An intelligent medicine management system based on multi-modal behavior confirmation

Abstract

This invention discloses an intelligent medication management system based on multimodal behavior verification, relating to the field of medication management technology. The system includes: a pressure distribution monitoring module, a removal location calculation module, a spatial matching index module, an attribute association mapping module, and a liquid level linkage verification module. In this invention, by constructing a base mechanical balance model and calculating the torque vector difference derived from the pressure distribution in real time, the planar geometric coordinates of the mass removal point are accurately derived using reverse physical logic, achieving spatial positioning of the medication removal action. Based on locking the drug compartment index and dynamically generating liquid level drop standards, the system forcibly associates drug characteristics with container liquid level change data, performing dual physical fact verification of medication accuracy and adequate water intake, ensuring that the medication process strictly follows pharmacological requirements to complete a closed-loop verification from accurate medication dispensing to scientific administration.

Description

Technical Field

[0001] This invention relates to the field of medication management technology, and in particular to an intelligent medication management system based on multimodal behavior confirmation. Background Technology

[0002] The field of medication management technology involves a systematic project that monitors and controls the entire process of prescription review, distribution, and patient medication use in clinical treatment in medical institutions and home rehabilitation care. Core aspects include using information technology to achieve precise medication dispensing, reminding patients of medication times, and tracking and recording medication dosage and frequency, thus building a closed-loop management system connecting doctors, pharmacists, and patients. Traditional intelligent medication management systems address the technical aspect of confirming whether a patient has actually taken their medication. This involves using intelligent medicine box devices with physical trigger detection capabilities. Microswitches or Hall effect sensors installed on the lid detect the opening and closing status of the lid, or weighing sensors at the bottom of the medicine compartment monitor the decrease in medication weight in real time, using this as a basis for determining whether the patient has taken the medication. Simultaneously, the system uses a built-in timer circuit to drive a buzzer or illuminate an indicator light at a preset time to remind the patient, and transmits the medicine box's open / closed status data or weight change data to a server for recording via a communication circuit.

[0003] Existing technologies rely solely on microswitch on / off signals or simple changes in total weight to determine medication administration. They cannot physically identify the confusion and misadministration of drugs with similar weights but different properties stored in different compartments of a multi-compartment integrated pillbox. This makes it difficult for the monitoring mechanism to establish a correspondence between medication location and drug type, thus failing to intercept the risk of accidental ingestion. Furthermore, such methods sever the intrinsic causal link between the physicochemical properties of the drug and the amount of water consumed, completely lacking quantitative assessment and effectiveness verification of the patient's drinking behavior after medication administration. As a result, drugs with special solubility often have their in vivo disintegration and absorption efficacy severely affected due to insufficient water intake, creating significant logical blind spots and security vulnerabilities in the compliance management of the entire medication administration process. Summary of the Invention

[0004] To address the problem that existing technologies cannot physically identify the confusion and mis-dosing of similar-weight but different-property medications stored in different compartments of multi-compartment integrated pillboxes, leading to difficulties in establishing a correspondence between medication location and type, thus failing to intercept the risk of accidental ingestion, and because such methods sever the intrinsic causal link between the physicochemical properties of the medication and the amount of water consumed, completely lacking quantitative assessment and effectiveness verification of the patient's drinking behavior after medication dispensing, resulting in medications with special solubility often having their in vivo disintegration and absorption efficacy severely affected due to insufficient water intake, creating significant logical blind spots and security vulnerabilities in the compliance management of the entire medication process, this invention provides an intelligent medication management system based on multimodal behavior confirmation. The technical solution is as follows:

[0005] On the one hand, an intelligent medication management system based on multimodal behavior confirmation is provided, the system comprising: The pressure distribution monitoring module collects multiple analog voltage signals output by the sensor array on the medicine box base, converts the voltage signals into pressure load data, calculates the product of load and horizontal axis, and the product of load and vertical axis, and establishes a static equilibrium state vector. The removal location calculation module, based on the static equilibrium state vector, monitors the real-time changes in the overall gravity data of the medicine box, captures the new steady-state pressure distribution data after the gravity step drop, and generates the geometric coordinates of the removal center; The spatial matching index module, for the geometric coordinates of the removed center, calls the pre-stored geometric center mapping map of the drug grid layout, calculates the Euclidean distance between the geometric coordinates of the removed center and the standard center point of the drug grid, compares the Euclidean distance with the determination range of the physical structure of the drug grid, and generates the target drug identity identifier. The attribute association mapping module calls the target drug identity identifier, retrieves the solubility attribute level of the corresponding drug from the built-in drug database, maps the solubility attribute level to the water volume standard required for scientific administration, and generates drinking water compliance requirement indicators.

[0006] Optionally, the static equilibrium state vector includes the horizontal axis static torque component, the vertical axis static torque component, and the initial total gravity load data; the removal center geometric coordinates include the lateral positioning data of the mass removal point and the longitudinal positioning data of the mass removal point; the target drug identification includes the matching drug grid physical sequence number and the drug grid matrix logical position index; and the drinking water compliance requirements include the minimum liquid level drop height limit and the standard liquid volume required for drug disintegration.

[0007] Optionally, the pressure distribution monitoring module includes: The signal conversion and mapping submodule collects multiple analog voltage signals output by the sensor array of the medicine box base, performs sampling and quantization processing on each signal, calls the preset piezoelectric response characteristic curve, maps the discrete digital voltage code to the corresponding physical pressure value, traverses the force status of each node position in the base array, and generates discrete point load data. The torque distribution calculation submodule calls the discrete point load data, reads the physical installation horizontal and vertical coordinates of the sensor in the base plane coordinate system, calculates the product of the single point load and the corresponding horizontal and vertical coordinates, obtains the torque contribution of the single point in the plane coordinate system, performs cumulative summation on the horizontal torque contribution and vertical torque contribution of the node, quantifies the rotation trend distribution of the entire medicine box in the two-dimensional plane, and generates directional torque components. The vector model aggregation submodule, based on the directional moment components, calls the discrete point load data to perform scalar summation on the force intensity of the nodes, obtains the total gravity term, combines the lateral moment accumulation result, the longitudinal moment accumulation result and the total gravity term to construct a multidimensional array describing the real-time mechanical distribution of the medicine box, and generates a static equilibrium state vector.

[0008] Optionally, the removal location calculation module includes: The steady-state change capture submodule calls the total gravity term in the static equilibrium state vector as the monitoring benchmark, continuously scans the real-time gravity data stream of the base sensor array, identifies events where the gravity value drops sharply and the fluctuation amplitude converges to the stable range, collects the pressure reading of the sensor at the stable moment, performs the product operation of pressure and lever arm according to the sensor physical coordinates and accumulates it to generate the post-steady-state torque data. The differential vector operation submodule, based on the post-steady-state torque data, extracts the total lateral torque, total longitudinal torque, and total gravity values ​​at two moments before and after drug retrieval, performs vector subtraction operation, calculates the numerical offset of the lateral torque, the numerical offset of the longitudinal torque, and the absolute reduction of the gravity load before and after the drug retrieval action, and generates a mechanical distribution difference vector. The coordinate inversion analysis submodule extracts the lateral moment offset, longitudinal moment offset, and reduction in gravity load from the mechanical distribution difference vector. It performs a division operation, dividing the lateral moment offset by the reduction in gravity load to calculate the lateral axis position of the removal point, and dividing the longitudinal moment offset by the reduction in gravity load to calculate the ordinate axis position of the removal point. By combining the lateral and ordinate axis position data, it reconstructs the spatial position of the drug leaving the medicine box plane and generates the geometric coordinates of the removal center.

[0009] Optionally, the total gravity term in the steady-state equilibrium vector is used as a fixed reference value to continuously sample the real-time gravity data stream of the base sensor array. When the change in gravity value in adjacent sampling periods exceeds a preset decrease threshold, and the peak fluctuation range of gravity value in multiple consecutive sampling periods is less than a preset stability threshold, it is determined that a stable moment has been reached. The pressure readings of the sensors at the steady-state moment are limited to the instantaneous pressure values ​​of all sensors within the sampling period corresponding to the steady-state moment. The product operation of pressure and lever arm is performed based on the sensor's physical coordinates and the summation is limited to the fixed lever arm length of the sensor relative to the reference origin in the sensor's physical coordinates. The corresponding pressure value is multiplied by the lever arm length one by one and then summed.

[0010] Optionally, the spatial matching index module includes: The distance measurement calculation submodule, for the geometric coordinates of the removal center, calls the pre-stored geometric center mapping map of the drug grid layout, extracts the standard physical center point data of each recorded storage unit, calculates the difference between the horizontal and vertical coordinates of the removal center and the standard center point on the two-dimensional plane, performs square summation and square root operation on the coordinate difference, quantifies the degree of deviation of the removal point from the center of the drug grid, and generates a spatial distance measurement sequence. The initial screening module, based on the spatial distance measurement sequence, calls the preset benchmark value for judging the physical structure of the drug grid, performs a comparison operation between each distance value and the benchmark value for judging the physical structure of the drug grid, logically judges whether the distance value falls within the physical boundary of the drug grid, retains the distance data items that meet the boundary constraint conditions, and generates a list of legal candidate drugs. The unique source point locking submodule retrieves the distance value items included in the list of legitimate candidate drugs, performs numerical sorting and minimum value comparison operations, locks the distance item with the smallest value, tracks the hardware number index corresponding to the distance item in the map mapping relationship, identifies the identity of the storage unit pointed to by this physical removal action, and generates the target drug identity identifier.

[0011] Optionally, the attribute association mapping module includes: The drug attribute retrieval submodule calls the target drug identification as the index key value, accesses the internally preset drug information database, traverses the storage unit record table in the drug information database, matches the drug entries that match the identifier, extracts the solubility classification data of the drug recorded under the drug entry, identifies the level of media conditions required for the drug to disintegrate in the digestive environment, and generates drug solubility characteristic parameters. The water volume benchmark mapping submodule, based on the drug dissolution characteristic parameters, calls the preset pharmaceutical service rule logic table, associates and matches the drug dissolution characteristic parameters with the standard drinking water level defined in the pharmaceutical service rule logic table, determines the minimum liquid volume corresponding to the drug, quantifies the physical water volume benchmark required by the drug, constructs a quantitative correspondence between drug attributes and ingested water volume, and generates a standard water volume value for administration. The liquid level threshold construction submodule performs a physical conversion operation between volume and height for the standard delivered water volume value. It divides the volume value by the container cross-sectional area to calculate the corresponding liquid level drop height value, sets the liquid level drop height value as the critical limit for determining the effectiveness of drinking behavior, constructs a quantitative standard to constrain the patient's drinking behavior to meet the standard, and generates drinking behavior compliance requirements indicators.

[0012] Optionally, the drug solubility characteristic parameters are composed of a multi-level parameter set formed by discretizing solubility classification data and medium condition levels, and the multi-level parameter set is limited to a finite number of preset level values; The standard drinking water level is a set of multiple non-continuous water volume intervals predefined in the pharmaceutical service rule logic table. Each non-continuous water volume interval corresponds one-to-one with a level value in the multi-level parameter set. The standard service water volume value is limited to the lower boundary value in the discontinuous water volume range.

[0013] Optionally, the system further includes a liquid level linkage verification module: The liquid level linkage verification module tracks the real-time fluctuation of the liquid level in the smart water cup based on the drinking water compliance requirement index, calculates the cumulative drop in liquid level after the medication dispensing action, compares the difference between the cumulative drop in liquid level and the drinking water compliance requirement index, and generates a medication completion judgment result. The medication completion determination result includes a logical flag indicating correct medication dispensing and a compliance check code for drinking water behavior.

[0014] Optionally, the liquid level linkage verification module includes: The liquid level change tracking submodule monitors the real-time height signal uploaded by the liquid level sensor inside the smart water cup through the drinking water compliance requirement index. It sets the time when the medicine dispensing action occurs as the timestamp start point, locks the initial liquid level height data at that time, continuously collects instantaneous liquid level height data within the time window, performs subtraction to calculate the drop difference between the instantaneous height and the initial height, accumulates and filters the continuous drop data, quantifies the spatial volume change of the liquid ingested by the patient in real time, and generates cumulative liquid level consumption data. The compliance difference comparison submodule, based on the cumulative liquid level consumption data and combined with the drinking water compliance requirement index as a comparison benchmark, performs numerical difference operation to calculate the gap between the real-time consumption data and the compliance requirement index, obtains the percentage coefficient of the real-time consumption data to the compliance requirement index, evaluates the degree of conformity between the real-time drinking behavior and the drug delivery requirements, and generates the drinking water compliance verification quantity. The behavior closed-loop judgment submodule calls the preset medication effectiveness judgment logic for the drinking water compliance verification quantity, checks whether the verification quantity has reached the preset minimum limit, verifies whether the patient has completed sufficient drinking water cooperation after taking the medicine, and generates a medication completion judgment result by combining the occurrence of the medication taking action and the compliance status of the drinking water behavior.

[0015] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, by constructing a mechanical balance model of the base and calculating the torque vector difference derived from the pressure distribution in real time, the planar geometric coordinates of the mass removal point are accurately derived using reverse physical logic, thereby realizing the spatial positioning of the drug removal action. Based on locking the drug grid index and the drug dissolution properties, a liquid level drop standard is dynamically generated, and the drug characteristics and container liquid level change data are forcibly associated. This provides dual physical fact confirmation of the correctness of drug dispensing and the adequacy of drinking water, ensuring that the medication process strictly follows pharmacological requirements to complete the closed-loop verification from accurate drug dispensing to scientific delivery. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A schematic diagram of the system provided by the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the pressure distribution monitoring module in this invention; Figure 4 This is a flowchart of the removal of the position calculation module in this invention; Figure 5 This is a flowchart of the spatial matching index module in this invention; Figure 6 This is a flowchart of the attribute association mapping module in this invention; Figure 7 This is a flowchart of the liquid level linkage verification module in this invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] This invention provides an intelligent medication management system based on multimodal behavior confirmation, such as... Figures 1-2 The diagram shown illustrates an intelligent medication management system based on multimodal behavior verification. The system includes: The pressure distribution monitoring module collects multiple analog voltage signals output by the sensor array on the medicine box base, converts the voltage signals into pressure load data, calculates the product of load and horizontal axis, the product of load and vertical axis, accumulates the torque data of support points, constructs an overall mechanical balance model of the base, summarizes the torque components and gravity components in multiple directions, and establishes a static equilibrium state vector. The removal location calculation module, based on the static equilibrium state vector, monitors the real-time changes in the overall gravity data of the medicine box, captures the new steady-state pressure distribution data after the gravity step drop, calculates the ratio of torque difference to gravity change, inverts the geometric center of the mass removal point, and generates the geometric coordinates of the removal center. The spatial matching index module, for removing the center geometric coordinates, calls the pre-stored drug grid layout geometric center mapping map, calculates the Euclidean distance between the removed center geometric coordinates and the standard center point of the drug grid, compares the Euclidean distance with the drug grid physical structure judgment range, filters the drug grid numbers with the minimum distance feature, and generates the target drug identity identifier. The attribute association mapping module calls the target drug identity identifier, retrieves the corresponding drug's solubility attribute level from the built-in drug database, maps the solubility attribute level to the scientifically required water volume standard for administration, constructs a drinking behavior verification benchmark for the characteristics of this drug, and generates drinking water compliance requirement indicators. The liquid level linkage verification module tracks the real-time fluctuation of the liquid level in the smart water cup based on the drinking water compliance requirements, calculates the cumulative drop in liquid level after the medication dispensing action, compares the difference between the cumulative drop in liquid level and the drinking water compliance requirements, and generates a medication completion judgment result.

[0020] The static equilibrium state vector includes the horizontal axis static torque component, the vertical axis static torque component, and the initial total gravity load data. The removal center geometric coordinates include the horizontal positioning data of the mass removal point and the vertical positioning data of the mass removal point. The target drug identification includes the matching drug grid physical serial number and the drug grid matrix logical position index. The drinking water compliance requirements include the minimum liquid level drop height limit and the standard liquid volume required for drug disintegration. The medication completion judgment result includes the medication correctness logical flag bit and the drinking water behavior compliance verification code.

[0021] Specifically, such as Figure 2 , Figure 3 As shown, the pressure distribution monitoring module includes: The signal conversion and mapping submodule collects multiple analog voltage signals output by the sensor array of the medicine box base, performs sampling and quantization processing on each signal, calls the preset piezoelectric response characteristic curve, maps the discrete digital voltage code to the corresponding physical pressure value, traverses the force status of each node position in the base array, and generates discrete point load data. A multi-channel physical connection is established with the high-density array pressure sensor embedded in the medicine box base. This array consists of 256 independent piezoresistive sensing units arranged in a 16x16 matrix, each measuring 5 mm by 5 mm, covering the load-bearing area of ​​the base. The 256 analog voltage signals output by the sensor array are acquired in parallel at a sampling frequency of 50 Hz. A built-in 12-bit analog-to-digital converter quantizes the analog voltage signals into digital voltage codes, which internally store pressure response characteristic curve data. This curve describes the linear or non-linear relationship between the sensor output voltage and the applied pressure. The digital voltage codes of each sensor node are iterated, and interpolation and correction operations are performed based on the pressure response characteristic curve, converting the dimensionless digital codes into values ​​in Newtons (N). The physical pressure value is expressed in units of 1.25 volts. For example, when the voltage of a sensor node is 1.25 volts... When ), based on the conversion factor of 2.0 Newtons per volt corresponding to the characteristic curve (), The calculated physical pressure at this node is 2.5 Newtons. The physical pressure values ​​of 256 nodes are fully scanned and formatted for reconstruction, and a 16-row, 16-column digital description matrix is ​​established. Each element in the matrix corresponds precisely to the real-time pressure value of a specific physical coordinate point on the base plane, thereby generating discrete point load data. As shown in Table 1, the collected voltage and converted pressure load data of some sensor nodes are listed.

[0022] Table 1: Example of Sensor Node Voltage to Pressure Load Conversion

[0023] The torque distribution calculation submodule calls the load data of discrete points, reads the physical installation horizontal and vertical coordinates of the sensors in the base plane coordinate system, calculates the product of the single-point load and the corresponding horizontal and vertical coordinates, obtains the torque contribution of the single point in the plane coordinate system, performs cumulative summation on the horizontal and vertical torque contributions of the nodes, quantifies the rotation trend distribution of the entire medicine box in the two-dimensional plane, and generates directional torque components. The system reads the base plane coordinate system parameters pre-stored in the system register. This coordinate system has the geometric center of the base as the origin, with the positive horizontal axis pointing to the right and the positive vertical axis pointing upwards. The torque distribution calculation submodule traverses each non-zero load data point in the matrix, extracts the physical installation horizontal and vertical coordinate values ​​of that point, and performs multiplication operations on each. The single-point pressure load data is multiplied by the corresponding physical installation horizontal coordinate to obtain the horizontal torque contribution of that point relative to the vertical axis. At the same time, the single-point pressure load data is multiplied by the corresponding physical installation vertical coordinate to obtain the vertical torque contribution of that point relative to the horizontal axis. After completing the single-point torque calculation of the effective nodes, an accumulation and summation operation is performed to algebraically add the horizontal torque contributions of the nodes to obtain the total horizontal torque of the entire medicine box. Similarly, the vertical torque contributions of the nodes are algebraically added to obtain the total vertical torque. Assuming that the pressure load of sensor S-045 is 2.40 Newtons, its horizontal coordinate is 30 mm, and its vertical coordinate is 40 mm, then the horizontal torque contribution of that point is 72 Newton-millimeters. The longitudinal torque contribution is 96 N / mm. If the base is only subjected to two points, S-001 and S-045, and the pressure load of S-001 is 1.00 Newtons with coordinates of 10 mm and 10 mm, then the total lateral torque calculated by the system is the sum of 10 N / mm and 72 N / mm, i.e., 82 N / mm, and the total longitudinal torque is the sum of 10 N / mm and 96 N / mm, i.e., 106 N / mm. Through such calculations, the rotational trend distribution of the entire medicine box on the two-dimensional plane is quantified, and directional torque components are generated.

[0024] The vector model aggregation submodule, based on directional moment components, calls discrete point load data to perform scalar summation on the force intensity of nodes, obtains the total gravity term, combines the lateral moment accumulation result, the longitudinal moment accumulation result and the total gravity term to construct a multidimensional array describing the real-time mechanical distribution of the medicine box, and generates a static equilibrium state vector; The discrete point load data is called again, and a scalar summation operation is performed on the force intensity of the nodes, that is, the pressure values ​​in the matrix are directly added together to obtain the total gravity term. For example, under the action of the two points mentioned above, the total gravity term is the sum of 1.00 Newtons and 2.40 Newtons, that is, 3.40 Newtons. The values ​​of the total lateral moment, the total longitudinal moment, and the total gravity term are combined in a predetermined order to construct a multidimensional array containing three components. This array is the static equilibrium state vector describing the complete mechanical distribution of the medicine box at the current moment. In the example above, the generated static equilibrium state vector is specifically as follows: , representing the lateral moment, longitudinal moment, and total gravity, respectively.

[0025] Specifically, such as Figure 2 , Figure 4 As shown, removing the location calculation module includes: The steady-state change capture submodule calls the total gravity term in the static equilibrium state vector as the monitoring benchmark, continuously scans the real-time gravity data stream of the base sensor array, identifies events where the gravity value drops sharply and the fluctuation amplitude converges to the stable range, collects the pressure readings of the sensors at the stable moment, performs the product calculation of pressure and lever arm according to the sensor physical coordinates and accumulates them to generate post-steady-state torque data. The system continuously scans the real-time gravity data stream uploaded by the base sensor array in 100-millisecond time windows and calculates the variance of the gravity data within the sliding time window. When a step drop in gravity is detected and the variance of gravity data within 5 consecutive time windows is less than 0.01 Newtons squared, the system is determined to have entered a new steady state, i.e., a drug removal action has occurred. The system locks the new steady state moment, collects the pressure readings of the sensors at that moment, and performs the product calculation of pressure and lever arm again based on the physical coordinates of each sensor and accumulates them. For example, if the initial gravity is 3.40 Newtons, and the gravity stabilizes at 3.35 Newtons after the drug is removed, and the variance meets the convergence condition, then subsequent processing is triggered to generate post-steady-state torque data.

[0026] The differential vector operation submodule, based on the post-steady-state torque data, extracts the total lateral torque, total longitudinal torque, and total gravity values ​​at two moments before and after drug retrieval, performs vector subtraction operations, calculates the numerical offset of the lateral torque, the numerical offset of the longitudinal torque, and the absolute reduction of the gravity load before and after the drug retrieval action, and generates a mechanical distribution difference vector. Calculate the numerical shift of the lateral moment before and after the drug-retrieving action, i.e., subtract the subsequent lateral total moment from the initial total lateral moment; calculate the numerical shift of the longitudinal moment, i.e., subtract the subsequent longitudinal total moment from the initial total longitudinal moment; calculate the absolute reduction of the gravitational load, i.e., subtract the subsequent total gravity from the initial total gravity. These three differences constitute the force distribution difference vector. Assuming that the total lateral moment after drug retrieval becomes 80 N / mm, the total longitudinal moment becomes 126 N / mm, and the total gravity becomes 3.35 N, then the lateral moment shift is: Newton-millimeter, longitudinal torque offset is Newton-millimeter, the reduction in gravitational load is Newton, the generated force distribution difference vector is specifically as follows .

[0027] The coordinate inversion analysis submodule extracts the lateral moment offset, longitudinal moment offset, and reduction in gravity load from the force distribution difference vector. It performs division operations, dividing the lateral moment offset by the reduction in gravity load to calculate the horizontal axis position of the removal point, and dividing the longitudinal moment offset by the reduction in gravity load to calculate the vertical axis position of the removal point. By combining the horizontal and vertical axis position data, it reconstructs the spatial position of the drug leaving the medicine box plane and generates the geometric coordinates of the removal center. The spatial location of the mass removal point is reconstructed using the reverse derivation logic of the torque balance principle. The lateral torque offset, longitudinal torque offset, and reduction in gravitational load are extracted from the tensor. A division operation is performed: the lateral torque offset is divided by the reduction in gravitational load to calculate the lateral axis coordinate of the removal point; the longitudinal torque offset is divided by the reduction in gravitational load to calculate the ordinate coordinate of the removal point. Based on the previous example data, the lateral torque offset of 2 N / mm is divided by the reduction in gravitational load of 0.05 N, resulting in a lateral axis position of 40 mm for the removal point. Dividing the longitudinal torque offset of 4 N / mm by the reduction in gravitational load of 0.05 N, the calculated removal point on the longitudinal axis is 80 mm. Combining these two calculation results, namely 40 mm on the horizontal axis and 80 mm on the vertical axis, the specific geometric center point of the drug leaving the plane of the medicine box is reconstructed, generating the geometric coordinates of the removal center. .

[0028] Specifically, such as Figure 2 , Figure 5 As shown, the spatial matching index module includes: The distance measurement calculation submodule, for the geometric coordinates of the removal center, calls the pre-stored geometric center mapping map of the drug grid layout, extracts the standard physical center point data of each recorded storage unit, calculates the difference between the horizontal and vertical coordinates of the removal center and the standard center point on the two-dimensional plane, performs square summation and square root operation on the coordinate difference, quantifies the degree of deviation of the removal point from the center of the drug grid, and generates a spatial distance measurement sequence. The pre-stored pouch layout geometric center mapping map stored in the system's non-volatile memory is invoked. This map details the hardware number of each individual pouch storage unit in the pouch and its standard physical center point data in the base coordinate system, as shown in Table 2. This table contains partial pouch layout mapping data stored within the system. For each pouch record, the difference between the removal center geometric coordinates and the horizontal and vertical coordinates of the standard center point on the two-dimensional plane is calculated. For example, for pharmacy compartment numbered G-02, its standard center is... Calculate the difference in the x-coordinates, i.e. Millimeters; calculate the difference in the ordinate, i.e. The coordinate difference is calculated by summing the squares and taking the square root, i.e., the sum of the squares of 2 and 2 is calculated to get 8. Then, the square root of 8 is taken to get the Euclidean distance of about 2.83 millimeters. The same calculation process is performed on the medicine grids in the map to quantify the degree of deviation of the removal point relative to the center of each medicine grid and generate a spatial distance metric sequence containing the distance values ​​of the medicine grids.

[0029] Table 2: Example Table of Geometric Center Mapping Map for Drug Grid Layout

[0030] The initial screening module, based on the spatial distance measurement sequence, calls the preset benchmark value for judging the physical structure of the drug grid, performs a comparison operation between each distance value and the benchmark value for judging the physical structure of the drug grid, logically determines whether the distance value falls within the physical boundary of the drug grid, retains the distance data items that meet the boundary constraints, and generates a list of legal candidate drugs. The baseline value for determining the physical structure range of the medicine grid is set as the effective physical radius of the medicine grid opening, i.e., 15 mm in Table 2. Each distance value in the sequence is compared with this baseline value, and a logical judgment is made as to whether the distance value is less than or equal to the determination radius. For the calculated result of 2.83 mm, since it is less than 15 mm, it is determined to fall within the physical boundary of medicine grid G-02, and this data item is retained. For medicine grid G-01, its center is... ,and The distance is approximately 58.03 mm, which is greater than 15 mm. Therefore, it is judged as a distance data item that does not meet the boundary constraint conditions and is removed. After traversing the sequence, a list containing only the drug grid numbers and their distance data that meet the boundary constraint conditions is constructed to generate a list of legal candidate drugs.

[0031] The unique source point locking submodule retrieves the distance value items included in the list of legitimate candidate drugs, performs numerical sorting and minimum value comparison operations, locks the distance item with the smallest value, tracks the hardware number index corresponding to the distance item in the map mapping relationship, identifies the identity of the storage unit pointed to by this physical removal action, and generates the target drug identity identifier. In most cases, the list contains only one candidate, but there are multiple candidates in the boundary area. The distance item with the smallest value is locked. If there are multiple candidates, the one with the smallest distance is selected as the only result. In the example above, assuming that only G-02 meets the condition, the distance item corresponding to G-02 is locked directly, and the hardware number index "G-02" corresponding to the distance item in the map mapping relationship is tracked. It is confirmed that the physical removal action did indeed occur in the drug cell G-02, and the number is output to generate the target drug identification.

[0032] Specifically, such as Figure 2 , Figure 6 As shown, the attribute association mapping module includes: The drug attribute retrieval submodule calls the target drug's identity identifier as the index key value, accesses the internally preset drug information database, traverses the storage unit record table in the drug information database, matches the drug entries that match the identifier, extracts the solubility classification data of the drug recorded under the drug entry, identifies the level of media conditions required for the drug to disintegrate in the digestive environment, and generates drug solubility characteristic parameters. The system receives the target drug identifier (e.g., "G-02") and uses it as the primary key index to access the pre-built drug information database within the system. This database uses a relational data structure to store the correspondence between drug grid numbers and detailed drug information. It traverses the storage unit record tables in the database and matches the drug entry that matches the identifier "G-02". Assuming that the drug corresponding to this entry is "Amoxicillin Capsules", it extracts the solubility classification data of the drug recorded under this drug entry. This data field is defined as "Solubility Grade", and the value range includes "Frequently Soluble", "Soluble", "Slightly Soluble", and "Poorly Soluble". It reads the value of this field to identify the level of media conditions required for the drug to disintegrate in the digestive environment. In this example, assuming that the solubility grade of Amoxicillin Capsules is recorded as "Soluble", it generates drug solubility characteristic parameters.

[0033] The water volume benchmark mapping submodule, based on drug solubility characteristic parameters, calls the preset pharmaceutical service rule logic table, associates and matches the drug solubility characteristic parameters with the standard drinking water level defined in the pharmaceutical service rule logic table, determines the minimum liquid volume corresponding to the drug, quantifies the physical water volume benchmark required by the drug, constructs a quantitative correspondence between drug attributes and ingested water volume, and generates a standard water volume value for administration. The system invokes a pre-defined pharmaceutical service rule logic table, which defines standard drinking water levels corresponding to different solubility grades. For example, the logic table specifies that "easily soluble" corresponds to 100 ml of water, "soluble" corresponds to 200 ml of water, "slightly soluble" corresponds to 300 ml of water, and "poorly soluble" corresponds to 500 ml of water. The parameter "soluble" is matched with the logic table to determine that the minimum liquid volume for administration of the drug is 200 ml. This process quantifies the physical water volume benchmark required for the drug to achieve its optimal dissolution rate in the body, establishes a quantitative correspondence between drug properties and ingested water volume, and generates a standard water volume value for administration.

[0034] The liquid level threshold construction submodule performs a physical conversion operation between volume and height for the standard delivered water volume value. It divides the volume value by the container cross-sectional area to calculate the corresponding liquid level drop height value, sets the liquid level drop height value as the critical limit for determining the effectiveness of drinking behavior, constructs a quantitative standard to constrain patients' drinking behavior to meet the standard, and generates drinking behavior compliance requirements indicators. Assuming the smart water cup configured in the system is cylindrical with an inner diameter of 80 millimeters, its inner cross-sectional area can be calculated using the formula for the area of ​​a circle, i.e. The calculated value is approximately 5024 square millimeters, which is equivalent to 50.24 square centimeters. To convert this to height, the standard volume of water (200 ml, or 200 cubic centimeters) is divided by the container's cross-sectional area (50.24 square centimeters) to obtain the corresponding drop in liquid level. The result was approximately 3.98 cm. This calculated height value (3.98 cm) was set as the critical threshold for determining the effectiveness of drinking behavior. A quantitative standard was constructed to constrain patients' drinking behavior to meet the standard. It was clarified that when patients take this specific drug, the water level in the cup must drop by at least 3.98 cm to be considered compliant, thus generating a drinking behavior compliance requirement index.

[0035] Specifically, such as Figure 2 , Figure 7 As shown, the liquid level linkage verification module includes: The liquid level change tracking submodule monitors the real-time height signal uploaded by the liquid level sensor inside the smart water cup through the drinking water compliance requirement index. It sets the time when the medicine is taken as the timestamp start point, locks the initial liquid level height data at that time, continuously collects the instantaneous liquid level height data within the time window, performs subtraction to calculate the drop difference between the instantaneous height and the initial height, accumulates and filters the continuous drop data, quantifies the spatial volume change of the liquid ingested by the patient in real time, and generates cumulative liquid level consumption data. The smart water cup's monitoring function is activated, communicating with a capacitive liquid level sensor installed on the inner wall of the cup via Bluetooth. It monitors the real-time uploaded height signal and verifies the moment the medication dispensing action occurs (i.e., the moment the aforementioned removal position calculation module determines a steady-state change) as the timestamp start point. The liquid level data uploaded by the sensor at this moment is locked as the initial liquid level height. Assuming the initial height is 12.00 cm, a monitoring time window of 5 minutes is opened, continuously collecting instantaneous liquid level data within this time window at a frequency of 1 Hz. For each collection point, a subtraction operation is performed to calculate the difference between the initial height and the instantaneous height, obtaining the instantaneous drop. To eliminate noise caused by water sloshing, Kalman filtering is applied to the continuous drop data to smooth data fluctuations. The maximum drop in liquid level is continuously tracked to quantify the spatial volume change of the liquid ingested by the patient in real time, generating cumulative liquid level consumption data. Assuming the liquid level stabilizes at 7.50 cm after 3 minutes, the cumulative liquid level consumption data is calculated as follows: That is, 4.50 centimeters.

[0036] The compliance difference comparison submodule, based on the cumulative liquid level consumption data and combined with the drinking water compliance requirement index as the comparison benchmark, performs numerical difference operation to calculate the gap between the real-time consumption data and the compliance requirement index, obtains the percentage coefficient of the real-time consumption data to the compliance requirement index, evaluates the degree of conformity between the real-time drinking behavior and the drug delivery requirements, and generates the drinking water compliance verification quantity. Based on the cumulative liquid level consumption data (4.50 cm), and combined with the drinking water compliance requirement index (3.98 cm) as a comparison benchmark, numerical difference calculation is performed to calculate the difference between the real-time consumption data and the compliance requirement index. This yields a surplus of 0.52 cm, or, by performing division, the following calculation can be performed: The result is approximately 1.13, which means that 113% of the drinking water requirements have been met. The accuracy of the drinking water compliance verification is then calculated by assessing the degree of consistency between real-time drinking behavior and drug administration requirements.

[0037] The behavior closed-loop judgment submodule calls the preset medication effectiveness judgment logic for the drinking water compliance verification quantity, checks whether the verification quantity has reached the preset minimum limit, verifies whether the patient has completed the drinking water cooperation after taking the medicine, and generates the medication completion judgment result by combining the occurrence of the medication taking action and the compliance status of the drinking behavior. For the compliance verification of drinking water (113% or 0.52 cm), the preset medication effectiveness judgment logic is invoked. This logic sets a minimum allowable limit, such as 90% or a drop gap of no more than 0.5 cm. It checks whether the verification amount reaches this preset limit. In this example, 113% is greater than 90%, and the surplus is positive, so the verification is passed, confirming that the patient has completed sufficient drinking water cooperation after taking the medication. The system combines the previously confirmed medication taking action (i.e., the successful generation of the target drug identification) with the currently confirmed compliance status of drinking water behavior, and performs a logical "AND" operation. Only when both conditions of "correct medication taking" and "compliant drinking water" are true, does the system determine that the medication behavior is complete and effective. This includes a Boolean success flag and detailed execution data logs for subsequent medical record archiving. Through the calculation of the above experimental data, this verification logic shows that after setting strict physical parameter constraints, the system can accurately distinguish between invalid drinking water and effective administration, and generate a medication completion judgment result.

[0038] In this embodiment of the invention, by constructing a mechanical balance model of the base and calculating the torque vector difference derived from the pressure distribution in real time, the planar geometric coordinates of the mass removal point are accurately derived using reverse physical logic, thereby realizing the spatial positioning of the drug removal action. Based on locking the drug grid index and the drug dissolution properties, a liquid level drop standard is dynamically generated, and the drug characteristics and container liquid level change data are forcibly associated. This provides dual physical fact confirmation of the correctness of drug dispensing and the adequacy of drinking water, ensuring that the medication process strictly follows pharmacological requirements to complete the closed-loop verification from accurate drug dispensing to scientific delivery.

[0039] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Unless otherwise specified, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0040] The use of terms such as "an embodiment," "an embodiment," "an exemplary embodiment," and "some embodiments" in the specification indicates that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the art.

[0041] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0042] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0043] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0044] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0045] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the technical solution.

Claims

1. An intelligent medication management system based on multi-modal behavior confirmation, characterized in that, The system includes: The pressure distribution monitoring module collects multiple analog voltage signals output by the sensor array on the medicine box base, converts the voltage signals into pressure load data, calculates the product of load and horizontal axis, and the product of load and vertical axis, and establishes a static equilibrium state vector. The removal location calculation module, based on the static equilibrium state vector, monitors the real-time changes in the overall gravity data of the medicine box, captures the new steady-state pressure distribution data after the step drop in gravity, and generates the geometric coordinates of the removal center. The spatial matching index module, for the geometric coordinates of the removal center, calls the pre-stored geometric center mapping map of the drug grid layout, calculates the Euclidean distance between the geometric coordinates of the removal center and the standard center point of the drug grid, compares the Euclidean distance with the determination range of the physical structure of the drug grid, and generates the target drug identity identifier. The attribute association mapping module calls the target drug identity identifier, retrieves the solubility attribute level of the corresponding drug in the built-in drug database, maps the solubility attribute level to the water volume standard required for scientific administration, and generates drinking water compliance requirement indicators. The liquid level linkage verification module tracks the real-time fluctuation of the liquid level in the smart water cup based on the drinking water compliance requirement index, calculates the cumulative drop in liquid level after the medication dispensing action, compares the difference between the cumulative drop in liquid level and the drinking water compliance requirement index, and generates a medication completion judgment result. The medication completion determination result includes a logical flag indicating correct medication dispensing and a compliance verification code for drinking water behavior. The pressure distribution monitoring module includes: The signal conversion and mapping submodule acquires multiple analog voltage signals output by the sensor array of the medicine box base, performs sampling and quantization processing on each signal, calls the preset piezoelectric response characteristic curve, maps the discrete digital voltage code to the corresponding physical pressure value, traverses the force status of each node position in the base array, and generates discrete point load data. The torque distribution calculation submodule calls the discrete point load data, reads the physical installation horizontal and vertical coordinates of the sensor in the base plane coordinate system, calculates the product of the single point load and the corresponding horizontal and vertical coordinates, obtains the torque contribution of the single point in the plane coordinate system, performs cumulative summation on the horizontal torque contribution and vertical torque contribution of the node, quantifies the rotation trend distribution of the entire medicine box in the two-dimensional plane, and generates directional torque components. The vector model aggregation submodule, based on the directional moment components, calls the discrete point load data to perform scalar summation on the force intensity of the nodes, obtains the total gravity term, combines the lateral moment accumulation result, the longitudinal moment accumulation result and the total gravity term to construct a multidimensional array describing the real-time mechanical distribution of the medicine box, and generates a static equilibrium state vector. The removal location calculation module includes: The steady-state change capture submodule calls the total gravity term in the static equilibrium state vector as the monitoring benchmark, continuously scans the real-time gravity data stream of the base sensor array, identifies events where the gravity value drops sharply and the fluctuation amplitude converges to the stable range, collects the pressure reading of the sensor at the stable moment, performs the product operation of pressure and lever arm according to the sensor physical coordinates and accumulates it to generate the post-steady-state torque data. Based on the post-steady-state torque data, the differential vector operation submodule extracts the total lateral torque, total longitudinal torque, and total gravity values ​​at two moments before and after drug retrieval, performs vector subtraction, calculates the numerical offset of the lateral torque before and after the drug retrieval action, the numerical offset of the longitudinal torque, and the absolute reduction of the gravity load, and generates a mechanical distribution difference vector. The coordinate inversion analysis submodule extracts the lateral moment offset, longitudinal moment offset, and reduction in gravity load from the mechanical distribution difference vector. It performs a division operation, dividing the lateral moment offset by the reduction in gravity load to calculate the lateral axis position of the removal point, and dividing the longitudinal moment offset by the reduction in gravity load to calculate the ordinate axis position of the removal point. By combining the lateral and ordinate axis position data, the spatial position of the drug leaving the medicine box plane is reconstructed, and the geometric coordinates of the removal center are generated.

2. The smart medication management system based on multi-modal behavior confirmation as claimed in claim 1, wherein, The static equilibrium state vector includes the horizontal axis static torque component, the vertical axis static torque component, and the initial total gravity load data. The removal center geometric coordinates include the lateral positioning data of the mass removal point and the longitudinal positioning data of the mass removal point. The target drug identification includes the matching drug grid physical serial number and the drug grid matrix logical position index. The drinking water compliance requirements include the minimum liquid level drop height limit and the standard liquid volume required for drug disintegration.

3. The smart medication management system based on multi-modal behavior confirmation as claimed in claim 1, wherein, Using the total gravity term in the steady-state equilibrium vector as a fixed reference value, the real-time gravity data stream of the base sensor array is continuously sampled. When the change in gravity value in adjacent sampling periods exceeds the preset decrease threshold, and the peak fluctuation range of gravity value in multiple consecutive sampling periods is less than the preset stability threshold, it is determined that a stable moment has been reached. The pressure readings of the sensors at the steady-state moment are limited to the instantaneous pressure values ​​of all sensors within the sampling period corresponding to the steady-state moment. The product operation of pressure and lever arm is performed based on the sensor's physical coordinates and the summation is limited to the fixed lever arm length of the sensor relative to the reference origin in the sensor's physical coordinates. The corresponding pressure value is multiplied by the lever arm length one by one and then summed.

4. The intelligent medication management system based on multimodal behavior confirmation according to claim 3, characterized in that, The spatial matching index module includes: The distance measurement calculation submodule, for the geometric coordinates of the removal center, calls the pre-stored geometric center mapping map of the drug grid layout, extracts the standard physical center point data of each recorded storage unit, calculates the difference between the horizontal and vertical coordinates of the removal center and the standard center point on the two-dimensional plane, performs square summation and square root operation on the coordinate difference, quantifies the degree of deviation of the removal point from the center of the drug grid, and generates a spatial distance measurement sequence. The initial screening module, based on the spatial distance measurement sequence, calls the preset benchmark value for determining the physical structure of the drug grid, performs a comparison operation between each distance value and the benchmark value for determining the physical structure of the drug grid, logically determines whether the distance value falls within the physical boundary of the drug grid, retains the distance data items that meet the boundary constraint conditions, and generates a list of legal candidate drugs. The unique source point locking submodule retrieves the distance value items included in the list of legitimate candidate drugs, performs numerical sorting and minimum value comparison operations, locks the distance item with the smallest value, tracks the hardware number index corresponding to the distance item in the map mapping relationship, identifies the identity of the storage unit pointed to by this physical removal action, and generates the target drug identity identifier.

5. The smart medication management system based on multi-modal behavior confirmation as claimed in claim 4, wherein, The attribute association mapping module includes: The drug attribute retrieval submodule calls the target drug identification as the index key value, accesses the internally preset drug information database, traverses the storage unit record table in the drug information database, matches the drug entries that match the identifier, extracts the solubility classification data of the drug recorded under the drug entry, identifies the level of media conditions required for the drug to disintegrate in the digestive environment, and generates drug solubility characteristic parameters. The water volume benchmark mapping submodule, based on the drug dissolution characteristic parameters, calls the preset pharmaceutical service rule logic table, associates and matches the drug dissolution characteristic parameters with the standard drinking water level defined in the pharmaceutical service rule logic table, determines the minimum liquid volume for drug administration, quantifies the physical water volume benchmark required for the drug, constructs a quantitative correspondence between drug attributes and ingested water volume, and generates a standard water volume value for administration. The liquid level threshold construction submodule performs a physical conversion operation between volume and height for the standard delivered water volume value. It divides the volume value by the container cross-sectional area to calculate the corresponding liquid level drop height value, sets the liquid level drop height value as the critical limit for determining the effectiveness of drinking behavior, constructs a quantitative standard to constrain the patient's drinking behavior to meet the standard, and generates drinking behavior compliance requirement indicators.

6. The smart medication management system based on multi-modal behavior confirmation as claimed in claim 5, wherein, The drug solubility characteristic parameters are composed of a multi-level parameter set formed by discretizing solubility classification data and media condition levels. The multi-level parameter set is limited to a finite number of preset level values. The standard drinking water level is a set of multiple non-continuous water volume intervals predefined in the pharmaceutical service rule logic table. Each non-continuous water volume interval corresponds one-to-one with a level value in the multi-level parameter set. The standard service water volume value is limited to the lower boundary value in the discontinuous water volume range.

7. The smart medication management system based on multi-modal behavior confirmation as claimed in claim 1, wherein, The liquid level linkage verification module includes: The liquid level change tracking submodule monitors the real-time height signal uploaded by the liquid level sensor inside the smart water cup through the drinking water compliance requirement index. It sets the moment when the medicine dispensing action occurs as the timestamp start point, locks the initial liquid level height data at that moment, continuously collects instantaneous liquid level height data within the time window, performs subtraction to calculate the drop difference between the instantaneous height and the initial height, accumulates and filters the continuous drop data, quantifies the spatial volume change of the liquid ingested by the patient in real time, and generates cumulative liquid level consumption data. The compliance difference comparison submodule uses the cumulative liquid level consumption data and the drinking water compliance requirement index as a comparison benchmark to perform numerical difference operation, calculate the gap between the real-time consumption data and the compliance requirement index, obtain the percentage coefficient of the real-time consumption data to the compliance requirement index, evaluate the degree of conformity between the real-time drinking behavior and the drug delivery requirements, and generate the drinking water compliance verification quantity. The behavior closed-loop judgment sub-module calls preset medicine-taking effectiveness judgment logic for the drinking water compliance check quantity, checks whether the check quantity reaches the preset minimum limit, verifies whether the patient completes sufficient drinking water cooperation after taking medicine, and generates a medicine-taking completion degree judgment result in combination with the occurrence fact of the medicine-taking action and the compliance state of the drinking water behavior.

Images