the existence of the aforementioned abnormality, Configuring the controller to generate an alert indicating the presence of the anomaly and to transmit it to a human-machine interface, wherein the alert provides the user with location information relating to the anomaly and / or instructions for correcting the anomaly. The controller is configured to activate one or more of the sensors to acquire subsequent data related to at least one of the fluid supply unit and the tool at a second time point after the first time point, The controller is configured to process the subsequent data and determine whether the anomaly has been corrected. A process that includes this.

56. The process according to claim 55, further comprising configuring the controller to transmit a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire the initial data.

57. The process according to claim 56, further comprising configuring the controller to transmit a signal to the discharge head to stop discharging the fluid from the discharge head when the controller determines that the presence of the abnormality is present.

58. The process according to claim 57, further comprising configuring the controller to send a signal to the discharge head to restart the discharge of the fluid from the discharge head when the controller determines that the abnormality has been corrected.

59. A system configured to determine the presence of an abnormality related to the dispensing system components of a dispenser, wherein the system A fluid supply unit for containing the fluid, Tools for supporting the parts, One or more sensors configured to detect the characteristics of the aforementioned discharge system components, A controller configured to activate one or more of the aforementioned sensors to acquire initial data related to the discharge system components at a first time point, Equipped with, The controller is further configured to process the initial data in a trained neural network to determine the presence of anomalies related to the discharge system components. The controller is further configured to log the presence of the abnormality, The controller is further configured to generate an alert indicating the presence of the anomaly and to transmit it to a human-machine interface, the alert providing the user with location information regarding the anomaly and / or instructions for correcting the anomaly. The controller is further configured to activate one or more sensors to capture subsequent data related to the discharge system components at a second time point following the first time point. The system further comprises a controller configured to process the subsequent data and determine whether the anomaly has been corrected.

60. The system according to claim 59, wherein the controller determines the presence of the anomaly using the trained neural network based on training using data related to a plurality of discharge system components.

61. The one or more sensors include a camera, The aforementioned controller, The camera is activated to capture an initial image of the dispensing system components of the dispenser at a first time point. The camera is activated to capture subsequent images of the dispensing system components of the dispenser at the second time point, The system according to claim 59, configured to perform the following:

62. The system according to claim 61, wherein the anomaly is the fluid surface height within the fluid supply unit, the trained neural network is trained on at least one image of the fluid supply unit having a fluid surface height above a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

63. The system according to claim 61, wherein the anomaly is the fluid accumulation height on the tool, the trained neural network is trained on at least one image of the tool having a fluid accumulation height below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

64. The system according to claim 61, wherein the abnormality is the presence of foreign matter in the working area of ​​the dispenser.

65. The system according to claim 64, wherein the trained neural network is trained on at least one image of the work area of ​​the dispenser which is free of foreign matter, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

66. The system according to claim 61, wherein the abnormality is at least one of a full purge cup and a soiled cleaning strip in the service station of the dispenser.

67. The system according to claim 66, wherein the trained neural network is trained on at least one image of at least one of the unfilled purge cup and the unstained cleaning strip of the service station of the dispenser, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial images.

68. The system according to claim 61, wherein the abnormality is the misloading of the component on the tool.

69. The system according to claim 68, wherein the trained neural network is trained based on at least one image of a part properly placed on the tool, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

70. The system according to claim 61, wherein the abnormality is the incorrect position of the bubble on the spirit level of the dispenser.

71. The system according to claim 70, wherein the trained neural network is trained on at least one image of a properly positioned bubble on the spirit level of the dispenser, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

72. The system according to claim 61, wherein the abnormality is an unauthorized user of the dispenser.

73. The system according to claim 72, wherein the trained neural network is trained on at least one of the authorized user's face recognition, biometric information, and unique badge, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

74. The system according to claim 61, wherein the abnormality is a characteristic of the discharge system component selected from the group consisting of leakage related to the discharge system component, damage or wear related to the discharge system component, and blockage or interruption related to the discharge system component.

75. The system according to claim 74, wherein the trained neural network is trained on at least one image of a discharge system component that is free from leaks, damage or wear and / or blockages or interruptions, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

76. The system according to claim 61, wherein the abnormality is the temperature and / or LED status of the discharge system components.

77. The system according to claim 76, wherein the trained neural network is trained on at least one image of the temperature and / or LED status of a dispensing system component having a temperature and / or operational LED status below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the abnormality based on the classification of the temperature and / or LED status of the dispensing system component.

78. The system according to claim 77, wherein the one or more sensors include an infrared camera configured to determine the temperature and / or the LED status of the discharge system components.

79. It also has a discharge head, The system according to claim 59, wherein the controller is further configured to transmit a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire the initial data.

80. The system according to claim 79, wherein, if the controller determines the presence of the abnormality, the controller is further configured to transmit a signal to the discharge head to cause the discharge head to stop discharging the fluid from the discharge head.

81. The system according to claim 80, wherein, if the controller determines that the abnormality has been corrected, the controller is further configured to send a signal to the discharge head to cause the discharge of the fluid from the discharge head to resume.

82. A process for determining the presence of an abnormality related to the dispensing system components of a dispenser, One or more sensors are configured to detect the characteristics of the aforementioned discharge system components, The controller is configured to activate one or more of the aforementioned sensors to acquire initial data related to the discharge system components at a first time point in time, The controller is configured to process the initial data in a trained neural network to determine the presence of anomalies related to the discharge system components, The controller is configured to log the existence of the aforementioned abnormality, Configuring the controller to generate an alert indicating the presence of the anomaly and to transmit it to a human-machine interface, wherein the alert provides the user with location information relating to the anomaly and / or instructions for correcting the anomaly. The controller is configured to activate one or more of the sensors to acquire subsequent data related to the discharge system components at a second time point after the first time point, The controller is configured to process the subsequent data and determine whether the anomaly has been corrected. A process that includes this.

83. The process according to claim 82, wherein the controller determines the presence of the anomaly using the trained neural network based on training with data related to a plurality of discharge system components.

84. The one or more sensors include a camera, The aforementioned controller, The camera is activated to capture an initial image of the dispensing system components of the dispenser at a first time point. The camera is activated to capture subsequent images of the dispensing system components of the dispenser at the second time point, The process according to claim 82, configured to perform the following:

85. The process according to claim 84, wherein the anomaly is the fluid surface height within the fluid supply unit, the trained neural network is trained on at least one image of the fluid supply unit having a fluid surface height greater than or equal to a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

86. The process according to claim 84, wherein the anomaly is a fluid accumulation height on a tool, the trained neural network is trained on at least one image of a tool having a fluid accumulation height below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

87. The process according to claim 84, wherein the abnormality is the presence of foreign matter in the working area of ​​the dispenser.

88. The process according to claim 87, wherein the trained neural network is trained on at least one image of the work area of ​​the dispenser which is free of foreign matter, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

89. The process according to claim 84, wherein the abnormality is at least one of a full purge cup and a soiled cleaning strip in the service station of the dispenser.

90. The process according to claim 89, wherein the trained neural network is trained on at least one image of at least one of the unfilled purge cup and the unstained cleaning strip of the service station of the dispenser, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial images.

91. The process according to claim 84, wherein the abnormality is the misloading of a part on the tool.

92. The process according to claim 91, wherein the trained neural network is trained based on at least one image of a part properly placed on the tool, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

93. The process according to claim 84, wherein the abnormality is the incorrect position of the bubble on the spirit level of the dispenser.

94. The process according to claim 93, wherein the trained neural network is trained on at least one image of a properly positioned bubble on the spirit level of the dispenser, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

95. The process according to claim 84, wherein the abnormality is an unauthorized user of the dispenser.

96. The process according to claim 95, wherein the trained neural network is trained on at least one of the authorized user's face recognition, biometric information, and unique badge, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the initial image.

97. The process according to claim 84, wherein the abnormality is a characteristic of the discharge system component selected from the group consisting of leakage related to the discharge system component, damage or wear related to the discharge system component, and blockage or interruption related to the discharge system component.

98. The process according to claim 97, wherein the trained neural network is trained on at least one image of a discharge system component that is free from leaks, damage or wear, and / or blockages or interruptions, and the controller is configured to use the trained neural network to determine the presence of an anomaly based on the classification of the initial images.

99. The process according to claim 84, wherein the abnormality is the temperature and / or LED status of the discharge system component.

100. The process according to claim 99, wherein the trained neural network is trained on at least one image of the temperature and / or LED status of a dispensing system component having a temperature and / or operable LED status below a predetermined threshold, and the controller is configured to use the trained neural network to determine the presence of the anomaly based on the classification of the temperature and / or LED status of the dispensing system component.

101. The process according to claim 99, wherein the one or more sensors include an infrared camera configured to determine the temperature and / or the LED status of the discharge system components.

102. The process according to claim 82, further comprising configuring the controller to transmit a signal to the discharge head to cause fluid to be discharged from the discharge head before activating one or more sensors to acquire the initial data.

103. The system according to claim 102, further comprising configuring the controller to transmit a signal to the discharge head to stop the discharge of the fluid from the discharge head when the controller determines that the presence of the abnormality is present.

104. The system according to claim 103, further comprising configuring the controller to send a signal to the discharge head to restart the discharge of the fluid from the discharge head when the controller determines that the abnormality has been corrected.

Images