Method and system for adjusting height of multifunctional vegetable basin

Abstract

The invention relates to the technical field of intelligent kitchen ware, in particular to a method and system for adjusting the height of a multifunctional vegetable washing basin, and the method comprises the steps: monitoring the operation force, water flow and gravity center change in real time through multiple sensors, providing data support for the height adjustment of the vegetable washing basin, and improving the adjustment adaptability; by automatically identifying the washing stage, quantifying the instability risk and dynamically matching the height strategy, the bending fatigue and the operation safety risk of the user are reduced; a safety height curve is constructed based on fluid motion prediction, shaking of the vegetable basin under dynamic load is effectively inhibited, and stability is enhanced; through weight sudden change detection and inclination angle spectrum analysis, compensation nodes are dynamically inserted, virtual damping is injected, height deviation is rapidly corrected, and stable adjustment of the vegetable basin in complex operation is ensured; by collecting electromyographic signals and task duration, user fatigue and work efficiency data are quantified, a height adjustment strategy is dynamically optimized, and human-computer interaction efficiency in long-term use is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent kitchen appliances, and in particular to a method and an adjustment system for adjusting the height of a multifunctional vegetable washing basin. Background Art

[0002] In the current smart kitchen appliance market, sink users, as the core of daily operations, face multiple challenges in optimizing their user experience and ensuring safety. Traditional sinks are fixed in height, requiring frequent bending or elevation during pre-rinsing, deep cleaning, and draining, which can lead to lumbar fatigue over time. Dynamic changes in the weight of ingredients and the impact of water flow within the sink can cause the sink to sway or tip over, posing a safety hazard. Traditional designs fail to fully consider ergonomic compensation, as users often lean forward or bend their lumbar spine during operation, resulting in insufficient operating height. Traditional sinks may only trigger adjustments using a single sensor (such as weight) and fail to integrate multi-dimensional data such as pressure, water flow, and posture. To address this issue, the present invention proposes a method and system for adjusting the height of a multifunctional sink. Summary of the Invention

[0003] The purpose of the present invention is to solve the problems in the background technology and to propose a method and an adjustment system for adjusting the height of a multifunctional vegetable washing basin.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions:

[0005] A method for adjusting the height of a multifunctional vegetable washing basin, comprising:

[0006] S1. Utilize a dynamic operation monitoring platform to capture multimodal operational behavior characteristics in real time, process the data through a data fusion algorithm, and output a dynamic load demand vector for the wash phase. The dynamic operation monitoring platform includes a pressure sensor array, a water flow measurement component, and a weight sensor array.

[0007] S2. Analyze the dynamic load demand vector and human posture characteristics, combine the washing stage markers to activate the corresponding strategy template, and quantitatively analyze the load instability risk coefficient and human body compensation value to ultimately generate a set of basin height offset factors;

[0008] S3: Integrate the basin height offset factor with the basin motion trajectory prediction, divide the risk level constraints into adjustment parameters, and construct a continuous height curve and discrete instruction sequence based on the probabilistic sloshing envelope of the random shallow water equation;

[0009] S4. Correct the altitude curve through weight mutation detection and inclination spectrum analysis, insert compensation nodes in real time and inject virtual damping, and output the anti-disturbance altitude instruction set;

[0010] S5. Collect electromyographic signals and task duration, quantify risk user fatigue and work efficiency improvement rate, locate inefficient strategy segments and provide feedback on optimized strategy templates.

[0011] Furthermore, the dynamic operation monitoring platform is used to capture multi-modal operation behavior characteristics in real time. After processing through the data fusion algorithm, the process of outputting the dynamic load demand vector with the washing stage includes:

[0012] The pressure sensor array at the edge of the basin monitors the pressure distribution on the basin during user operation in real time, with each pressure sensor node outputting a pressure-time waveform. Simultaneously, a water flow measurement component monitors the turbine speed, generates a raw pulse frequency signal, calculates the instantaneous flow value, and classifies the water flow state, i.e., continuous spray or intermittent dripping. Simultaneously, a weight sensor array at the four corners of the basin bottom records the total mass change of the ingredients in the basin in real time. The weight sensor array consists of four piezoelectric units arranged at the four corners of the basin bottom.

[0013] A sliding window Fourier transform is performed on the pressure-time waveform to extract the local pressure extreme points of each node, and noise signals smaller than a preset amplitude are filtered out. The extreme points of all nodes are aligned by timestamp to generate a spatiotemporal pressure event sequence. The spatiotemporal pressure event sequence is density clustered: in the spatial dimension, adjacent extreme points within adjacent time windows are merged; in the spatial dimension, the pressure values ​​of adjacent sensor nodes are aggregated; and the aggregated pressure peak sequence is output to characterize the distribution characteristics of user operation force.

[0014] Normalize G consecutive instantaneous flow values ​​to eliminate interference from water pressure fluctuations; perform fast Fourier transform on the standardized flow sequence to extract the energy distribution of the frequency band; classify the water flow state into continuous spraying and intermittent dripping based on the spectral characteristics; and finally output the water flow frequency spectrum vector, i.e., the spectrum entropy value.

[0015] Output force data in real time through weight sensor array ;According to the basin geometry model, establish the moment balance equation to calculate the instantaneous center of mass offset ; Calculate the instantaneous offset vector Apply Kalman filter to suppress random fluctuations caused by food shaking and output the center of mass offset after stabilization ;

[0016] Dynamically assign feature weights through the attention mechanism: assign operation intensity weights based on the stage-sensitive function marked by the operation stage ; Obtain the flow state weight based on the nonlinear mapping of spectral entropy value ; Obtain the center of gravity stability weight based on the exponential decay function of the mass center offset ;

[0017] Extract the pressure peak sequence, water flow frequency spectrum and mass center offset; perform spatiotemporal coding on the pressure peak sequence to obtain the force distribution code ; The center of gravity stability index is calculated based on the normalized mass center offset ;

[0018] Generating dynamic load demand vectors through weighted fusion ,in is the force distribution code, W is the water flow state label, is the center of gravity stability index, T is the matrix transpose symbol;

[0019] At the same time, the current washing stage is marked including the pre-rinsing stage, the deep cleaning stage and the draining stage.

[0020] Furthermore, the dynamic load demand vector is analyzed, the corresponding strategy template is activated in combination with the washing stage marker, and the load instability risk coefficient and human body compensation value are quantitatively analyzed to finally generate a set of basin height offset factors. The process includes:

[0021] Receive dynamic load demand vector ; Get the washing stage mark and activate the height adjustment strategy template according to the washing stage mark: load the corresponding preset height strategy primitives in combination with the identified washing stage mark, including: when entering the pre-rinse stage, automatically activate the high-level primitive; if the deep cleaning stage is detected, switch to the middle-level primitive; and when it is identified as the draining stage, enable the low-level primitive; and then synchronously output the strategy label ;

[0022] Combined with dynamic load demand vector Calculate the load instability risk factor ;

[0023] The user's joint points are captured by a depth camera to locate the user's shoulder and hip joints; the body compensation value is calculated based on the joint points, that is, the vertical height loss caused by the user's lumbar spine bending. When the trunk leans forward When the torso forward lean angle is greater than the preset standard value, the waist protection mode is triggered and the vertical height loss Increase compensation margin; output basin height offset factor set .

[0024] Furthermore, the basin height offset factor is integrated with the basin motion trajectory prediction. By dividing the risk level constraints and adjusting the parameters, the process of constructing a continuous height curve and discrete instruction sequence based on the probabilistic sloshing envelope of the random shallow water equation includes the following:

[0025] Input basin height offset factor set ;

[0026] Obtain data related to the volume of the fluid in the basin. This data comes from the total mass change of the ingredients in the basin recorded by the weight sensor array; calculate the volume of the fluid in the basin based on the known density of the ingredients; and convert the calculated volume of the fluid in the basin into the depth of the fluid in the basin: depth of the fluid in the basin = volume of the fluid in the basin / area of ​​the basin bottom;

[0027] Based on the load instability risk factor The values ​​are divided into differentiated risk response levels to determine the constraint parameters for adjusting the height of the vegetable sink: setting the risk response threshold range ,in are the lower and upper limits of the risk response threshold respectively; when the load instability risk coefficient The value is less than or equal to the risk response threshold If the load instability risk factor is The value is to Between, it is considered a medium risk state, and the damping constraint mechanism is enabled; and when the load instability risk coefficient Value reaches or exceeds When the device is in a high-risk state, the height adjustment function of the sink will be frozen and the safety lock mechanism will be activated.

[0028] Retrieve the historical operation trajectory library, load the historical operation trajectory of the current washing stage, and extract the historical disturbance pattern; obtain the pre-built random shallow water equation model; substitute the depth of the fluid in the basin into the random shallow water equation to solve it, calculate the center of mass motion of the fluid in the basin in the next J seconds, and predict the center of mass motion envelope in the next J seconds ;

[0029] Vertical height loss As the base height, construct the target height function:

[0030] , where is the target height; is the geometric center of the basin; is the envelope following weight; under the constraints of maximum adjustment speed and maximum acceleration, the target height function is optimized to generate a continuous height curve , that is, the height change curve, whose discretized instruction sequence is marked as ,in is the time series index.

[0031] Furthermore, the altitude curve is corrected through weight mutation detection and inclination spectrum analysis, compensation nodes are inserted in real time, virtual damping is injected, and the process of outputting the anti-disturbance altitude instruction set includes:

[0032] Input discrete instruction sequence ; Real-time acquisition of weight sensor data and basin tilt sensor flow data; Among them, the weight sensor data reflects the real-time weight changes of the load in the vegetable washing basin, and the basin tilt sensor flow data is used to monitor the changes in the tilt angle of the basin under different operations;

[0033] Monitor the weight sensor value in real time. When a sudden change in weight is detected, a dynamic backtracking mechanism is triggered: freeze the current height of the sink. ; Recalculate the load instability risk coefficient based on the latest dynamic load demand vector V ;

[0034] If the recalculated load instability risk factor indicates an increased risk level, then The remaining sequence is inserted into the compensation node to obtain the target height after compensation ; Perform spectrum analysis on the basin tilt sensor flow data. If a resonance frequency band is detected, inject a virtual damping factor into the corresponding period: modify the speed command to , where is the speed command after injecting damping; is the original speed instruction; is the damping ratio, which is used to control the damping strength; is the detected resonance frequency; is the sampling frequency; the compensation node and the damping correction term are combined to integrate the instructions after weight mutation compensation and virtual damping injection to generate the anti-disturbance height instruction set .

[0035] Furthermore, the process of collecting electromyographic signals and task duration, quantifying risk user fatigue and work efficiency improvement, locating inefficient strategy segments, and providing feedback for optimizing strategy templates includes:

[0036] Input anti-disturbance altitude command set Execution log;

[0037] Obtain the user's muscle surface electromyographic signal and extract the fatigue accumulation index from the electromyographic signal: collect the erector spinae surface electromyographic signal through the armrest electrode; perform multi-layer discrete wavelet decomposition on the collected electromyographic signal; extract low-frequency energy from the decomposed signal. Low-frequency energy is related to the fatigue state of the muscle, and its energy ratio is used to reflect the degree of fatigue accumulation of the muscle; calculate the fatigue accumulation index ;

[0038] Quantitative evaluation of work efficiency: record the actual time it takes for users to complete the task of washing the vegetable sink , the actual task duration Compared with historical duration benchmark Compare and calculate the efficiency improvement rate When the efficiency improvement rate Less than the preset ergonomic threshold and fatigue accumulation index Greater than the preset fatigue threshold When , the current operation phase is determined to be an inefficient strategy phase;

[0039] Reversely optimize the policy template parameters corresponding to the washing stage: locate the washing stage mark corresponding to the inefficient policy segment; adjust the policy template parameters according to the washing stage mark corresponding to the located inefficient policy segment; finally generate a policy template parameter update package and synchronize it to step S2.

[0040] A multifunctional vegetable sink height adjustment system, comprising:

[0041] Multimodal fusion processing module: Utilizes the dynamic operation monitoring platform to capture multimodal operation behavior characteristics in real time. After processing through the data fusion algorithm, it outputs the dynamic load demand vector with the washing stage.

[0042] Load risk strategy analysis module: This module analyzes the dynamic load demand vector, activates the corresponding strategy template based on the wash phase marker, and quantitatively analyzes the load instability risk coefficient and human body compensation value to ultimately generate a set of basin height offset factors.

[0043] Sloshing trajectory dynamic analysis module: This module integrates the basin height offset factor and basin motion trajectory prediction, constrains and adjusts parameters by dividing risk levels, and constructs a continuous height curve and discrete instruction sequence based on the probabilistic sloshing envelope of the random shallow water equation;

[0044] Anti-disturbance altitude real-time correction module: This module corrects the altitude curve through weight mutation detection and inclination spectrum analysis, inserts compensation nodes and injects virtual damping in real time, and outputs an anti-disturbance altitude instruction set;

[0045] Strategy feedback update optimization module: collects electromyographic signals and task duration, quantifies risk user fatigue and work efficiency improvement rate, locates inefficient strategy segments and provides feedback for optimization strategy templates.

[0046] Compared with the existing technology, the advantages of the method and adjustment system for adjusting the height of a multifunctional vegetable washing basin provided by the present invention are:

[0047] The present invention uses a pressure sensor array, a water flow measurement component, and a weight sensor to monitor operational behavior in real time, identify washing stages (pre-rinse, deep clean, drain) and load changes, provide a data basis for height adjustment, and avoid operational safety hazards caused by unstable loads. It combines the dynamic load demand vector with human posture characteristics to quantify the load instability risk factor and human compensation value, and automatically activates the corresponding height strategy templates (high, medium, and low positions), reducing user muscle fatigue caused by bending or improper operation force, and improving user comfort. It constructs the center of mass motion envelope based on the random shallow water equation, divides risk levels, and generates a continuous height curve to achieve smooth adjustment of the sink under dynamic loads, avoiding the risk of splashing or tipping due to fluid shaking. Through weight mutation detection and inclination spectrum analysis, compensation nodes are dynamically inserted and virtual damping is injected to quickly correct height adjustment errors, ensure the stability of the sink in complex operating environments, and improve user experience. It collects electromyographic signals and task duration, quantifies user fatigue and work efficiency improvement rate, locates inefficient strategy segments, and iteratively optimizes strategy templates, thereby reducing user labor intensity. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flow chart of a method for adjusting the height of a multi-functional vegetable washing basin proposed by the present invention.

[0049] Figure 2 This is a module diagram of a multifunctional vegetable sink height adjustment system proposed by the present invention. DETAILED DESCRIPTION

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the implementation regulations described are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0051] See also Figure 1 The present invention provides a method for adjusting the height of a multifunctional vegetable washing basin, comprising:

[0052] S1. Utilize a dynamic operation monitoring platform to capture multimodal operational behavior characteristics in real time, process the data through a data fusion algorithm, and output a dynamic load demand vector for the wash phase. The dynamic operation monitoring platform includes a pressure sensor array, a water flow measurement component, and a weight sensor array.

[0053] S2. Analyze the dynamic load demand vector and human posture characteristics, combine the washing stage markers to activate the corresponding strategy template, and quantitatively analyze the load instability risk coefficient and human body compensation value to ultimately generate a set of basin height offset factors;

[0054] S3: Integrate the basin height offset factor with the basin motion trajectory prediction, divide the risk level constraints into adjustment parameters, and construct a continuous height curve and discrete instruction sequence based on the probabilistic sloshing envelope of the random shallow water equation;

[0055] S4. Correct the altitude curve through weight mutation detection and inclination spectrum analysis, insert compensation nodes in real time and inject virtual damping, and output the anti-disturbance altitude instruction set;

[0056] S5. Collect electromyographic signals and task duration, quantify risk user fatigue and work efficiency improvement rate, locate inefficient strategy segments, and provide feedback for iterative optimization strategy templates.

[0057] It should be further explained that, in the specific implementation process, the dynamic operation monitoring platform is used to capture multi-modal operation behavior characteristics in real time. After processing through the data fusion algorithm, the process of outputting the dynamic load demand vector with the washing stage includes:

[0058] The pressure distribution of the basin during user operation is monitored in real time by a pressure sensor array on the basin edge (sampling at a 100Hz rate). Each pressure sensor node outputs a pressure-time waveform. A water flow measurement component simultaneously monitors the turbine speed, generates a raw pulse frequency signal, calculates the instantaneous flow value, and classifies the water flow state as continuous spray or intermittent dripping. Simultaneously, a weight sensor array at the four corners of the basin floor records the total mass change of the ingredients in the sink in real time. The weight sensor array consists of four piezoelectric units arranged at the four corners of the basin floor.

[0059] Perform a sliding window Fourier transform (e.g., window width 0.1 seconds) on the pressure-time waveform to extract the local pressure extreme points of each node, and filter out noise signals less than a preset amplitude (e.g., 5 kPa). Align the extreme points of all nodes by timestamp to generate a spatiotemporal pressure event sequence: {timestamp, sensor ID, pressure value}. Density clustering is performed on the spatiotemporal pressure event sequence: in the spatial dimension, adjacent extreme points within adjacent time windows (e.g., 300 ms) are merged; in the spatial dimension, the pressure values ​​of adjacent sensor nodes (e.g., spacing < 2 cm) are aggregated. The aggregated pressure peak sequence is output: {peak start time, duration, peak intensity mean, action area coordinates}, which is used to characterize the distribution characteristics of user operation force.

[0060] Normalize G (e.g., 30) consecutive instantaneous flow values ​​to eliminate interference from water pressure fluctuations. Perform a fast Fourier transform on the standardized flow sequence to extract the energy distribution of the frequency band (e.g., 0.5-10 Hz). Classify the flow state based on the spectral characteristics: when the main frequency energy accounts for >85% (e.g., 2-4 Hz), it is marked as continuous jetting; when the energy exhibits a multi-band energy distribution (e.g., 0.5-1 Hz and 3-5 Hz), it is marked as intermittent dripping. Finally, the water flow frequency spectrum vector, i.e., the spectrum entropy value, is output.

[0061] Output force data in real time through weight sensor array ;According to the basin geometry model, establish the moment balance equation to calculate the instantaneous center of mass offset: total mass , the horizontal coordinate offset of the center of mass , the horizontal coordinate offset of the center of mass ,in are the length and width of the pelvic floor, and g is the acceleration of gravity; calculate the instantaneous offset vector: , where X0, Y0 are the centroid references of the empty basin;

[0062] Calculate the instantaneous offset vector Apply the Kalman filter to suppress random fluctuations caused by food shaking (such as water impact and temporary displacement caused by user operation) and output the center of mass offset after stabilization. ;

[0063] Dynamically assign feature weights through the attention mechanism: Assign operation strength weight Wp based on the stage-sensitive function marked by the operation stage: , where Binary mark for the washing stage (e.g. pre-rinsing stage = 1, other stages = 0), is the stage sensitivity threshold (e.g. pre-flushing stage = 1.5, others = 0.5); is the Sigmoid activation function;

[0064] The water flow state weight Ww is obtained based on the nonlinear mapping of the spectrum entropy value: ,in , where The hyperbolic tangent function maps the spectrum entropy value to the (-1,1) interval, and then normalizes it by the normalization coefficient Adjust the range and offset to get a weight value of (-1,1); is the normalization coefficient; is the water flow spectrum entropy value (range [0,1]); For the The energy proportion of each frequency band; is the frequency band index, is the total number of frequency bands;

[0065] Obtain the center of gravity stability weight Wm based on the exponential decay function of the center of mass offset: , where is the center of gravity stability reference parameter, which represents the reference value of the ideal center of mass offset; The parameter that controls the decay speed of the Sigmoid function determines the sensitivity of the center of gravity stabilization weight to the change of the offset;

[0066] Extract the pressure peak sequence, water flow frequency spectrum, and mass center offset; perform spatiotemporal encoding on the pressure peak sequence: , where Force distribution encoding is used to characterize the spatiotemporal distribution characteristics of user operation force; is the peak timestamp; is the normalized pressure value; The pressure action area is the area of ​​the pressure peak sequence. During the density clustering process, the pressure values ​​of adjacent sensor nodes are merged to form a continuous pressure action area. The geometric area of ​​the area is estimated based on the physical layout of the sensor nodes (i.e., the coordinates of the four corners of the basin bottom) and the coverage after clustering. For spatiotemporal encoding, the timestamps, pressure values, and spatial action areas in the pressure peak sequence are mapped into structured feature vectors. The center of gravity stability index is calculated based on the normalized mass center offset: , where is the maximum and minimum value of the center of mass offset;

[0067] Generating dynamic load demand vectors through weighted fusion , where P is the force distribution code, W is the water flow state label (continuous spray or intermittent dripping), M is the center of gravity stability index, and T is the matrix transpose symbol;

[0068] At the same time, the current washing stage (pre-rinse, deep clean, drain) is marked. The washing stage marking is determined as follows: if and only if the following conditions are met at the same time, it is determined to be the pre-rinse stage: the water flow state is identified as continuous spray (the main frequency energy accounts for >85% and the frequency band is concentrated in 2-4Hz), the average pressure peak intensity collected by the pressure sensor array is lower than the preset pressure peak threshold (for example, 8kPa, indicating gentle operation), if the center of gravity stability index exceeds the preset center of gravity stability threshold (for example, 0.7, reflecting that the food has not undergone significant displacement); if and only if the following conditions are met at the same time, it is determined to be the deep cleaning stage: the water flow state is identified as intermittent dripping (energy is distributed in the dual frequency bands of 0.5-1Hz and 3-5Hz) The pressure sensing array detects a pressure peak sequence (reflecting a sudden change in force) that lasts longer than Z (Z is the time parameter detected by the pressure sensing array, which is used to define the duration threshold of the pressure peak sequence in the pressure sensing array detection, such as 3 seconds), and the food mass offset recorded by the weight sensor is greater than the preset offset threshold (for example, 15mm indicates that the food is displaced due to water impact). When the following feature combination is detected, it is determined to be the draining stage: the water flow state is identified as intermittent dripping (main frequency <1Hz), the pressure peak disappears, the pressure sensing array has no valid pressure peak input (the operation is completely stopped), and the weight sensor detects that the mass center offset is showing a continuous downward trend (water discharge causes food to settle).

[0069] It should be further explained that, in the specific implementation process, the dynamic load demand vector is parsed, the corresponding strategy template is activated in combination with the washing stage marker, and the load instability risk coefficient and human body compensation value are quantitatively analyzed to finally generate the basin height offset factor set. The process includes:

[0070] Receive dynamic load demand vector ; Get the washing stage mark, and activate the height adjustment strategy template according to the washing stage mark: Combine the recognized washing stage mark to load the corresponding preset height strategy primitive, including: when entering the pre-rinsing stage, automatically activate the high-position primitive (the default height from the ground is set to 110cm to effectively reduce the inconvenience of users bending over to operate); if the deep cleaning stage is detected, switch to the middle primitive (the height from the ground is adjusted to 90cm, while ensuring the operating force and maintaining the stability of the vegetable washing basin); and when it is identified as the draining stage, enable the low-position primitive (the height from the ground is reduced to 70cm, making full use of gravity to achieve efficient drainage); and then synchronously output the strategy label containing the primitive ID and the corresponding height value ;

[0071] Combined with dynamic load demand vector Calculate the load instability risk factor :

[0072] ,

[0073] Where, is the Sigmoid activation function; the feature embedding layer includes multi-layer perceptron embedding of the strength distribution code P , the unique hot encoding of the water flow state label W And the embedded representation of the center of gravity stability index M ,in The embedding dimension of the force distribution encoding is set to 128, the one-hot encoding dimension of the water flow state label is set to 2, and the embedding dimension of the center of gravity stability index is set to 64. It is a dataset used to declare that the embedded features belong to the real space; attention fusion mechanism ,in The goal of the function is to calculate the attention weight matrix, which is used to weight the importance of different features. They are query matrix (used to calculate attention weights), key matrix (used to match queries), and value matrix (used for weighted aggregation). is the time series dimension index, is the dimension of the key vector and query vector (i.e. the dimension of a single attention head), is the time series risk sensitivity coefficient; when the load instability risk coefficient When the preset risk threshold is exceeded, it indicates that the current load is at risk of instability and an overturning warning sign is added;

[0074] The depth camera is used to capture the user's joint points and locate the user's shoulder joint (T1 vertebra projection point) and hip joint (L4 vertebra projection point). The depth camera can accurately obtain the position information of the human joints in three-dimensional space, providing accurate data for subsequent calculations; the human body compensation value is calculated based on the joint points, that is, the vertical height loss caused by the user's lumbar spine bending. :

[0075] ,

[0076] Where, is the shoulder joint coordinate; is the hip joint coordinate; is the forward lean angle of the trunk; is the biomechanical correction factor; is the spinal curvature compensation term (obtained by fitting a function to magnetic resonance imaging data). This formula calculates the distance between the projection points of the shoulder joint and the hip joint and the forward lean angle of the trunk to obtain the vertical height lost by the user due to lumbar curvature. When the angle is greater than the preset standard value of the trunk forward tilt (for example, 25°), it indicates that the user's trunk is leaning forward at a large angle, and the pressure on the lumbar spine increases. At this time, the waist protection mode is triggered, and the vertical height loss Increase the compensation margin to further adjust the basin height and reduce the burden on the user's waist; output the basin height offset factor set ; It can be understood that this set integrates multiple factors such as height adjustment strategy, load instability risk factor and human body compensation value, providing comprehensive and accurate data support for the subsequent precise adjustment of the basin height.

[0077] It should be further explained that, in the specific implementation process, the basin height offset factor is integrated with the basin motion trajectory prediction, the parameters are adjusted by dividing the risk level constraints, and the probabilistic sloshing envelope based on the random shallow water equation is constructed into a continuous height curve and a discrete instruction sequence. The process includes:

[0078] Input basin height offset factor set The strategy label specifies the height adjustment strategy corresponding to the current washing stage, the load instability risk coefficient reflects the stability of the sink load, and the human body compensation value reflects the vertical height loss caused by human posture.

[0079] Obtain data related to the volume of the fluid in the basin. This data comes from the total mass change of the ingredients in the basin recorded by the weight sensor array. Combined with the known density of the ingredients, the volume of the fluid in the basin is calculated for subsequent center of mass motion prediction. The calculated volume of the fluid in the basin is converted to the depth of the fluid in the basin: the depth of the fluid in the basin = the volume of the fluid in the basin / the area of ​​the basin bottom.

[0080] Based on the load instability risk factor The values ​​are divided into differentiated risk response levels to determine the constraint parameters for adjusting the height of the vegetable sink: setting the risk response threshold range (e.g. [0.3, 0.6]), where are the lower and upper limits of the risk response threshold respectively; when the load instability risk coefficient The value is less than or equal to the risk response threshold When the load instability risk factor is , it is judged to be a low risk state, and the vegetable washing basin is allowed to be raised and lowered at a maximum adjustment speed of 8cm / s; The value is to If the load instability risk factor is between 0.01 and 0.11, it is considered a medium risk state, and the damping constraint mechanism is activated to limit the maximum adjustment speed to 5cm / s and ensure that the maximum acceleration does not exceed 0.4g to ensure the stability of the adjustment process; Value reaches or exceeds When the height adjustment function of the sink is frozen, the safety lock mechanism is activated to prevent accidents caused by load instability.

[0081] Retrieve the historical operation trajectory library, load the historical operation trajectory of the current washing stage, and extract historical disturbance patterns (such as the sinusoidal shaking caused by rapid scrubbing and the attenuated oscillation caused by water injection shock) as the external force input of the random shallow water equation. It can be understood that the historical operation trajectory records the user's operating habits and methods of the vegetable washing basin in different washing stages. By analyzing these trajectories, historical disturbance patterns can be extracted, such as the sinusoidal shaking caused by rapid scrubbing and the attenuated oscillation caused by water injection shock. These typical disturbance patterns can provide a reference for the current center of mass motion prediction.

[0082] Obtain a pre-built stochastic shallow water equation model. This model is based on the geometry of the sink and the physical properties of the fluid, and can describe the sloshing behavior of the fluid in the sink under different external forces.

[0083] Substitute the depth of the fluid in the basin into the random shallow water equation and solve it to calculate the center of mass motion of the fluid in the basin in the next J seconds (J is the prediction time window), and predict the envelope of the center of mass motion in the next J seconds. :

[0084] , ,

[0085] Where, is the expected value; is the variance; is the position of the fluid centroid, which is obtained by solving the random shallow water equation statistically multiple times; is the fluid state information as of time t; is the current moment; the envelope reflects the possible movement range of the center of mass of the fluid in the basin in the future, providing important information for the subsequent generation of the height curve;

[0086] It is understandable that the random shallow water equation is specifically: based on the geometric shape of the sink (such as rectangular or circular cross-section) and the depth of the fluid, a two-dimensional shallow water equation is established , where is the depth of the fluid in the basin, Level and vertical Directional flow velocity, is the acceleration due to gravity, is the fluid density, is the fluid pressure field, The external force term, i.e., the external disturbance force (e.g., the sinusoidal force generated by the user brushing, the impact force of water injection), is integrated with the historical disturbance patterns in the historical operation trajectory library, such as sinusoidal shaking and decaying oscillation.

[0087] Vertical height loss As the base height, construct the target height function:

[0088] ,

[0089] Where, is the target height; The geometric center of the basin, that is, the fixed reference point of the vegetable sink; The envelope following weight is used to adjust the degree to which the target height follows the envelope of the center of mass motion. This function can make the height of the vegetable sink meet the ergonomic requirements while better adapting to the movement changes of the center of mass of the fluid in the basin, thereby improving the stability of the adjustment. Under the constraints of maximum adjustment speed and maximum acceleration, the target height function is optimized to generate a continuous height curve. , that is, the height change curve, whose discretized instruction sequence is marked as ,in is the time series index; the discretization process is: set the control period ,Will According to the time scale Discretize into a sequence .

[0090] It should be further explained that, in the specific implementation process, the altitude curve is corrected through weight mutation detection and inclination spectrum analysis, compensation nodes are inserted in real time, virtual damping is injected, and the process of outputting the anti-disturbance altitude instruction set includes:

[0091] Input discrete instruction sequence ; Real-time acquisition of weight sensor data and basin tilt sensor flow data; Among them, the weight sensor data reflects the real-time weight changes of the load in the vegetable washing basin, and the basin tilt sensor flow data is used to monitor the changes in the tilt angle of the basin under different operations;

[0092] Monitor the weight sensor value in real time. When a sudden change in weight is detected (i.e. the weight sensor value is greater than the standard weight value), a dynamic backtracking mechanism is triggered: the current height of the sink is frozen. , to prevent further aggravation of load instability due to improper height adjustment when the load changes suddenly; recalculate the load instability risk coefficient based on the latest dynamic load demand vector V It is understandable that by recalculating the load instability risk coefficient, the stability of the sink after a sudden load change can be reflected in a timely manner, providing an accurate basis for subsequent adjustment strategies.

[0093] If the recalculated load instability risk factor indicates an increased risk level, then Insert the remaining sequence into the compensation node:

[0094] ,

[0095] Where, is the target height after compensation; They are proportional gain (i.e. the proportional coefficient for amplifying the error signal, used to quickly respond to deviations) and differential gain (i.e. the coefficient for suppressing the error change rate, used to suppress overshoot); is the height tracking error, ,in are the target center of mass offset and the actual center of mass offset respectively; is the sampling period, that is, the time interval between collecting sensor data and executing control instructions; it is understandable that by inserting the compensation node, the height of the sink can be appropriately adjusted when the load changes suddenly, reducing the risk of load instability; the basin tilt sensor stream data is spectrally analyzed to detect possible resonance frequency bands (for example, during the scrubbing operation, a resonance frequency band of 4-6 Hz caused by scrubbing vibration may appear; the resonance phenomenon will aggravate the shaking of the basin and affect the stability of the sink); during the detected resonance period, a virtual damping factor is injected to suppress mechanical resonance: the speed instruction is modified to

[0096] ,

[0097] Where, is the speed command after injecting damping; is the original speed instruction; is the damping ratio, which is used to control the damping strength; is the detected resonance frequency; is the sampling frequency, that is, the frequency of collecting sensor data; it is understandable that by injecting virtual damping, the damping characteristics of the system can be changed, the resonance amplitude can be reduced, and the stability of the vegetable sink can be improved; the compensation node and the damping correction term are combined to integrate the instructions after weight mutation compensation and virtual damping injection to generate an anti-disturbance height instruction set. Among them, the instruction set comprehensively considers the impact of factors such as load mutation and mechanical resonance on the height adjustment of the sink, and can provide more stable and accurate control instructions.

[0098] It should be further explained that, during the specific implementation process, the process of collecting electromyographic signals and task duration, quantifying risk user fatigue and work efficiency improvement, locating inefficient strategy segments, and providing feedback for iterative optimization of strategy templates includes:

[0099] Input anti-disturbance altitude command set Execution logs, which record detailed information about the height adjustment of the sink at different time points according to the optimized instructions, providing basic data for subsequent analysis of user fatigue and ergonomics during operation;

[0100] Obtain the user's muscle surface electromyographic signal and extract the fatigue accumulation index (low-frequency energy ratio) in the electromyographic signal: collect the erector spinae surface electromyographic signal through the armrest electrode (the sampling rate is set to 1kHz); it is understandable that the erector spinae muscles play an important supporting and movement function during the operation of the vegetable sink, and their electromyographic signals can better reflect the user's muscle fatigue state; perform multi-layer (for example, 5-layer) discrete wavelet decomposition on the collected electromyographic signal. Discrete wavelet decomposition can decompose the signal into different frequency bands, which is convenient for extracting characteristic information of specific frequency bands; extract low-frequency energy (for example, 0.5-2Hz) from the decomposed signal. Low-frequency energy is related to the fatigue state of the muscle, and its energy ratio is used to reflect the degree of fatigue accumulation of the muscle; calculate the fatigue accumulation index :

[0101] ,

[0102] Where, Indicates that the integral time range is the start of the task To the end ; The total duration of a single operation; It is low-frequency myoelectric energy, such as 0.5-2Hz low-frequency energy; is the total myoelectric energy; is the maximum trunk forward lean angle change; The critical flexion angle of the lumbar spine (e.g., 30°) is used. This index can be used to quantify the user's muscle fatigue level, providing a basis for subsequent identification of inefficient strategy segments.

[0103] Quantitative evaluation of work efficiency: record the actual time it takes for users to complete the task of washing the vegetable sink , the actual task duration Compared with historical duration benchmark Compare and calculate the efficiency improvement rate :

[0104] ,

[0105] Where, is the work efficiency weight coefficient, is the time weight coefficient; when the efficiency improvement rate Less than the preset ergonomic threshold and fatigue accumulation index Greater than the preset fatigue threshold When , the current operation phase is determined to be an inefficient strategy phase;

[0106] Reversely optimize the policy template parameters corresponding to the washing stage: locate the washing stage mark corresponding to the inefficient policy segment; it is understandable that by analyzing the execution log of the anti-disturbance height instruction set Hopt[k], the specific position of the inefficient policy segment in the entire washing process of the vegetable washing basin is determined, providing accurate positioning information for the subsequent policy template parameter adjustment; according to the washing stage mark corresponding to the located inefficient policy segment, adjust the policy template parameters. Specific adjustment measures include reducing the default height of high-risk operations to reduce the probability of the vegetable washing basin overturning and ensure operational safety; increase the human body compensation weight, and by optimizing the height adjustment strategy, reduce the load on the user's lumbar spine during operation and improve operational comfort; finally generate a policy template parameter update package and synchronize it to step S2.

[0107] See also Figure 2 The present invention provides a multifunctional vegetable sink height adjustment system, comprising:

[0108] Multimodal fusion processing module: Utilizes the dynamic operation monitoring platform to capture multimodal operation behavior characteristics in real time. After processing through the data fusion algorithm, it outputs the dynamic load demand vector with the washing stage.

[0109] Load risk strategy analysis module: This module analyzes the dynamic load demand vector, activates the corresponding strategy template based on the wash phase marker, and quantitatively analyzes the load instability risk coefficient and human body compensation value to ultimately generate a set of basin height offset factors.

[0110] Sloshing trajectory dynamic analysis module: This module integrates the basin height offset factor and basin motion trajectory prediction, constrains and adjusts parameters by dividing risk levels, and constructs a continuous height curve and discrete instruction sequence based on the probabilistic sloshing envelope of the random shallow water equation;

[0111] Anti-disturbance altitude real-time correction module: This module corrects the altitude curve through weight mutation detection and inclination spectrum analysis, inserts compensation nodes and injects virtual damping in real time, and outputs an anti-disturbance altitude instruction set;

[0112] Strategy feedback update optimization module: collects electromyographic signals and task duration, quantifies risk user fatigue and work efficiency improvement rate, locates inefficient strategy segments and provides feedback to iteratively optimize strategy templates.

[0113] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0114] It should be understood that determining B based on A does not mean determining B based solely on A. B can also be determined based on A and / or other information.

[0115] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0116] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for adjusting the height of a multifunctional vegetable washing basin, characterized by: S1. Utilize a dynamic operation monitoring platform to capture multimodal operational behavior characteristics in real time, process the data through a data fusion algorithm, and output a dynamic load demand vector for the wash phase. The dynamic operation monitoring platform includes a pressure sensor array, a water flow measurement component, and a weight sensor array. S2. Analyze the dynamic load demand vector and human posture characteristics, combine the washing stage markers to activate the corresponding strategy template, and quantitatively analyze the load instability risk coefficient and human body compensation value to ultimately generate a set of basin height offset factors; S3: Integrate the basin height offset factor with the basin motion trajectory prediction, divide the risk level constraints into adjustment parameters, and construct a continuous height curve and discrete instruction sequence based on the probabilistic sloshing envelope of the random shallow water equation; S4. Correct the altitude curve through weight mutation detection and inclination spectrum analysis, insert compensation nodes in real time and inject virtual damping, and output the anti-disturbance altitude instruction set; S5. Collect electromyographic signals and task duration, quantify risk user fatigue and work efficiency improvement rate, locate inefficient strategy segments and provide feedback on optimized strategy templates.

2. The method for adjusting the height of a multifunctional vegetable washing basin according to claim 1, characterized in that: The process of using the dynamic operation monitoring platform to capture multi-modal operation behavior characteristics in real time and then processing them through the data fusion algorithm to output the dynamic load demand vector for the washing phase includes: The pressure sensor array at the edge of the basin monitors the pressure distribution on the basin during user operation in real time, with each pressure sensor node outputting a pressure-time waveform. Simultaneously, a water flow measurement component monitors the turbine speed, generates a raw pulse frequency signal, calculates the instantaneous flow value, and classifies the water flow state, i.e., continuous spray or intermittent dripping. Simultaneously, a weight sensor array at the four corners of the basin bottom records the total mass change of the ingredients in the basin in real time. The weight sensor array consists of four piezoelectric units arranged at the four corners of the basin bottom. A sliding window Fourier transform is performed on the pressure-time waveform to extract the local pressure extreme points of each node, and noise signals smaller than a preset amplitude are filtered out. The extreme points of all nodes are aligned by timestamp to generate a spatiotemporal pressure event sequence. The spatiotemporal pressure event sequence is density clustered: in the spatial dimension, adjacent extreme points within adjacent time windows are merged; in the spatial dimension, the pressure values ​​of adjacent sensor nodes are aggregated; and the aggregated pressure peak sequence is output to characterize the distribution characteristics of user operation force. Normalize G consecutive instantaneous flow values ​​to eliminate interference from water pressure fluctuations; perform fast Fourier transform on the standardized flow sequence to extract the energy distribution of the frequency band; classify the water flow state into continuous spraying and intermittent dripping based on the spectral characteristics; and finally output the water flow frequency spectrum vector, i.e., the spectrum entropy value. Output force data in real time through weight sensor array ;According to the basin geometry model, establish the moment balance equation to calculate the instantaneous center of mass offset ; Calculate the instantaneous offset vector Apply Kalman filter to suppress random fluctuations caused by food shaking and output the center of mass offset after stabilization ; Dynamically assign feature weights through the attention mechanism: assign operation intensity weights based on the stage-sensitive function marked by the operation stage ; Obtain the flow state weight based on the nonlinear mapping of spectral entropy value ; Obtain the center of gravity stability weight based on the exponential decay function of the mass center offset ; Extract the pressure peak sequence, water flow frequency spectrum and mass center offset; perform spatiotemporal coding on the pressure peak sequence to obtain the force distribution code ; The center of gravity stability index is calculated based on the normalized mass center offset ; Generating dynamic load demand vectors through weighted fusion ,in is the force distribution code, W is the water flow state label, is the center of gravity stability index, T is the matrix transpose symbol; At the same time, the current washing stage is marked including the pre-rinsing stage, the deep cleaning stage and the draining stage.

3. The method for adjusting the height of a multifunctional vegetable washing basin according to claim 1, characterized in that: The process of analyzing the dynamic load demand vector, activating the corresponding strategy template based on the wash phase marker, and quantitatively analyzing the load instability risk factor and human body compensation value to ultimately generate a set of basin height offset factors includes: Receive dynamic load demand vector ; Get the washing stage mark and activate the height adjustment strategy template according to the washing stage mark: load the corresponding preset height strategy primitives in combination with the identified washing stage mark, including: when entering the pre-rinse stage, automatically activate the high-level primitive; if the deep cleaning stage is detected, switch to the middle-level primitive; and when it is identified as the draining stage, enable the low-level primitive; and then synchronously output the strategy label ; Combined with dynamic load demand vector Calculate the load instability risk factor ; The user's joint points are captured by a depth camera to locate the user's shoulder and hip joints; the body compensation value is calculated based on the joint points, that is, the vertical height loss caused by the user's lumbar spine bending. When the trunk leans forward When the torso forward lean angle is greater than the preset standard value, the waist protection mode is triggered and the vertical height loss Increase compensation margin; output basin height offset factor set .

4. The method for adjusting the height of a multifunctional vegetable washing basin according to claim 1, characterized in that: The process of integrating the basin height offset factor with the basin motion trajectory prediction, dividing the risk level constraints and adjusting the parameters, and constructing a continuous height curve and discrete instruction sequence based on the probabilistic sloshing envelope of the random shallow water equation includes the following: Input basin height offset factor set ; Obtain data related to the volume of the fluid in the basin. This data comes from the total mass change of the ingredients in the basin recorded by the weight sensor array; calculate the volume of the fluid in the basin based on the known density of the ingredients; and convert the calculated volume of the fluid in the basin into the depth of the fluid in the basin: depth of the fluid in the basin = volume of the fluid in the basin / area of ​​the basin bottom; Based on the load instability risk factor The values ​​are divided into differentiated risk response levels to determine the constraint parameters for adjusting the height of the vegetable sink: setting the risk response threshold range ,in are the lower and upper limits of the risk response threshold respectively; when the load instability risk coefficient The value is less than or equal to the risk response threshold If the load instability risk factor is The value is to Between, it is considered a medium risk state, and the damping constraint mechanism is enabled; and when the load instability risk coefficient Value reaches or exceeds When the device is in a high-risk state, the height adjustment function of the sink will be frozen and the safety lock mechanism will be activated. Retrieve the historical operation trajectory library, load the historical operation trajectory of the current washing stage, and extract the historical disturbance pattern; obtain the pre-built random shallow water equation model; substitute the depth of the fluid in the basin into the random shallow water equation to solve it, calculate the center of mass motion of the fluid in the basin in the next J seconds, and predict the center of mass motion envelope in the next J seconds ; Vertical height loss As the base height, construct the target height function: , where is the target height; is the geometric center of the basin; is the envelope following weight; under the constraints of maximum adjustment speed and maximum acceleration, the target height function is optimized to generate a continuous height curve , that is, the height change curve, whose discretized instruction sequence is marked as ,in is the time series index.

5. The method for adjusting the height of a multifunctional vegetable washing basin according to claim 1, characterized in that: The process of correcting the altitude curve through weight mutation detection and inclination spectrum analysis, inserting compensation nodes and injecting virtual damping in real time, and outputting the anti-disturbance altitude instruction set includes: Input discrete instruction sequence ; Real-time acquisition of weight sensor data and basin tilt sensor flow data; Among them, the weight sensor data reflects the real-time weight changes of the load in the vegetable washing basin, and the basin tilt sensor flow data is used to monitor the changes in the tilt angle of the basin under different operations; Monitor the weight sensor value in real time. When a sudden change in weight is detected, a dynamic backtracking mechanism is triggered: freeze the current height of the sink. ; Recalculate the load instability risk coefficient based on the latest dynamic load demand vector V ; If the recalculated load instability risk factor indicates an increased risk level, then The remaining sequence is inserted into the compensation node to obtain the target height after compensation ; Perform spectrum analysis on the basin tilt sensor flow data. If a resonance frequency band is detected, inject a virtual damping factor into the corresponding period: modify the speed command to , where is the speed command after injecting damping; is the original speed instruction; is the damping ratio, which is used to control the damping strength; is the detected resonance frequency; is the sampling frequency; the compensation node and the damping correction term are combined to integrate the instructions after weight mutation compensation and virtual damping injection to generate the anti-disturbance height instruction set .

6. The method for adjusting the height of a multifunctional vegetable washing basin according to claim 1, characterized in that: The process of collecting electromyographic signals and task duration, quantifying risk user fatigue and work efficiency improvement, locating inefficient strategy segments, and providing feedback for optimizing strategy templates includes: Input anti-disturbance altitude command set Execution log; Obtain the user's muscle surface electromyographic signal and extract the fatigue accumulation index from the electromyographic signal: collect the erector spinae surface electromyographic signal through the armrest electrode; perform multi-layer discrete wavelet decomposition on the collected electromyographic signal; extract low-frequency energy from the decomposed signal. Low-frequency energy is related to the fatigue state of the muscle, and its energy ratio is used to reflect the degree of fatigue accumulation of the muscle; calculate the fatigue accumulation index ; Quantitative evaluation of work efficiency: record the actual time it takes for users to complete the task of washing the vegetable sink , the actual task duration Compared with historical duration benchmark Compare and calculate the efficiency improvement rate When the efficiency improvement rate Less than the preset ergonomic threshold and fatigue accumulation index Greater than the preset fatigue threshold When , the current operation phase is determined to be an inefficient strategy phase; Reversely optimize the policy template parameters corresponding to the washing stage: locate the washing stage mark corresponding to the inefficient policy segment; adjust the policy template parameters according to the washing stage mark corresponding to the located inefficient policy segment; finally generate a policy template parameter update package and synchronize it to step S2.

7. A multifunctional vegetable sink height adjustment system, characterized in that: A method for adjusting the height of a multifunctional vegetable washing basin as claimed in any one of claims 1 to 6, the system comprising: Multimodal fusion processing module: Utilizes the dynamic operation monitoring platform to capture multimodal operation behavior characteristics in real time. After processing through the data fusion algorithm, it outputs the dynamic load demand vector with the washing stage. Load risk strategy analysis module: This module analyzes the dynamic load demand vector, activates the corresponding strategy template based on the wash phase marker, and quantitatively analyzes the load instability risk coefficient and human body compensation value to ultimately generate a set of basin height offset factors. Sloshing trajectory dynamic analysis module: This module integrates the basin height offset factor and basin motion trajectory prediction, constrains and adjusts parameters by dividing risk levels, and constructs a continuous height curve and discrete instruction sequence based on the probabilistic sloshing envelope of the random shallow water equation; Anti-disturbance altitude real-time correction module: This module corrects the altitude curve through weight mutation detection and inclination spectrum analysis, inserts compensation nodes and injects virtual damping in real time, and outputs an anti-disturbance altitude instruction set; Strategy feedback update optimization module: collects electromyographic signals and task duration, quantifies risk user fatigue and work efficiency improvement rate, locates inefficient strategy segments and provides feedback for optimization strategy templates.

Images