OPERATIONAL DATA CAPTURE

FIELD OF THE INVENTION

This invention in general relates to handlooms, and specifically relates to authenticating a handloom machine.

BACKGROUND OF THE INVENTION

Handloom fabrics represent the rich culture and tradition of a particular location. These fabrics signify the heritage incorporated from generations in the making of fabrics and the techniques which are passed down traditionally. Making of a handloom fabric requires artistic skills, patience and dedication on the part of the skilled weavers for the manifestation of the final fabric product.

Handloom fabrics are created on traditional looms like the pit looms or the frame looms, which are located at the weaver’s homes or the traditional weaver workplaces. These are hand operated tools which require energy input from the weaver for the operation. Therefore, these looms operate at a pace set by the weaver. As a result, the creation of a handloom fabric demands a lot of time from the skilled weaver. Due to these factors the handloom fabric or handwoven designs are rare, unique and highly priced.

On the other hand, power looms are electrically coupled devices which are organized for quick operation and mass production of fabric. The fabric produced by the power looms does not require much human inputs and skills. Due to the artistic attributes and weaver-mediated skills involved in the production of handloom products, these products are unique and are highly in demand in the market. As a result of the quality associated with the handloom products, a lot of counterfeit products manufactured on power looms are sold in the market as handloom products. This is detrimental to the livelihood of the skilled weavers because of the uniqueness and artistic inputs they incorporate in every product.

Also, it is difficult for a non-expert customer to distinguish between a handloom and a power loom product. The customer has to rely on the seller for the authenticity of the product and the probability of falling into a false trap is high. In view of the counterfeit market linked to the handloom products, there is a need to identify the authenticity of the handloom products. Anti-counterfeit systems for detection of counterfeit products are known in the art.

Chinese patent application CN110414635A to Huizhou University describes an anti-counterfeiting and tracing system for authentication of product. Chinese Patent discloses an (RFID)-based product counterfeiting and traceability system. The user query subsystem includes a user terminal, a (RFID) tag, an interactive query module for the user to check the authenticity of a product and trace the production information. The user terminal is a mobile phone or an electronic computer or a smart wearable device. US patent US2015370333A1 to THALMIC LABS INC discloses a wearable gesture identification device including a band (worn on the limb of a user) having a sensor, processor, non-transitory processor-readable storage medium. An accelerometer, a gyroscope, and an inertial measurement unit (IMU) are also components of the gesture recognition device. The gesture identification system includes a minimum of one sensor responsive to detect, sense, measure, or transduce a physical gesture performed by a user of the gesture identification system.

These systems are used to authenticate products and detect counterfeit items. But the uniqueness of a handloom product is attributed to the loom on which it is produced and the skilled weaver involved in the production. Therefore, there is a further need to authenticate and authorize the identity of the skilled weaver and identify the origin of the product like the handloom on which it is created.

A major distinguishing factor between a power loom and a handloom product is the speed at which the fabric is woven. Power looms operate at higher speed as compared to handlooms. Therefore, the average picks per unit time for handloom is less as compared to that of power looms. Also, the beats that a handloom produces during its working are unique to that handloom and the weaver who is working on the handloom.

Therefore, there is a need to identify the loom operation characteristics like beat detection, vibrations, etc. that are specific to the working of a particular handloom. Also, there is a need to create a handloom profile which is specific to a particular handloom which can be stored and easily retrieved. There is a further need to identify the weaver associated with a particular handloom product and validate the timeframe in which a particular skilled article was made. An accurate validation of handlooms is essential to reduce the counterfeit handloom products. Since handlooms are installed in geographically remote locations, it presents additional challenges for customers and intermediate entities to identify the handloom products.

SUMMARY OF THE INVENTION

A system for authenticating a handloom machine includes a field detection device coupled to the handloom machine that further includes a sensor module configured to detect inputs to the system, a processor that is configured to process the inputs received from the sensor module and processing the signals, a weaver identifier detection module for capturing the unique identifier of the weaver and an analytics module residing in said cloud that generates a loom profile of said machine, wherein the said loom profile classifies the machine as a handloom or power loom and the subtypes of the loom.

The sensor module is advantageously one or more of vibration sensor, sound sensor, motion sensor, proximity sensors, magnetic sensors, image and video sensors, location sensors and time stamps. The vibration sensor captures the mechanical vibrations from loom. The motion and proximity sensor captures the moving patterns of said loom machine parts including the shuttle and beater during loom operation.

The analytics module residing in the cloud also includes a raw data module, a database and an artificial intelligence and machine learning components.

The method of authenticating a machine as a handloom or a powerloom wherein authenticating the machine includes measuring the variation of vibration and sound data, applying a statistical method and/or a machine learning process to analyze the time-series and frequency vibration and sound data, classifying said data and providing an output for distinguishing the machine. The method of authenticating a machine as a handloom or a powerloom wherein authenticating the machine includes analyzing the images captured during loom operation to identify hand movement and type of the loom.

A method of profiling a weaver operating a handloom includes the steps of determining the total number of picks inserted in a predetermined time and/or a predetermined length of fabric, a second step of determining the total productive hours as the total hours the loom was active divided by the total available hours, a next step of determining the productivity as the number of average picks per unit time, a next step of determining the skill of weaver based on the variation characteristics of the beats and a final step of analyzing said beat data; and by using machine learning algorithms, the unique pattern of weaving for each weaver is identified.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing summary, as well as the following detailed description of the invention, is better understood when read in conjunction with the appended drawings. For illustrating the invention, exemplary constructions of the invention are shown in the drawings. However, the invention is not limited to the specific components disclosed herein. The description of a component referenced by a numeral in a drawing is applicable to the description of that component shown by that same numeral in any subsequent drawing herein.

FIG. 1 shows system for profiling a handloom in accordance with the present invention;

FIG. 2 shows a cloud supporting multiple field devices in accordance with the present invention;

FIG. 3 shows a process steps involved in creating a loom profile and validating the handloom process using primary sensors for vibrations and sound; FIG. 4 shows a schematic diagram of the loom;

FIG. 5 shows the process of beat detection;

FIG. 6 shows the sample raw data from the loom;

FIG. 7 shows the data after smoothening;

FIG. 8 shows the beat threshold; and

FIG. 9 shows exemplary embodiment for authenticating a handmade product in accordance with the present invention.

DETAILED DESCRIPTION OF THE INVENTION

The invention described herein is explained using specific exemplary details for better understanding. However, the invention disclosed can be worked on by a person skilled in the art without the use of these specific details.

References in the specification to "one embodiment" or "an embodiment" means that particular feature, structure, characteristic, or function described in connection with the embodiment is included in at least one embodiment of the invention. The appearances of the phrase “in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment.

References in the specification to “preferred embodiment” means that a particular feature, structure, characteristic, or function described in detail hereby omitting known constructions and functions for clear description of the present invention. The foregoing description of specific embodiments of the present invention has been presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the present invention to the precise forms disclosed, and obviously many modifications and variations are possible in light of the above teaching.

FIG. 1 shows a system 100 for profiling a handloom machine 101 in accordance with the present invention. Accordingly, the system 100 includes a field device 107 that is connected to a remote environment for example cloud environment 103. The field device 107 includes a sensor module 102, a processor 104 and weaver identifier detector module 106. The sensor module 102 includes a plurality of sensors for example, audio sensors, camera, proximity sensor, location sensor, and time module. The sensors of the sensor module 102 are configured to detect inputs to the system 100 that are processed by the processor 104.

The field device 107 includes the processor 104 and the sensor module 102 such that the field device 107 is installed on the handloom machine 101. The field device 107 is also connected to the cloud 103 through a network and periodically transmits the sensor data to the cloud 103 in the network that is preferably a secured network. The sensor module 102 is a combination of primary sensors and secondary sensors that include audio, vibration, proximity, image, video and time stamp. It is known in the art that handlooms 101 are not manufactured based on any standardized designs or guidelines and hence one or more of the aforementioned sensors are used to collect information from respective handlooms.

In accordance with the present invention, the processor 104 sends communicates with the sensor module 102 to interpret the respective inputs. It is noted that the action trigger is periodic or as per inputs from one of the sensors. The field device 107 performs the initial validation of data and uploads to a cloud storage location or database. In case of no connectivity to the network, the field device 107 stores data locally until the network is available again and then forwards data.

Each of the weavers 105 is assigned a Unique Identification Code, hereinafter referred as a Unique ID. The Unique ID detection module 106 captures the unique ID of the weaver 105 and/or customer reference or a design. The unique ID helps to link all operational data to the weaver 105. Weaver 105, customer, or design is associated with a unique ID. The unique ID is creatable and readable using bar code or NFC or Bluetooth. The weaver 105 presents the unique ID to the field device 107. The field device 107 captures the Unique ID details and tags that along with the product and stored in the data base.

The processor 104 is configured to process the inputs received from the sensors and process the signals. The field device receives a weaver ID that is validated by the sensor module and received by the processor. It is noted that the processor 104 advantageously creates a handloom profile in accordance with the present invention. The inputs received from the loom 101 and weaver 105 are used by the processor to create the handloom profile. In accordance with the present invention, the weaver 105 is uniquely identified using a unique identification code that is stored using any electronic means such as such as barcode, RFID (Radio Frequency Identification) tag or Bluetooth tag or biometrics.

It is noted that the inputs needed to detect the type of the loom are captured from the loom using the sensor module 102 of the field device 107. The processor analyses the inputs using the modules in the cloud environment. These inputs are the stored and analyzed in the data base of the cloud environment 103. The unique loom profile is generated as an output from the analysis. In accordance with the present invention, the primary sensors include vibration sensor, sound sensor, and motion or proximity sensor. The vibration sensor captures the mechanical vibrations from loom. The motion or proximity sensor captures the moving patterns of a loom parts including shuttle and beater during the loom operation. The secondary sensors include image and video sensors, location sensors and time stamp. The camera captures images and videos. The GPS location is captured to validate the location of loom. The time of loom action is captured using a clock on the field device 107. The secondary sensors are used to augment findings from the analysis of primary sensor inputs.

FIG. 2 shows the cloud 103 communicating with multiple field devices 107. The cloud 103 includes modules such as raw data module, database module, and machine learning module. The machine learning module has artificial intelligence module, machine learning module, deep learning module, analytics module and other codes. The raw data module has sensor inputs from the field device 107 that are received in the form of files or objects. Cloud based storage solutions are used for the raw data module. The raw data from the cloud storage is transformed and stored in the database for analysis. The Machine Learning Modules learn from the input data, classify, and features are generated.

FIG. 3 shows steps involved in the process of creating a loom profile, and validating the handloom in accordance with the present invention. Accordingly, the process of creating loom profile and the validating process is performed using primary sensors for vibrations and sound. The process of detection of the state of the loom includes following activities. The loom has two states namely active state and idle state. When the weaving activity starts on the loom machine 101, the loom-state is considered to be ‘active’. Using the inferences from the beat detection process, the loom is said to be ‘active ’ when a predetermined number of beats are recorded in a predetermined time, referred to as threshold beat rate, for example, two beats per minute. The loom-state is considered ‘idle’ when the number of beats recorded are below the threshold beat rate. The loom-state is determined by data from the primary sensors. The loom-state is used to activate various modules of the field device, such as to activate other sensors and processes, optimize power consumption, memory and other resources. On detecting an active loom-state, the processor 104 triggers capturing signals from Motion sensor, visual signals through a camera sensor optionally. This is done, on a periodic basis or on random basis.

The processor 104 advantageously stores the loom characteristics in internal storage. The internal storage is a persistent data that saves the loom characteristics temporarily until it is uploaded. The processor initiates the cloud upload activity on a periodic basis. The raw data from the field device is stored as data files in a cloud location. The cloud 103 initiates data extract from data files and loads to a relational or object database for further analysis. The input raw data is further processed to reduce the harmonic components. The loom identification step comprises identifying the loom based on the threshold number of beats per minute, motion variability, location, images and video analysis. The authentication step refers to the method to confirm if the specific product is woven on a specified loom/handloom or shuttle looms at a specific location. The authentication step is executed using the following methods in isolation or in a combination.

The step of inferring from the threshold number of beats per minute is described herein. Powerlooms are capable of running at higher speeds than handlooms. For each of the handlooms a threshold beats per minute will be defined. Higher beats than the threshold number indicates that the loom on which fabric is produced is a powerloom. For example, if the beats per minute are more than 150, it’s a machine-driven powerloom. The step of inferring based on motion variability is described herein. Due to the machine driven motion, the sensor data recorded on a powerloom has a different pattern than the pattern of sensor data on a handloom machine 101. This variability of data depends on various factors such as type of weave, type of loom, and attachments. The motion of machine driven looms have less variability compared to a handloom 101. For example, the coefficient of variation of vibration data was found to be high for handloom and low for a powerloom. Different movements result in motion variability in the frequency domain. A statistical analysis is done and/or machine learning process is used to analyze the time-series and frequency data, classify the data and provide an output that can distinguish the loom uniquely.

The step of image and video analysis, which is an optional process, is described herein. Images captured during handloom operation at a regular or random intervals should have hand movement. The images captured are analyzed manually or by image recognition software.

Once the loom is identified as handloom 101 or a powerloom, a notification is generated for the users. In addition, the method and system disclosed herein can also be used to identify powerlooms, shuttle looms, shuttle less looms, and other types of looms.

The picks estimation step is described herein. While ends per inch (EPI) remains same all throughout the length, picks per inch (PPI) varies across the length of fabric especially on handlooms. Number of beats is a closer representation of the number of picks inserted. Based on the number of beats, the total number of picks is estimated within a certain time range as well as for a pre-determined length.

Once a total number of picks and length of the fabric is known, the pick density may be calculated. Picks Per Inch (PPI) is one of the indicators of pick density, is calculated as follows. The average PPI is equal to the total picks per length of fabric in inches. Further, the variations in data corresponding to beats is used to estimate the variability of PPI along the length of fabric. Variation in PPI will indicate the quality of the fabric.

The weaver profiling is a combination of the characteristics of “Total productive hours”, productivity, skill level of weaver 105 and uniqueness of weaving, which can together be referred as “Weaver Fingerprinting”. The total productive hours (%) is equal to (Total hours the loom was active/Total available hours). The productivity is the number of average picks per/ unit time or number of products produced in a predetermined time. The skill level of weaver 105 is assessed based on the variability of beats, such as standard deviation, coefficient of variation, or by factoring in other statistical parameters. Uniqueness may be identified by analyzing the signature data related to lifting, picking, beating with respect to effort and speed to determine the weaver’s unique signature of weaving

“Product fingerprinting” is accomplished by analyzing the data; and by using machine learning algorithms, the unique pattern of weave and type/sub- type of loom for each weaver 105 is identified. Examples of subtype of looms include plain weave, dobby and jacquard weave. Supply chain traceability is enabled by the information gathered from the labels. Labels are attached at a predetermined place on the fabric, typically at the start or towards the end of the product. The label is scanned at the time of weaving for each product. The number of labels scanned is counted to derive number of products produced remotely. This provides the real time production status. This data is analyzed to provide the status on the quantity produced at a given point time as well as to forecast the time required to complete manufacturing of a specific quantity. Similarly, data on picks estimation is compiled for a set of products to prepare a report on both quantity and quality. This data can be utilized to manage vendor development and pricing. Further when the same label is scanned along the supply chain, the precise status from manufacturing to sale of the product, including location is determined.

FIG. 4 shows a schematic diagram of the handloom 101. The weaver 105 operates the loom 101 manually to weave fabrics. A set of warp yams 402 is let off from warp beam 401, these warp yarns 402 are passed through different sets of harnesses 403. Further warp 402 passes through a reed 404. Next to the reed 404, the weft carrier, that is generally a shuttle 405 operates. The following are the sub activities performed in the loom 101. Lifting 407 refers to lifting of set of warp yams 402 passing through harness 403 to create a shed 408. Picking refers to shuttle 405 carrying weft from one end to another end to insert weft yarn across the shed. Beating 409 refers to the movement of the beater 406 to push the weft yarn securely into place.

FIG. 5 shows the process of beat detection in accordance with the present invention. Each sub-activity in the process of beat detection creates specific signature data attributes on the data from primary and secondary sensors. These signature data attributes uniquely identify the sub-activity and are used to define the sub-activity while analyzing the data generated during the weaving activity. The following process steps of data collection 501, data smoothen- ing or filtering 502, beat data threshold setting 503 and beat detection 504 is followed to detect the beating activity of the loom.

In the beat detection 504 step, given the beat threshold, a beat detection algorithm is employed that works by observing for crossings of the threshold in the downward or upward direction. A beat is defined as occurring if there is a negative slope of the acceleration plot, when the acceleration curve crosses below the beat threshold. In case of data from motion detection, the beats are detected directly from the motion detected from the beater or as a proxy using the data from the shuttle movement. FIG. 6 shows the sample raw data from the loom. Vibration, sound, motion, location and timestamp of operation data is collected. The sensors fixed on or near the loom will capture the raw data from the loom. In FIG. 6 and FIG. 7, X axis represents time and Y axis represents amplitude.

FIG. 7 shows the data after smoothening. The data smoothening step is required for processing sound and vibration data. The raw data, i.e. the vibration and/or sound and/or motion data is smoothened by taking the moving average of readings to make the signal clear. The number of values considered would vary upon the individual loom. The number of values considered will be determined on experimental basis for each loom to arrive at the optimal signal.

FIG. 8 shows the beat threshold. The beat threshold 801 setting step comprises setting a threshold value for separating out frequencies related to beating action from other frequencies. This process is required for processing sound and vibration data.

Referring to FIG. 9, in an exemplary embodiment of the system for authentication of of handloom by operational data capture is discussed. Accordingly, an IOT Device 900 that is field device 107 is mounted on a handloom beating bar. Now, when the bar moves, motion of the bar is captured by the IOT device’s 900 motion sensor. The motion of the bar is sampled at fixed time intervals or as triggered by an application running on remote device 904. The IOT device 900 also captures any non-loom moments using magnetic sensors or IMU sensors for tamper detection. The IOT device 900 sends the motion data samples and tamper detection data with timestamp to the remote device 904.

The remote device 904 connects to IOT device 900 on periodic basis to collect data. The remote device 904 connects only to the IOT devices 900 allocated to a weaver. The remote device 904 allows the weaver to capture the time of production^ start and end time) and to attach a digital identity using a unique label for example QR Code. These operations are permissible only when the remote device 904 is close to the IOT device 900, generally less than 75 centrimeters. The remote device 904 also captures the location and timestamp of the operation using the inbuilt GPS sensors. The authenticity of the weaver is verified using the phone number of the weaver. The remote device 904 sends all the authentication and production data to a coud. The machine learning and statistical algorithms validate the motion samples by classifying them as handmade or otherwise. A Consumer the unique code code attached to the product, and is able to see the authentication information.

Referring to FIG. 1-8, in operation the field device 107 of the system 100 is mounted on the handloom 101 beating bar. The sensor module 102 of the field device 107 captures the motion of the bar. The sensor module 102 also captures any non-loom movements for tamper detection and communicates the same to the processor 104 for interpretation. The field device 107 performs initial validation of data and uploads to a cloud storage location or database. The data is sampled at fixed time intervals and to a mobile application.

The field device 107 sends the motion data samples and tamper detection data with timestamp to the mobile. The mobile connects to the field device 107 at periodic intervals to collect data. The system 100 using the unique identification code of the weaver 105 connects only to the allocated looms 101. The mobile application displays the time of operation, timestamp of operation and location of the loom.

All this data is also stored on the cloud environment 103 binded to the unique identification code of the weaver 105. The machine learning and statistical algorithms validate the samples by classifying them in categories such as handmade, machine made and the like. During sale, the consumer scans the unique code attached to the product to access the product data. The foregoing examples have been provided merely for explanation and are in no way to be construed as limiting of the system and method for validating that a machine is a handloom and not a powerloom counterfeiting as a handloom disclosed herein. While the system and method has been described with reference to particular embodiments, it is understood that the words, which have been used herein, are words of description and illustration, rather than words of limitation. Furthermore, although the system and method has been described herein with reference to particular means, materials, and embodiments, the system and method is not intended to be limited to the particulars disclosed herein; rather, the design and functionality of the system and method extends to all functionally equivalent structures and uses, such as are within the scope of the appended claims.

While particular embodiments are disclosed, it will be understood by those skilled in the art, having the benefit of the teachings of this specification, that the system and method disclosed herein is capable of modifications and other embodiments may be effected and changes may be made thereto, without departing from the scope and spirit of the system and method disclosed herein.