Pet house system

Abstract

A pet house system and a method for using the pet house system are provided. The method for using the pet house system according to an embodiment of the present invention comprises: a step in which a sensor unit acquires a pressure signal of a pet; a step in which a processor processes a signal measured through the sensor unit; and a step in which the processor analyzes the heart rate variability (HRV) included in the signal processing result to estimate the stress of the pet. The acquisition step involves acquiring pressure signals from a plurality of channels and selecting an optimal signal among the pressure signals of each channel. By doing so, functions for analyzing the pet's heart rate, respiratory rate, and stress, classifying sleep states, and analyzing body water and body fat are provided, thereby contributing to the health management of the pet. Furthermore, by utilizing sensors from multiple channels to collect sensor values ​​and selecting the optimal sensor value among them, it is possible to measure biosignals with high accuracy even in various postures. Additionally, by measuring the pet's heart rate variability (HRV), deriving peaks of heart rate and respiration, and correcting the heart rate and respiratory rate, the heart rate and respiratory rate can be measured with higher accuracy than conventional technology.

Description

Technology Field

[0001] The present invention relates to a pet house system for companion animals, and more specifically, to a pet house system capable of measuring biosignals. Background Technology

[0003] Ballistocardiogram (BCG) refers to a signal that measures the ballistics resulting from changes in blood flow in the heart and blood vessels according to the contraction and relaxation of the heart, and this ballistocardiogram (BCG) signal, together with the electrocardiogram (ECG) signal, represents information about the user's heartbeat as an electrical signal.

[0004] These ballistocardiogram (BCG) signals measure minute vibrations of the heartbeat; while they cannot represent the movement of the heart muscle as an electrical signal, they contain information such as heart rate per minute. Furthermore, BCG signals have the advantage of measuring the user's heart rate signal non-restrictive and unconsciously, as there is no need to attach special electrodes or sensors to the body.

[0005] Meanwhile, as the number of people interested in the health of pets has recently increased, technologies for measuring the biometric information of pets for health management, such as Korean Registered Patent No. 10-2352071, have emerged.

[0006] However, conventional technology has limitations as it is limited to collecting fragmentary biometric data or measuring the pet's weight based on it. Prior art literature

[0008] Korean Registered Patent No. 10-2352071 (Title of Invention: Biometric Information Measurement Device for Pets) Korean Registered Patent No. 10-1223889 (Title of Invention: Stress Analysis Device and Method) The problem to be solved

[0009] The present invention has been devised to solve the above-mentioned problems, and the objective of the present invention is to provide a pet house system capable of analyzing the heart rate, respiratory rate, and stress of a pet, classifying sleep status, and analyzing body water and body fat, and a method for using the pet house system.

[0010] Another objective of the present invention is to provide a pet house system and a method of using the pet house system that can collect sensor values ​​using multiple channels of sensors and select the optimal sensor value among them, thereby enabling high-accuracy biosignal measurement even in various postures, and can measure the heart rate variability (HRV) of a pet to derive peaks of heart rate and respiration and correct heart rate and respiration rate. means of solving the problem

[0012] A method for using a pet house system according to an embodiment of the present invention for achieving the above objective comprises: a step in which a sensor unit acquires a pressure signal of a pet; a step in which a processor processes a signal measured through the sensor unit; and a step in which the processor analyzes heart rate variability (HRV) included in the signal processing result to estimate the stress of the pet, wherein the acquisition step acquires a pressure signal from each of a plurality of channels and selects an optimal signal among the pressure signals of each channel.

[0013] And the sensor unit may include a multi-array piezoelectric sensor composed of a plurality of piezoelectric sensors embedded in the bottom of the pet house.

[0014] Additionally, the processor may include a filter circuit that performs a filtering operation on a pressure signal measured through multiple piezoelectric sensors; and an amplifier circuit that performs an amplification operation on the filtered signal.

[0015] In the acquisition step, when filtering and amplification operations are performed on pressure signals measured through multiple piezoelectric sensors, an optimal signal is selected from the signals that have undergone filtering and amplification operations. In the signal processing step, the selected optimal signal is separated into preset time units using a window function, a heart rate reference filter based on the heart rate frequency band of a normal pet is applied to the signals separated into preset time units to perform a Fast Fourier Transform (FFT), and a respiration reference filter based on the respiration frequency band of a normal pet is applied to the signals separated into preset time units to perform a Fast Fourier Transform (FFT).

[0016] In addition, the acquisition step involves, when selecting the optimal signal, obtaining the SNR value by calculating the ratio of the signal with the largest signal frequency band to the total signal band of the signal that has undergone filtering and amplification operations, calculating the peak value of the signal obtained through cross-correlation with the signals of each channel using the signal of the channel with the least noise as the reference signal, and calculating the average by normalizing the signal-to-noise ratio value and the cross-correlation peak value for each channel to be between 0 and 1, and selecting the channel signal with the highest calculated average as the optimal signal.

[0017] And, the method of using the pet house system according to the present embodiment may further include the step of a processor calculating the heart rate and respiratory rate of a pet per minute based on the signal processing result.

[0018] In addition, the step of calculating the heart rate and respiratory rate per minute can calculate the heart rate peak value within the heart rate frequency range and the respiratory rate peak value within the respiratory frequency range based on the signal processing results.

[0019] And in the step of estimating stress, a peak detection operation is performed on a time-domain signal to which a heart rate reference filter is applied to derive heart rate variability based on the variability of peak values, and stress can be estimated by performing frequency analysis of the heart rate variability.

[0020] In addition, the method of using the pet house system according to the present embodiment further includes a step in which a processor classifies the sleep state of a pet by applying a signal processing result to a CNN-LSTM model; and the step of classifying the sleep state may classify the sleep state of the pet into NREM (Non-REM sleep state), REM (REM sleep state), or WAKE (non-sleep state) by applying a signal processing result to a trained CNN-LSTM model.

[0021] Meanwhile, according to another embodiment of the present invention, a pet house system comprises: a sensor unit embedded in the bottom of the pet house to acquire a pressure signal of a pet; and a processor that processes the signal acquired through the sensor unit and analyzes the heart rate variability (HRV) included in the signal processing result to estimate the stress of the pet. The sensor unit acquires a pressure signal from each of a plurality of channels, and the processor selects and processes the optimal signal among the pressure signals of each channel.

[0022] A method of using a pet house system according to another embodiment of the present invention comprises: a step in which a sensor unit acquires an analog signal for measuring the weight and analyzing the body composition of a pet; a step in which a processor converts the analog signal acquired through the sensor unit into a digital signal; a step in which the processor corrects the digital signal to measure the weight; and a step in which the processor calculates a bioimpedance value based on the digital signal and measures the body composition of the pet based on the weight measurement result and the bioimpedance value.

[0023] In addition, a pet house system according to another embodiment of the present invention includes: a sensor unit embedded in the bottom of the pet house to acquire an analog signal for measuring the weight and analyzing the body composition of a pet; and a processor that converts the analog signal acquired through the sensor unit into a digital signal, corrects the digital signal to measure the weight, calculates a bioimpedance value based on the digital signal, and measures the body composition of the pet based on the weight measurement result and the bioimpedance value. Effects of the invention

[0025] As described above, according to the embodiments of the present invention, functions for analyzing the heart rate, respiratory rate, and stress of a pet, classifying sleep status, and analyzing body water and body fat are provided, thereby contributing to the health management of pets.

[0026] In addition, according to embodiments of the present invention, sensor values ​​can be collected using multiple channels of sensors and the optimal sensor value among them can be selected, thereby enabling high-accuracy biosignal measurement even in various postures. Furthermore, by measuring the heart rate variability (HRV) of a pet to derive peaks of heart rate and respiration and correcting heart rate and respiration rate, heart rate and respiration rate can be measured with higher accuracy than conventional technology. Brief explanation of the drawing

[0028] FIG. 1 is a diagram provided for the description of the configuration of a pet house system according to an embodiment of the present invention; FIG. 2 is a diagram illustrating the configuration of a pet house system according to an embodiment of the present invention in which multiple piezoelectric sensors, biosignal electrodes, and load cells are provided; FIG. 3 is a diagram provided for the description of the configuration of multiple pressure sensors according to an embodiment of the present invention; FIG. 4 is a flowchart provided for the description of a method of using a pet house system according to an embodiment of the present invention; FIG. 5 is a diagram illustrating the derivation of a signal peak value through cross-correlation with signals of each channel according to an embodiment of the present invention; FIG. 6 is a diagram provided for the description of the normalization and scoring process of the signal-to-noise ratio and cross-correlation according to an embodiment of the present invention; and FIG. 7 is a flowchart provided for the description of a method of measuring the weight and body composition of a pet using a pet house system according to an embodiment of the present invention. Specific details for implementing the invention

[0029] The present invention will be described in more detail below with reference to the drawings.

[0030] A pet house system according to one embodiment of the present invention can provide functions for analyzing the heart rate, respiratory rate, and stress of a pet, classifying sleep status, and analyzing body water and body fat.

[0031] In addition, a pet house system according to one embodiment can collect sensor values ​​using multiple channels of sensors and select the optimal sensor value among them to provide functions for analyzing the heart rate, respiratory rate, and stress of a pet, as well as classifying the sleep state. By measuring the heart rate variability (HRV) of the pet to derive peaks of heart rate and respiration and correcting the heart rate and respiratory rate, the heart rate and respiratory rate can be measured with higher accuracy than conventional technology.

[0032] FIG. 1 is a drawing provided for the description of the configuration of a pet house system according to one embodiment of the present invention, FIG. 2 is a drawing illustrating an example in which multiple piezoelectric sensors, biosignal electrodes, and load cells are provided in a pet house system according to one embodiment of the present invention, and FIG. 3 is a drawing provided for the description of the configuration of multiple pressure sensors according to one embodiment of the present invention.

[0033] Referring to FIG. 1, the pet house system may include a sensor unit (110), a processor (120), a communication unit (130), and a storage unit (140).

[0034] The sensor unit (110) can acquire the biometric information of the pet.

[0035] For example, the sensor unit (110) is embedded in the bottom of the pet house and can obtain a pressure signal of the pet, a bioelectrical impedance for measuring body composition (body fat and body water), and an impedance for measuring body weight.

[0036] To this end, the sensor unit (110) may include a multi-piezoelectric sensor (111), an electrode (112), and a load cell (113) embedded in the bottom of the pet house as illustrated in FIG. 2.

[0037] The multi-piezoelectric sensor (111) is composed of multiple piezoelectric sensors embedded in the bottom of the pet house and can acquire pressure signals of the pet. At this time, the multi-piezoelectric sensor (111) can acquire pressure signals from each of the multiple channels.

[0038] And the multiple pressure sensors (111) can be embedded at multiple points (e.g., at least two points) within the pet house floor, and each of the multiple pressure sensors (111) embedded at each point may include a piezo sensor for measuring differential pressure and a force sensing resistor (FSR) sensor for measuring absolute pressure, as illustrated in FIG. 3. That is, each multiple pressure sensor (111) may have the piezo sensor and the FSR sensor assembled into a single module.

[0039] The electrode (112) is embedded in the bottom of the pet house to obtain bioelectrical impedance for measuring the body composition (body fat and body water) of the pet, and the load cell (113) is embedded in the bottom of the pet house to obtain impedance for measuring body weight.

[0040] The communication unit (130) is a communication means provided for transmitting and receiving data necessary for the operation of the processor (120), and the storage unit (150) is a storage medium for storing programs and data necessary for the operation of the processor (120).

[0041] The processor (120) is provided to handle all matters for providing functions such as heart rate, respiratory rate and stress analysis of the pet, classification of sleep state, and analysis of body water and body fat.

[0042] To this end, the processor (120) can perform the following functions by processing the signal measured through the sensor unit (110) and analyzing the processed signals.

[0043] - Estimation of pet stress

[0044] - Calculate pet's heart rate and respiratory rate per minute

[0045] - Classification of pet sleep states

[0046] - Measurement of pet weight and body composition

[0047] For example, when estimating stress of a pet, the processor (120) can select the optimal signal among the pressure signals of each channel obtained through the multiple piezoelectric sensors (111), process the signal, and analyze the processed signals to estimate the stress.

[0048] Here, the processor (120) may include a filter circuit (not shown) that performs a filtering operation on a pressure signal measured through multiple piezoelectric sensors, and an amplifier circuit (not shown) that performs an amplification operation on the filtered signal.

[0049] Additionally, the processor (120) can select the optimal signal among the signals that have undergone filtering and amplification when filtering and amplification are performed on pressure signals measured through multiple piezoelectric sensors (111).

[0050] And the processor (120) can perform a Fast Fourier Transform (FFT) operation in which the selected optimal signal is separated into preset time units using a window function, and a heart rate reference filter based on the heart rate frequency band of a normal pet is applied to the signal separated into preset time units, and a Fast Fourier Transform (FFT) operation in which a respiration reference filter based on the respiration frequency band of a normal pet is applied to the signal separated into preset time units.

[0051] And the processor (120) can perform a peak-detection operation on a time domain signal to which a heart rate reference filter is applied to derive heart rate variability based on the variability of the peak value, and estimate stress by performing frequency analysis of the heart rate variability.

[0052] In another example, the processor (120) processes the pressure signal of the pet obtained through the multiple piezoelectric sensors (111), as in the case of calculating the pet's heart rate and respiratory rate per minute or estimating the pet's stress, and when filtering and amplification operations are performed on the pressure signal measured through the multiple piezoelectric sensors (111), the optimal signal among the signals that have undergone filtering and amplification operations can be selected.

[0053] Additionally, the processor (120) can perform a Fast Fourier Transform (FFT) operation in which the selected optimal signal is separated into preset time units using a window function, and a heart rate reference filter based on the heart rate frequency band of a normal pet is applied to the signal separated into preset time units, and a Fast Fourier Transform (FFT) operation in which a respiration reference filter based on the respiration frequency band of a normal pet is applied to the signal separated into preset time units.

[0054] And the processor (120) can calculate the heart rate peak value within the heart rate frequency range and the respiratory rate peak value within the respiratory frequency range based on the signal processing results, and calculate the heart rate and respiratory rate per minute.

[0055] As another example, when classifying the sleep state of a pet, the processor (120) can classify the sleep state of the pet by selecting the optimal signal among the pressure signals of each channel obtained through the piezoelectric sensor (111) as mentioned above, processing the signal, and applying the processed signals to a CNN (Convolutional Neural Networks)-LSTM (Long short-term memory) model.

[0056] Here, the CNN-LSTM model is a model that combines a CNN model for extracting local features and an LSTM model for extracting continuous features based on the local features extracted through the CNN model.

[0057] Specifically, the processor (120) can classify the sleep state of the pet as NREM (Non-REM sleep state), REM (REM sleep state), or WAKE (non-sleep state) by applying the signal processing result to the learned CNN-LSTM model.

[0058] As another example, when measuring the weight and body composition of a pet, the processor (120) can convert an analog signal obtained through the electrode (112) and load cell (113) into a digital signal and measure the weight and body composition of the pet based on the digital signal.

[0059] Meanwhile, the method of using the pet house system according to the present embodiment can provide: ⑴ a stress estimation function for a pet, ⑵ a function to calculate the heart rate and respiratory rate per minute of a pet, ⑶ a sleep state classification function for a pet, and ⑷ a weight and body composition measurement function for a pet by using the pet house system described above with reference to FIGS. 1 to 3.

[0060] FIG. 4 is a flowchart provided in the description of a method for providing (1) a stress estimation function of a pet, (2) a heart rate and respiratory rate calculation function of a pet per minute, and (3) a sleep state classification function of a pet among the methods for using a pet house system according to the present embodiment.

[0061] Referring to FIG. 4, the method of using the pet house system according to the present embodiment may include a pressure signal acquisition process (A), a signal processing process (B), a stress estimation process (C), a heart rate and respiratory rate calculation process (D), and a sleep state classification process (E) to provide (1) a stress estimation function of a pet, (2) a heart rate and respiratory rate calculation function of a pet, and (3) a sleep state classification function of a pet.

[0062] In the pressure signal acquisition process (A), pressure signals are acquired from multiple channels through multiple piezoelectric sensors (111), and filtering and amplification operations are performed on the pressure signals of each channel through a processor (120), and an optimal signal can be selected from the signals that have undergone filtering and amplification operations.

[0063] Specifically, when selecting an optimal signal, the processor (120) obtains an SNR value by calculating the ratio of the signal with the largest signal frequency band to the total signal band of the signal that has undergone filtering and amplification operations, and calculates the peak value of the signal obtained through cross-correlation with the signals of each channel using the signal of the channel with the least noise as a reference signal. Here, FIG. 5 is a diagram illustrating an example of how the peak value of the signal is derived through cross-correlation with the signals of each channel according to an embodiment of the present invention.

[0064] And the processor (120) can calculate the average by normalizing the signal-to-noise ratio value and the cross-correlation peak value for each channel to be 0 to 1, and select the signal of the channel with the highest calculated average as the optimal signal.

[0065] That is, as exemplified in Figure 6, the signal-to-noise ratio value and the cross-correlation peak value for each channel are normalized to have values ​​ranging from 0 to 1, and the average is calculated, and the signal of the channel with the highest score among them is selected as the optimal signal.

[0066] In the signal processing process (B), the processor (120) can perform a Fast Fourier Transform (FFT) operation to separate the selected optimal signal into preset time units using a window function, apply a heart rate reference filter based on the heart rate frequency band of a normal pet to the signal separated into preset time units, and perform a Fast Fourier Transform (FFT) operation to apply a respiration reference filter based on the respiration frequency band of a normal pet to the signal separated into preset time units.

[0067] And in the stress estimation process (C), the processor (120) can estimate stress by performing a peak-detection operation on a time domain signal to which a heart rate reference filter is applied, deriving heart rate variability based on the variability of the peak value, and performing frequency analysis of the heart rate variability.

[0068] In the process of calculating the heart rate and respiratory rate per minute (D), the processor (120) can calculate the heart rate frequency peak value within the heart rate frequency range and the respiratory rate peak value within the respiratory frequency range based on the signal processing results, thereby calculating the heart rate and respiratory rate per minute.

[0069] At this time, the processor (120) can analyze the heart rate and respiratory rate values ​​derived by estimating the peak value and the heart rate and respiratory rate values ​​obtained based on the peak value to calculate the heart rate and respiratory rate per minute.

[0070] In the sleep state classification process (E), the processor (120) can classify the sleep state of the pet as NREM (Non-REM sleep state), REM (REM sleep state), or WAKE (non-sleep state) by applying the signal processing result to the learned CNN-LSTM model.

[0071] FIG. 7 is a flowchart provided in the description of the method of providing the function of measuring the weight and body composition of a pet among the methods of using the pet house system according to the present embodiment.

[0072] Referring to FIG. 7, the method of using the pet house system according to the present embodiment may include (4) a weight measurement process (F) and a body composition measurement process (G) to provide a function for measuring the weight and body composition of a pet.

[0073] In the weight measurement process (F), when an analog signal (impedance) for weight measurement is obtained through the load cell (113), the processor (120) converts the analog signal into a digital signal and can measure the weight of the pet based on the digital signal.

[0074] And in the body composition analysis process (G), when an analog signal (bioelectrical impedance) for measuring body composition (body fat and body water) is obtained through the electrode (112), the processor (120) converts the analog signal into a digital signal and calculates a bioimpedance value based on the digital signal, thereby allowing the body composition of the pet to be measured based on the weight measurement result and the bioimpedance value.

[0075] Meanwhile, it goes without saying that the technical concept of the present invention may also be applied to a computer-readable recording medium containing a computer program that enables the device and method according to the present embodiment to perform their functions. Furthermore, the technical concept according to various embodiments of the present invention may be implemented in the form of computer-readable code recorded on a computer-readable recording medium. A computer-readable recording medium may be any data storage device that can be read by a computer and store data. For example, a computer-readable recording medium may be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, hard disk drive, etc. Additionally, computer-readable code or a program stored on a computer-readable recording medium may be transmitted through a network connected between computers.

[0076] Furthermore, although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above. Various modifications are possible by those skilled in the art without departing from the essence of the invention as claimed in the claims, and such modifications should not be understood individually from the technical spirit or perspective of the present invention. Explanation of the symbols

[0078] 110 : Sensor part 111 : Multi-array Piezoelectric Sensor 112 : Electrode 113 : Load Cell 120 : Processor 130 : Communications Department 140 : Storage section

Claims

Claim 1 A sensor unit acquires a pressure signal of a pet; a processor processes the signal measured through the sensor unit; The processor includes a step of estimating stress in a pet by analyzing heart rate variability (HRV) included in the signal processing result; the step of estimating stress comprises performing a peak detection operation on a time domain signal to which a heart rate reference filter is applied to derive heart rate variability based on the variability of the peak value, and estimating stress by performing frequency analysis of the derived heart rate variability; and the step of acquiring, when pressure signals are acquired from a plurality of channels, performing filtering and amplification operations on the pressure signals acquired from each channel, obtaining an SNR value by calculating the ratio of the signal with the largest signal frequency band to the total signal band of the signal after filtering and amplification, calculating the peak value of the signal obtained through cross-correlation with the signals of each channel using the signal of the channel with the least noise as a reference signal, normalizing the signal-to-noise ratio value and the cross-correlation peak value for each channel so that they are each between 0 and 1 to calculate the average, and selecting the signal of the channel with the highest calculated average as the signal processing target. Selected as a signal, the sensor unit includes a multi-array piezoelectric sensor composed of a plurality of piezoelectric sensors embedded in the bottom of the pet house; and the processor includes a filter unit circuit that performs a filtering operation on a pressure signal measured through the multi-array piezoelectric sensor; and an amplifier unit circuit that performs an amplification operation on the filtered signal.The method of using a pet house system includes a signal processing step in which a selected signal is separated into preset time units using a window function, a Fast Fourier Transform (FFT) operation is performed by applying a heart rate reference filter based on the heart rate frequency band of a normal pet to the signal separated into preset time units, and a Fast Fourier Transform (FFT) operation is performed by applying a respiration reference filter based on the respiration frequency band of a normal pet to the signal separated into preset time units; the method of using a pet house system further includes a step in which a processor classifies the sleep state of a pet by applying the signal processing result to a CNN-LSTM model that combines a CNN model for extracting local features and an LSTM model for extracting continuous features based on the local features extracted through the CNN model; and the step of classifying the sleep state is characterized by classifying the sleep state of the pet into NREM (Non-REM sleep state), REM (REM sleep state), or WAKE (non-sleep state) by applying the signal processing result to a trained CNN-LSTM model. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 delete Claim 7 delete Claim 8 delete Claim 9 delete Claim 10 A sensor unit embedded in the bottom of the pet house to acquire pressure signals from the pet; A processor that processes signals acquired through a sensor unit and analyzes heart rate variability (HRV) included in the signal processing result to estimate the stress of a pet; wherein the sensor unit acquires pressure signals from a plurality of channels, and the processor, when estimating stress, performs a peak detection operation on a time domain signal to which a heart rate reference filter is applied to derive heart rate variability based on the variability of the peak value, and estimates stress by performing frequency analysis of the derived heart rate variability; wherein, when pressure signals are acquired from a plurality of channels, the processor performs filtering and amplification operations on the pressure signals acquired from each channel, obtains an SNR value by calculating the ratio of the signal with the largest signal frequency band to the total signal band of the signal after filtering and amplification, calculates the peak value of the signal obtained through cross-correlation with the signals of each channel using the signal of the channel with the least noise as a reference signal, and such that the signal-to-noise ratio value and the cross-correlation peak value for each channel are each 0 to 1. Calculate an average by normalizing, and select the signal of the channel with the highest calculated average as the signal to be processed; the sensor unit includes a multi-array piezoelectric sensor composed of a plurality of piezoelectric sensors embedded in the bottom of the pet house; and the processor includes a filter unit circuit that performs a filtering operation on a pressure signal measured through the multi-array piezoelectric sensor; and an amplifier unit circuit that performs an amplification operation on the filtered signal.A pet house system comprising: a processor, when performing signal processing, separates a selected signal into preset time units using a window function, performs a Fast Fourier Transform (FFT) operation by applying a heart rate reference filter based on the heart rate frequency band of a normal pet to the signals separated into preset time units, and performs a Fast Fourier Transform (FFT) operation by applying a respiration reference filter based on the respiration frequency band of a normal pet to the signals separated into preset time units; the processor classifies the sleep state of a pet by applying the signal processing results to a CNN-LSTM model, which combines a CNN model for extracting local features and an LSTM model for extracting continuous features based on the local features extracted through the CNN model; and when classifying the sleep state, the processor classifies the sleep state of the pet into NREM (Non-REM sleep state), REM (REM sleep state), or WAKE (non-sleep state) by applying the signal processing results to a trained CNN-LSTM model. Claim 11 delete Claim 12 delete

Images