Intelligent plastic clothes hanger control method and system based on artificial intelligence

Abstract

The invention discloses an intelligent plastic clothes hanger control method and system based on artificial intelligence, and relates to the field of intelligent household electrical appliance data processing.The method comprises the following steps that original data of aired clothes are collected, and a multi-dimensional clothes characteristic spectrum is constructed through the tensor decomposition technology; based on the multi-dimensional clothes characteristic spectrum, generating a clothes airing strategy by using an ant colony algorithm; and determining a specific control scheme of the intelligent plastic clothes hanger by using a multi-objective optimization algorithm according to the clothes airing strategy. According to the method, the original data of the aired clothes are collected, the multi-dimensional clothes characteristic spectrum is constructed by using the tensor decomposition technology, then the clothes airing strategy is generated by using the ant colony algorithm based on the spectrum, and finally the specific control scheme is determined according to the clothes airing strategy, so that the intelligent management and optimization of the clothes airing process are realized; a personalized clothes airing scheme can be formulated according to different clothes characteristics and environmental conditions, and the clothes airing efficiency and the user experience are greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of smart home data processing, and in particular to an artificial intelligence-based smart plastic clothes drying rack control method and system. Background Art

[0002] With the improvement of people's living standards and the development of science and technology, home intelligence has gradually become a new trend in modern life. Various home appliances are gradually shifting from manual operation to automation and intelligent control to improve the quality and efficiency of life. Taking smart plastic clothes drying racks as an example, as an important part of smart home, it has attracted more and more attention from consumers. The traditional way of drying clothes often relies on manual judgment of weather and clothes drying conditions, which is not only time-consuming and laborious, but also the effect of drying clothes is affected by many factors. Therefore, researching and developing a smart clothes drying rack technology that can automatically adjust the clothes drying strategy and optimize the clothes drying process has become an important research direction in the current smart home field.

[0003] Currently, most smart plastic clothes drying racks on the market use simple sensors and preset rules to control the drying process, such as automatically retracting the clothes drying rack according to temperature and humidity. Although these technologies have improved the automation level of clothes drying to a certain extent, they can usually only handle limited parameters and cannot fully consider the complex factors in the clothes drying process, such as the material of the clothes, the drying history, and environmental changes. Therefore, these smart clothes drying rack control technologies still have a lot of room for improvement in adaptability and clothes drying effect.

[0004] Although the existing clothes drying rack control technology has improved the convenience of drying clothes to a certain extent, they generally have problems such as insufficient data processing capabilities and poor strategy adaptability. Existing clothes drying rack control technologies can often only make decisions based on limited parameter data, and cannot fully consider the multi-dimensional factors in the drying process, such as the material, thickness, and drying history of the clothes; in addition, the control strategies of these technologies are usually fixed and lack the ability to adapt to dynamic environmental changes, resulting in unsatisfactory drying effects and efficiency; due to the inability to effectively process and integrate multi-source data, existing clothes drying rack control technologies often cannot accurately judge the drying state of clothes and the best time to dry clothes, resulting in uneven drying of clothes and even possible damage to clothes; in addition, fixed clothes drying strategies cannot adapt to changing environmental conditions and clothing characteristics, which greatly reduces the effect of drying clothes.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In response to the problems in the related technology, the present invention proposes an intelligent plastic clothes drying rack control method and system based on artificial intelligence, which has the advantages of constructing a multi-dimensional clothing feature map using tensor decomposition technology and generating a clothes drying strategy using an ant colony algorithm, thereby realizing intelligent management and optimization of the clothes drying process, and being able to formulate personalized clothes drying plans according to different clothing characteristics and environmental conditions, greatly improving the clothes drying efficiency and user experience, thereby solving the problems in the prior art that decisions can only be made based on limited parameters, and the multi-dimensional factors in the clothes drying process cannot be fully considered, the control strategy is usually fixed, and there is a lack of adaptability to dynamic environmental changes, resulting in unsatisfactory clothes drying effects and efficiency.

[0007] To this end, the specific technical solution adopted by the present invention is as follows: According to one aspect of the present invention, a method for controlling an intelligent plastic clothes drying rack based on artificial intelligence is provided, and the method for controlling an intelligent plastic clothes drying rack based on artificial intelligence comprises the following steps: S1. Collect the original data of clothes drying, and use tensor decomposition technology to construct a multi-dimensional clothing feature map; S2, based on the multi-dimensional clothing feature map, using the ant colony algorithm to generate the clothes drying strategy; S3. According to the clothes drying strategy, use the multi-objective optimization algorithm to determine the specific control plan of the intelligent plastic clothes drying rack.

[0008] Furthermore, collecting the original data of clothes drying and using tensor decomposition technology to construct a multi-dimensional clothing feature map includes the following steps: S11. Based on sensors, combined with weather forecast data and historical drying records, original data of drying clothes is constructed, where the original data of drying clothes includes weight change data, local environmental parameters, weather forecast information, time information and historical drying records; S12, based on the original data of clothes drying, analyzing the key features of drying, and generating an initial multi-dimensional clothing feature tensor; S13. Utilize high-order singular value decomposition technology to compress the initial multi-dimensional clothing feature tensor and perform feature mapping to establish a multi-dimensional clothing feature map.

[0009] Furthermore, based on sensors, combined with weather forecast data and historical drying records, constructing the original data of drying clothes includes the following steps: S111, using sensors to obtain clothing weight change data and local environmental parameters; S112. Use Internet of Things technology to obtain real-time weather forecast information and time information; S113. Based on the search parameters, search the cloud database and extract relevant historical drying records to generate original data of the drying clothes. The search parameters include the initial weight of the current clothes and local environmental parameters.

[0010] Furthermore, based on the original data of clothes drying, analyzing the key features of clothes drying, and generating the initial multi-dimensional clothes feature tensor include the following steps: S121, standardize the original data of clothes drying to obtain a washing data set; S122, analyzing key drying features based on the cleaning data set, where the key drying features include clothing drying feature parameters, drying suitability index, and drying expectation parameters; S123. Based on the key features of drying clothes, an initial multi-dimensional clothing feature tensor is established using a tensor construction algorithm.

[0011] Furthermore, based on the cleaning data set, analyzing the key features of drying includes the following steps: S1221, calculating clothing drying characteristic parameters based on the washing data set, where the clothing drying characteristic parameters include initial moisture content and average drying rate; The expression for calculating the characteristic parameters of clothes drying is: ; In the formula, η is the initial moisture content, W 0 is the initial weight, W d To estimate dry weight; v is the average drying rate, W t For current clothes drying t Weight after hours, t For drying time; S1222. Calculate the drying suitability index according to environmental parameters and weather forecast data; S1223. Calculate expected drying parameters based on historical drying records. The expected drying parameters include expected drying time and user predicted satisfaction.

[0012] Furthermore, based on the key features of drying clothes, the tensor construction algorithm is used to establish the initial multi-dimensional clothing feature tensor, including the following steps: S1231. Establish a drying feature vector based on the drying key features; S1232. Generate a dynamic feature matrix using time series sampling technology; S1233. Based on the dynamic feature matrix and combined with the environmental dimension information, an initial multi-dimensional clothing feature tensor is constructed.

[0013] Furthermore, using high-order singular value decomposition technology, compressing the initial multi-dimensional clothing feature tensor and feature mapping, and establishing a multi-dimensional clothing feature map includes the following steps: S131, compressing the initial multi-dimensional clothing feature tensor to obtain a core tensor and a factor matrix; S132, generating a compressed feature map based on the core tensor and the factor matrix; S133. Construct a multi-dimensional clothing feature map based on the compressed feature map.

[0014] Furthermore, based on the multi-dimensional clothing feature map, the ant colony algorithm is used to generate a clothes drying strategy, which includes the following steps: S21. Construct a clothes drying strategy search space based on a multi-dimensional clothing feature map; S22. Based on the pheromone mechanism, the path selection is continuously optimized to explore the optimal combination of clothes drying strategies in the search space; S23. Utilize the real-time feedback information of each clothes-drying position to dynamically adjust the optimal clothes-drying strategy combination.

[0015] Furthermore, based on the pheromone mechanism, the path selection is continuously optimized and the optimal combination of clothes drying strategies is explored in the search space, including the following steps: S211, generating an initial clothes-drying strategy set based on the initial pheromone distribution and heuristic information; S212, using the strategy evaluation function to obtain the quality score of each clothes drying strategy; S213. Based on the strategy quality scoring results, update the pheromone distribution and generate a new set of optimized clothes-drying strategies.

[0016] According to another aspect of the present invention, there is also provided an intelligent plastic clothes drying rack control system based on artificial intelligence, the intelligent plastic clothes drying rack control system based on artificial intelligence comprising: The clothing feature construction module is used to collect the original data of clothes drying and construct a multi-dimensional clothing feature map using tensor decomposition technology; A clothes drying strategy generation module is used to generate clothes drying strategies based on a multi-dimensional clothing feature map using an ant colony algorithm; The control scheme optimization module is used to determine the specific control scheme of the intelligent plastic clothes drying rack based on the clothes drying strategy and using a multi-objective optimization algorithm.

[0017] The beneficial effects of the present invention are: (1) The present invention collects the original data of clothes drying and constructs a multi-dimensional clothes feature map using tensor decomposition technology. Based on the map, the ant colony algorithm is used to generate a clothes drying strategy. Finally, the specific control plan is determined according to the clothes drying strategy, thereby realizing intelligent management and optimization of the clothes drying process. The method makes full use of multi-source data, including real-time data collected by sensors, weather forecast information and historical drying records, to construct a multi-dimensional feature space that comprehensively reflects the clothes drying status. At the same time, through efficient data collection and processing, the feature tensor is compressed and mapped using high-order singular value decomposition technology, which not only reduces the data complexity but also retains key information, providing high-quality input for subsequent strategy generation. This data processing method based on multi-dimensional feature maps significantly improves the accuracy and adaptability of clothes drying strategies, and can formulate personalized clothes drying plans according to different clothing characteristics and environmental conditions, greatly improving clothes drying efficiency and user experience.

[0018] (2) A significant advantage of the present invention lies in the tensor decomposition technology and multi-dimensional clothing feature map construction method adopted. Traditional clothes drying control systems usually only consider a limited number of parameters, such as temperature and humidity, while ignoring the complex dynamic changes and multi-dimensional factors in the clothes drying process. The present invention constructs an initial multi-dimensional clothing feature tensor, integrates information of multiple dimensions such as time series, clothing characteristics, and environmental parameters into a unified mathematical model, and then uses high-order singular value decomposition technology to compress and feature map this high-dimensional tensor, which not only greatly reduces the complexity of data storage and calculation, but also extracts the most representative feature patterns. This data processing method can capture the subtle changes and potential laws in the clothes drying process, providing a rich and accurate information basis for the subsequent generation of clothes drying strategies. The construction of multi-dimensional clothing feature maps provides a new data representation and analysis framework for intelligent clothes drying, greatly improving the understanding and adaptability to complex clothes drying scenarios.

[0019] (3) The method for generating a clothes-drying strategy based on an ant colony algorithm adopted in the present invention is also a major innovation. Compared with traditional fixed rules or simple optimization algorithms, the ant colony algorithm has powerful global search capabilities and adaptability to dynamic environments. This method first constructs a search space suitable for the ant colony algorithm based on a multi-dimensional clothing feature map, transforming the complex clothes-drying problem into a path optimization problem. By intelligently processing and analyzing the data and designing a reasonable pheromone distribution and heuristic information, it is possible to quickly find a clothes-drying solution that is close to the optimal one in the huge strategy space. In addition, this method also introduces a real-time feedback mechanism that can dynamically adjust the pheromone distribution according to the actual status of each clothes-drying position, so that the generated clothes-drying strategy can adapt to environmental changes and the clothes drying process in a timely manner. This dynamic collaborative strategy generation method based on real-time data processing greatly improves the intelligence level of the clothes-drying process, which can not only optimize the drying efficiency, but also balance multiple goals such as energy consumption and clothes protection, providing users with an efficient, energy-saving and considerate smart clothes-drying experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 is a flow chart of a method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to an embodiment of the present invention; Figure 2 The present invention is a block diagram of an intelligent plastic clothes drying rack control system based on artificial intelligence. DETAILED DESCRIPTION

[0022] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0023] According to an embodiment of the present invention, a method and system for controlling an intelligent plastic clothes drying rack based on artificial intelligence are provided.

[0024] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1As shown, according to one embodiment of the present invention, a control method of an intelligent plastic clothes drying rack based on artificial intelligence is provided, and the control method of an intelligent plastic clothes drying rack based on artificial intelligence comprises the following steps: S1. Collect the original data of clothes drying, and use tensor decomposition technology to construct a multi-dimensional clothing feature map; S2, based on the multi-dimensional clothing feature map, using the ant colony algorithm to generate the clothes drying strategy; S3. According to the clothes drying strategy, use the multi-objective optimization algorithm to determine the specific control plan of the intelligent plastic clothes drying rack.

[0025] Specifically, considering the characteristics and requirements of the intelligent plastic clothes drying rack control method, a suitable multi-objective optimization algorithm should be able to effectively handle multiple objectives and have good convergence and computational efficiency. In this embodiment, the NSGA-II algorithm (Non-dominated Sorting Genetic Algorithm II) is used as the multi-objective optimization algorithm; the NSGA-II algorithm is specially designed to handle multi-objective optimization problems, and can simultaneously consider multiple objectives that may conflict with each other (including drying efficiency, energy consumption, and clothing protection). The NSGA-II algorithm uses a fast non-dominated sorting method and can efficiently find the Pareto optimal solution set. Using NSGA-II, multiple objectives such as clothes drying efficiency, energy consumption, and clothing protection can be effectively balanced to generate the optimal control solution for the intelligent plastic clothes drying rack. This method not only takes into account the complexity of the problem, but also ensures the practicality and efficiency of the algorithm, and is very suitable for this kind of intelligent home appliance control application based on artificial intelligence.

[0026] Specifically, in this embodiment, ① the clothes drying strategy generated by S2 is used as input, including the optimal temperature of 26°C, the optimal humidity of 55%, the drying suitability index of 0.82, and the expected drying time of 3.5 hours; ② the decision variables are defined, including (a) the height of the clothes drying rack is in the range of 1.5m to 2.0m; (b) the angle of the clothes drying rack is in the range of 0° to 30°; (c) the distance between the clothes drying racks is in the range of 20cm to 40cm; ③ the optimization goals are set, including (a) maximizing the drying efficiency f1=1 / drying time; (b) minimizing the energy consumption f2=fan energy consumption+height adjustment energy consumption; (c) maximizing the clothing protection f3=1-|current temperature-optimal temperature| / 10; ④ the constraints are defined, including (a) 1.5m≤height of clothes drying rack≤2.0m; (b) 0°≤angle of clothes drying rack≤30°; (c) 20cm≤distance between clothes drying racks≤40cm; (d) 0≤auxiliary fan speed≤3; Specifically, in the above embodiment, ⑤ configure the NSGA-II algorithm parameters, including a population size of 100, an iteration number of 50, a crossover probability of 0.9, and a mutation probability of 0.1; ⑥ execute the NSGA-II algorithm, including (a) initializing the population and randomly generating 100 solutions that meet the constraints; (b) evaluating the fitness of each solution (the value of the three objective functions); (c) performing non-dominated sorting and crowding calculations; (d) generating a new population through selection, crossover, and mutation operations; (e) repeating steps bd until 50 iterations are reached; ⑦ selecting the final solution from Par A balanced solution is selected from the ETO optimal solution set as the final control solution, and the selection criterion is the solution with the largest weighted sum of the three objective functions;⑧ Output control solution. In the above embodiment, the final solution selected is: clothes rack height 1.8m; clothes rack angle 15°; clothes rack spacing 30cm;⑨ Converted into specific control instructions, including (a) sending instructions to the height adjustment motor to raise the clothes rack to 1.8m; (b) sending instructions to the angle adjustment mechanism to tilt the clothes rack 15°; (c) sending instructions to the spacing adjustment mechanism to set the clothes rack spacing to 30cm.

[0027] Specifically, through the above process, using the NSGA-II algorithm, based on the clothes-drying strategy generated by S2, considering multiple objectives (drying efficiency, energy consumption and clothes protection), and under the condition of satisfying physical constraints, a specific control scheme for the smart plastic clothes drying rack was obtained. This scheme balances multiple objectives and provides optimized control parameters for the clothes-drying process.

[0028] In one embodiment, collecting the original data of clothes drying and constructing a multi-dimensional clothing feature map using tensor decomposition technology includes the following steps: S11, constructing original data of clothes drying based on sensors, in combination with weather forecast data and historical drying records, wherein the original data of clothes drying includes weight change data, local environmental parameters, weather forecast information, time information and historical drying records; S12, based on the original data of clothes drying, analyzing the key features of drying, and generating an initial multi-dimensional clothing feature tensor; S13. Utilize high-order singular value decomposition technology to compress the initial multi-dimensional clothing feature tensor and perform feature mapping to establish a multi-dimensional clothing feature map.

[0029] In one embodiment, based on sensors, combined with weather forecast data and historical drying records, constructing the original data of drying clothes includes the following steps: S111, using sensors to obtain clothing weight change data and local environmental parameters; Specifically, based on the built-in sensor of the smart plastic clothes drying rack, the weight change data of the clothes and the local environmental parameters are obtained. The local environmental parameters include the temperature data and humidity data of the surrounding environment. The weight sensor and the environmental sensor are used to collect the weight of the clothes and the temperature and humidity data of the surrounding environment in real time.

[0030] Specifically, in this embodiment, when the user hangs a piece of wet clothes on the smart plastic clothes drying rack, the weight sensor immediately records the initial weight as 500 grams and is set to record the weight data every 15 minutes; at the same time, the environmental sensor continuously monitors the local environment, recording the current temperature as 26°C and the relative humidity as 55%; in the next 3 hours, a total of 12 sets of weight data and 36 sets of temperature and humidity data were recorded.

[0031] S112. Use Internet of Things technology to obtain real-time weather forecast information and time information; Specifically, the Internet of Things technology is used to obtain real-time weather forecast information and time information; and through the network connection function of the smart plastic clothes drying rack, short-term weather forecasts are obtained from the meteorological service API and the system time is synchronized.

[0032] Specifically, in this embodiment, the smart plastic clothes drying rack is connected to the Internet via Wi-Fi, accesses the local meteorological service API, and obtains the weather forecast for the next 6 hours, which shows that the weather is sunny, the temperature will fluctuate between 25-28°C, the humidity is expected to remain in the range of 50-60%, and there is no probability of precipitation; at the same time, the start time of drying is recorded as 14:30, and a timer is set to track the duration of drying.

[0033] S113, based on the search parameters, searching the cloud database, extracting relevant historical drying records, and generating original data of the drying clothes, wherein the search parameters include the initial weight of the current clothes and local environmental parameters.

[0034] Specifically, based on the cloud database, relevant historical drying records are retrieved and extracted; using the Internet of Things platform, the user's historical drying database is accessed to screen out drying records similar to the current clothing characteristics and environmental conditions, obtain relevant historical drying records, and combine weight change data, local environmental parameters, weather forecast information and time information to establish the original data of drying clothes.

[0035] Specifically, in the above embodiment, the initial weight of the current clothes (500 grams) and the environmental conditions (26°C, 55% humidity) are used as search parameters to search for drying records under similar conditions in the past month from the cloud database; finally, 5 relevant historical drying records are successfully retrieved, including the average drying time of these clothes (3.5 hours), the final dry weight (about 320 grams) and the drying effect score (average 4.2 / 5).

[0036] In one embodiment, based on the original data of clothes drying, analyzing the key features of clothes drying and generating an initial multi-dimensional clothes feature tensor comprises the following steps: S121, standardize the original data of clothes drying to obtain a washing data set; Specifically, the original data of the clothes drying obtained in step S11 is processed using a data cleaning algorithm and a standardization method to eliminate abnormal values ​​and unify the data format.

[0037] Specifically, in this embodiment, outlier detection is performed on the weight change data to eliminate instantaneous abnormal readings caused by the shaking of the hanger; then normalization (maximum and minimum normalization is used in this embodiment) is used to convert the temperature data from degrees Celsius to a standardized 0-1 range, 26°C is converted to 0.6, and the drying effect scores in the historical drying records are also normalized, converting 4.2 / 5 to 0.84; finally, a cleaned data set containing time series weight data, standardized environmental parameters, and normalized historical data is obtained.

[0038] S122, analyzing key drying features based on the cleaning data set, where the key drying features include clothing drying feature parameters, drying suitability index, and drying expectation parameters; S123. Based on the key features of drying clothes, an initial multi-dimensional clothing feature tensor is established using a tensor construction algorithm.

[0039] In one embodiment, based on the cleaning data set, analyzing the key characteristics of drying includes the following steps: S1221, calculating clothing drying characteristic parameters based on the washing data set, wherein the clothing drying characteristic parameters include initial moisture content and average drying rate; The expression for calculating the clothes drying characteristic parameter is: ; In the formula, η is the initial moisture content, W 0 is the initial weight, W d To estimate dry weight; v is the average drying rate, W t For current clothes drying t Weight after hours, t For drying time; S1222. Calculate the drying suitability index according to environmental parameters and weather forecast data; Specifically, the expression for calculating the drying suitability index is: ; In the formula, S is the drying suitability index,f ( T ), g ( H ), h ( W ) are the scoring functions for temperature, humidity and weather conditions, respectively. w T , w H , w W are the corresponding weights respectively, and w T + w H + w W =1.

[0040] Specifically, the temperature scoring function f ( T )=1-| T - T op | / T rg , where T is the current temperature, T op is the optimal drying temperature (set to 25°C in this embodiment), T rg is the acceptable temperature range (set to 20°C in this embodiment); Specifically, the humidity score function g ( H )=1- H / 100, where H is the current relative humidity percentage; Specifically, the weather condition scoring function h ( W ) is expressed as follows: .

[0041] Specifically, in the above embodiment, the current temperature is 26°C, the relative humidity is 55%, and the weather forecast is sunny, then (a) the weight distribution w T =0.4, w H =0.3, w W =0.3; (b) Calculation process: f (26)=1-|26-25| / 20=0.95; g (55)=1-55 / 100=0.45; h (“Sunny”) = 1.0; Finally, the drying suitability index is obtained. S =0.4*0.95+0.3*0.45+0.3*1.0=0.815.

[0042] S1223. Calculate expected drying parameters based on historical drying records, where the expected drying parameters include expected drying time and user predicted satisfaction.

[0043] Specifically, the expression for calculating the expected parameters of drying is: ; In the formula, T ep For the expected drying time, T i is the drying time in the historical drying records, w i is the similarity weight, S pr Predicting user satisfaction, S i Score user satisfaction in historical drying records.

[0044] Specifically, the similarity weight is calculated using temperature and humidity as comparison factors, and the calculation expression is: w i =1-(| T i - T c | / 10+| H i - H c | / 20) / 2, where T i and H i It is a historical record i temperature and humidity, T c and H c are the current temperature and humidity, 10 and 20 are normalization factors. In this embodiment, the temperature difference is set within ±10°C, and the humidity difference is set within ±20%.

[0045] Specifically, in the above embodiment, the current conditions are temperature 25°C and humidity 60%, and 3 similar historical records are found from the historical drying records: Record 1—Temperature 24°C, humidity 58%, drying time 3.2 hours, satisfaction 4.5; Record 2—Temperature 26°C, humidity 62%, drying time 3.5 hours, satisfaction 4.2; Record 3—Temperature 23°C, humidity 65%, drying time 3.8 hours, satisfaction 4.0; Then, the similarity weights corresponding to the three similar historical records are calculated respectively, and we get w 1=0.95, w 2=0.95, w 3=0.875; Finally, the expected drying time T ep =(3.2*0.8+3.5*0.9+3.8*0.7) / (0.8+0.9+0.7)=3.49 hours; User Prediction of Satisfaction S pr =(4.5*0.8+4.2*0.9+4.0*0.7) / (0.8+0.9+0.7)=4.25.

[0046] In one embodiment, based on the key features of drying clothes, using a tensor construction algorithm, establishing an initial multi-dimensional clothing feature tensor includes the following steps: S1231. Establish a drying feature vector based on the drying key features; Specifically, the clothing drying characteristic parameters, the drying suitability index, and the drying expectation parameters calculated in step S122 are used to generate a characteristic vector representing the current drying state.

[0047] Specifically, in the above embodiment, the result obtained in step S122 is: Initial moisture content η =36%; Average drying rate v =9% / hour; Drying suitability index S =0.815; Expected drying time T ep =3.49 hours; User predicted satisfaction Spr=4.25; Then construct the feature vector F =[ η , v , S , T ep , S pr ]=[0.36,0.09,0.815,3.49,4.25].

[0048] S1232. Generate a dynamic feature matrix using time series sampling technology; Specifically, based on the drying feature vector and a preset time interval, the drying process is sampled multiple times to generate a feature matrix reflecting the dynamic changes of the drying process.

[0049] Specifically, in this embodiment, the preset sampling interval is 30 minutes, the total number of sampling times is 8 times (covering 4 hours), and then at each time point t, the dynamic change value (moisture content, drying rate) in the feature vector is updated, and other features are kept unchanged. The generated dynamic feature matrix M is: M=[ [0.36,0.09,0.815,3.49,4.25] / / t=0 [0.31,0.08,0.815,3.49,4.25] / / t=30min [0.27,0.07,0.815,3.49,4.25] / / t=60min ... [0.05,0.02,0.815,3.49,4.25] / / t=210min ].

[0050] S1233. Based on the dynamic feature matrix and combined with the environmental dimension information, an initial multi-dimensional clothing feature tensor is constructed.

[0051] Specifically, a third-order tensor containing time, feature, and environment dimensions is constructed using the dynamic feature matrix and additional environment information.

[0052] Specifically, the environmental dimension (temperature, humidity) is added, and the dynamic feature matrix M is expanded to a third-order tensor T; in the above embodiment, the temperature and humidity recorded at each time point are [26, 25, 25, 24, 24, 23, 23, 22] and [55, 54, 53, 52, 51, 50, 49, 48] respectively; The initial multi-dimensional clothing feature tensor T generated is: T1=[M,[26,25,25,24,24,23,23,22]^T]; T2=[M,[55,54,53,52,51,50,49,48]^T]; Where, the dimension of T is 8×7×2, which means 8 time points, 7 features (5 original features plus temperature and humidity), and 2 environmental dimensions (temperature and humidity).

[0053] Specifically, the key features of drying are first organized into feature vectors, and then based on the time series, the feature vectors are expanded into dynamic feature matrices. Finally, the dynamic feature matrix is ​​further expanded into a multidimensional feature tensor to comprehensively capture the dynamic changes and environmental impacts of the drying process. Through the above process, the complex drying data is successfully converted into a structured, information-rich multidimensional feature tensor, which provides an ideal input for subsequent high-order singular value decomposition and feature map generation.

[0054] In one embodiment, using high-order singular value decomposition technology, compressing the initial multi-dimensional clothing feature tensor and feature mapping, and establishing a multi-dimensional clothing feature map includes the following steps: S131, compressing the initial multi-dimensional clothing feature tensor to obtain a core tensor and a factor matrix; Specifically, the initial multi-dimensional clothing feature tensor is compressed, and a compressed core tensor and factor matrix are obtained by using a high-order singular value decomposition technique.

[0055] Specifically, step S1233 obtains an initial multidimensional clothing feature tensor T with a dimension of 8×7×2, and uses the Tucker decomposition algorithm to decompose T, and the steps include (a) performing singular value decomposition on each mode of the tensor T; (b) selecting the main singular vectors of each mode to form factor matrices A, B, and C; (c) obtaining a core tensor G by projecting the original tensor T into the factor matrix space; wherein G represents the core tensor (with a dimension of 4×3×2 in this embodiment), and A (with a dimension of 8×4), B (with a dimension of 7×3), and C (with a dimension of 2×2) represent the factor matrices of the time dimension, feature dimension, and environment dimension, respectively.

[0056] S132, generating a compressed feature map based on the core tensor and the factor matrix; Specifically, the core tensor and factor matrix are used to generate a compressed feature map; based on the compressed core tensor and factor matrix, a low-dimensional feature space that can effectively represent the original data is constructed.

[0057] Specifically, in the above embodiment, the core tensor G (4×3×2) is used as the main feature representation after compression, the factor matrix A (8×4) is used to represent the main mode of the time dimension, the factor matrix B (7×3) is used to represent the main mode of the feature dimension, and the factor matrix C (2×2) is used to represent the main mode of the environment dimension, thereby obtaining a compressed feature map F, expressed as F={G,A,B,C}. This mapping allows the original high-dimensional data to be represented and operated in a low-dimensional space, greatly reducing the computational complexity.

[0058] S133. Construct a multi-dimensional clothing feature map based on the compressed feature map.

[0059] Specifically, based on the compressed feature mapping, a multi-dimensional clothing feature map is constructed; using the compressed feature representation, an intuitive, multi-dimensional clothing drying status visualization map is generated.

[0060] Specifically, the slices of the core tensor G are used as the main body of the graph to represent the drying status under different time and feature combinations; first, the first slice G1 of the core tensor G is used as the basic plane; the column vectors of the factor matrix A are used as the time axis (x-axis) to represent the main time pattern of the drying process; the column vectors of the factor matrix B are used as the feature axis (y-axis) to represent the main pattern of the key drying features; on the xy plane, the value of G1 is displayed in the form of a heat map; the same process is repeated for G2 to generate a second heat map; finally, the two heat maps are superimposed and represented by different color channels to generate the final multi-dimensional clothing feature map.

[0061] In one embodiment, generating a clothes drying strategy using an ant colony algorithm based on a multi-dimensional clothing feature map includes the following steps: S21. Construct a clothes drying strategy search space based on a multi-dimensional clothing feature map; Specifically, based on the multi-dimensional clothing feature map constructed in step S13, a search space of an ant colony algorithm is established, clothing features and environmental parameters are mapped to "paths" and "nodes" in the ant colony algorithm, and a clothes drying strategy search space suitable for ant colony optimization is constructed.

[0062] Specifically, the main dimensions in the multidimensional clothing feature map are used as the search dimensions of the ant colony algorithm, each dimension is discretized into several levels, and then the initial pheromone distribution is defined, and the initial value is set based on the historical clothes drying data.

[0063] Specifically, in this embodiment, based on the multidimensional clothing feature map constructed in step S13, the following dimensions are selected as the search space of the ant colony algorithm: ① temperature, ranging from 20-30°C, discretized into 10 levels, 1°C per level; ② humidity, ranging from 40-80%, discretized into 8 levels, 5% per level; ③ drying suitability index, ranging from 0-1, discretized into 10 levels, 0.1 per level; ④ expected drying time, ranging from 1-6 hours, discretized into 10 levels, 0.5 hours per level.

[0064] Specifically, in the above embodiment, the initial pheromone distribution is set based on the historical drying records calculated in step S1223, wherein a higher initial pheromone value is set for a path with a temperature of 25°C, a humidity of 60%, a drying suitability index of 0.815, and an expected drying time of 3.49 hours, and ultimately the initial pheromone value of the path is set to 1.5; the initial pheromone values ​​of other paths are set to 1.0.

[0065] S22. Based on the pheromone mechanism, the path selection is continuously optimized to explore the optimal combination of clothes drying strategies in the search space; S23. Utilize the real-time feedback information of each clothes-drying position to dynamically adjust the optimal clothes-drying strategy combination.

[0066] Specifically, based on the optimization results of the ant colony algorithm and combined with the real-time feedback mechanism, a dynamic collaborative clothes-drying strategy is generated; by utilizing the real-time feedback information of each clothes-drying position, the results of the ant colony algorithm are dynamically adjusted and optimized.

[0067] Specifically, ① set a feedback cycle. In this embodiment, the status information of each clothes drying position is collected every 15 minutes; ② define feedback evaluation indicators, including actual drying rate and energy consumption. In this embodiment, the actual drying rate is set as the percentage of weight reduction every 15 minutes, and the energy consumption is set as the power consumption every 15 minutes; ③ dynamically adjust the pheromone distribution according to the feedback information. The specific adjustment rules include a. if the drying effect of a certain clothes drying position is good, increase the pheromone concentration of the position; b. if the effect is not good, reduce the pheromone concentration; ④ based on the adjusted pheromone distribution, generate a new optimization strategy. In the above embodiment, the original strategy is temperature 26°C, humidity 55%, drying suitability index 0.815, and expected drying time 3.5 hours; the adjusted strategy is temperature 27°C, humidity 53%, drying suitability index 0.83, and expected drying time 3.3 hours.

[0068] In one embodiment, based on the pheromone mechanism, the path selection is continuously optimized and the optimal combination of clothes drying strategies is explored in the search space, including the following steps: S221, generating an initial clothes-drying strategy set based on the initial pheromone distribution and heuristic information; Specifically, 20 initial clothes drying strategies are generated using the initial pheromone distribution and heuristic information set in step S21, including: (a) Pheromone distribution: the initial pheromone value of the strategy with a temperature of 25°C, a humidity of 60%, a drying suitability index of 0.815, and an expected drying time of 3.49 hours is 1.5, and the initial pheromone values ​​of other strategies are 1.0; (b) Heuristic information, η(i,j)=1 / |current value-optimal value|, where i represents the decision variable and j represents the specific value of the variable; In this embodiment, for temperature (i=1), η(1,25)=1 / |25-25|=1 (25°C is set as the optimal temperature), η(1,26)=1 / |26-25|=1; For humidity (i=2), η(2,60)=1 / |60-55|=0.2 (55% is set as the optimal humidity), η(2,55)=1 / |55-55|=1; (c) Strategy selection probability: P(i,j)=[τ(i,j)]^α*[η(i,j)]^β / Σ[τ(i,k)]^α*[η(i,k)]^β, where α=1, β=2, τ(i,j) is the pheromone value, η(i,j) is the heuristic information value, and k represents all possible values ​​of the decision variable i; (d) Use the roulette wheel method to select 20 initial drying strategies, each strategy contains parameters such as temperature, humidity, drying suitability index and expected drying time; In this embodiment, strategy 1—temperature 26°C, humidity 55%, drying suitability index 0.82, expected drying time 3.5 hours; Strategy 2—Temperature 25°C, humidity 58%, drying suitability index 0.81, expected drying time 3.7 hours... (20 strategies in total).

[0069] S222, using the strategy evaluation function to obtain the quality score of each clothes drying strategy; Specifically, based on the drying suitability index formula in step S1222, the quality score of each drying strategy is calculated. ① Strategy evaluation function Q =w1× S +w2×(1 / T ep )+w3× E , where S is the drying suitability index, T ep For the expected drying time, E is the energy efficiency index; ② weight setting, w1=0.5, w2=0.3, w3=0.2; ③ energy efficiency index: E =1-(current energy consumption / maximum energy consumption); In this embodiment, the clothes drying strategy is temperature 26°C, humidity 55%, drying suitability index 0.815, expected drying time 3.5 hours, energy efficiency index 0.9, and finally the strategy evaluation function is solved to obtain the strategy quality score: Q =0.5×0.815+0.3×(1 / 3.5)+0.2×0.9=0.6375.

[0070] S223. Based on the strategy quality scoring results, update the pheromone distribution and generate a new set of optimized clothes-drying strategies.

[0071] Specifically, the strategy quality scoring results are used to update the pheromone distribution and generate the optimized clothes-drying strategy set for the next iteration, including ① the pheromone update formula τ(i,j)=(1-ρ)τ(i,j)+Δτ(i,j), where ρ is the pheromone volatility coefficient (set to 0.1) and Δτ(i,j) is the pheromone increment; ② the pheromone increment calculation formula Δτ(i,j)=Q / L , where Q Score the quality of the strategy, L is the number of strategy parameters (4 in this embodiment).

[0072] Specifically, in this embodiment, the above formula is used to calculate: original pheromone value - τ (26 ° C, 55%) = 1.0; strategy quality score Q =0.6375; new pheromone value τ(26°C,55%)=(1-0.1)*1.0+0.6375 / 4=1.059375; thus, based on the updated pheromone distribution, 20 new optimized clothes drying strategies are generated using the same method as step S221.

[0073] like Figure 2 As shown, according to another embodiment of the present invention, there is also provided an intelligent plastic clothes drying rack control system based on artificial intelligence, and the intelligent plastic clothes drying rack control system based on artificial intelligence includes: Clothing feature construction module 1, used to collect original data of clothes drying, and construct a multi-dimensional clothing feature map using tensor decomposition technology; A clothes drying strategy generation module 2 is used to generate clothes drying strategies based on a multi-dimensional clothing feature map using an ant colony algorithm; The control scheme optimization module 3 is used to determine the specific control scheme of the intelligent plastic clothes drying rack according to the clothes drying strategy using a multi-objective optimization algorithm.

[0074] 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 principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent plastic clothes drying rack control method based on artificial intelligence, characterized in that: The artificial intelligence-based intelligent plastic clothes drying rack control method comprises the following steps: S1. Collect the original data of clothes drying, and use tensor decomposition technology to construct a multi-dimensional clothing feature map; S2, based on the multi-dimensional clothing feature map, using the ant colony algorithm to generate the clothes drying strategy; S3. According to the clothes drying strategy, use the multi-objective optimization algorithm to determine the specific control plan of the intelligent plastic clothes drying rack.

2. The method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to claim 1, characterized in that: The collecting of original data of clothes drying and the construction of a multi-dimensional clothes feature map using tensor decomposition technology include the following steps: S11, constructing original data of clothes drying based on sensors, in combination with weather forecast data and historical drying records, wherein the original data of clothes drying includes weight change data, local environmental parameters, weather forecast information, time information and historical drying records; S12, based on the original data of clothes drying, analyzing the key features of drying, and generating an initial multi-dimensional clothing feature tensor; S13. Utilize high-order singular value decomposition technology to compress the initial multi-dimensional clothing feature tensor and perform feature mapping to establish a multi-dimensional clothing feature map.

3. The method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to claim 2, characterized in that: The method of constructing the original data of clothes drying based on the sensor, in combination with the weather forecast data and the historical drying records, comprises the following steps: S111, using sensors to obtain clothing weight change data and local environmental parameters; S112. Use Internet of Things technology to obtain real-time weather forecast information and time information; S113, based on the search parameters, searching the cloud database, extracting relevant historical drying records, and generating original data of the drying clothes, wherein the search parameters include the initial weight of the current clothes and local environmental parameters.

4. The method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to claim 2, characterized in that: The method of analyzing the key features of drying clothes based on the original data of drying clothes and generating an initial multi-dimensional clothing feature tensor comprises the following steps: S121, standardize the original data of clothes drying to obtain a washing data set; S122, analyzing key drying features based on the cleaning data set, where the key drying features include clothing drying feature parameters, drying suitability index, and drying expectation parameters; S123. Based on the key features of drying clothes, an initial multi-dimensional clothing feature tensor is established using a tensor construction algorithm.

5. The method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to claim 4, characterized in that: The analysis of the key features of drying based on the cleaning data set includes the following steps: S1221, calculating clothing drying characteristic parameters based on the washing data set, wherein the clothing drying characteristic parameters include initial moisture content and average drying rate; The expression for calculating the clothes drying characteristic parameter is: ； In the formula, η is the initial moisture content, W 0 is the initial weight, W d To estimate dry weight; v is the average drying rate, W t For current clothes drying t Weight after hours, t For drying time; S1222. Calculate the drying suitability index according to environmental parameters and weather forecast data; S1223. Calculate expected drying parameters based on historical drying records, where the expected drying parameters include expected drying time and user predicted satisfaction.

6. The method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to claim 4, characterized in that: The method of establishing an initial multi-dimensional clothing feature tensor based on the key features of drying clothes by using a tensor construction algorithm comprises the following steps: S1231. Establish a drying feature vector based on the drying key features; S1232. Generate a dynamic feature matrix using time series sampling technology; S1233. Based on the dynamic feature matrix and combined with the environmental dimension information, an initial multi-dimensional clothing feature tensor is constructed.

7. The method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to claim 2, characterized in that: The method of compressing the initial multi-dimensional clothing feature tensor and mapping the feature by using the high-order singular value decomposition technology to establish the multi-dimensional clothing feature map includes the following steps: S131, compressing the initial multi-dimensional clothing feature tensor to obtain a core tensor and a factor matrix; S132, generating a compressed feature map based on the core tensor and the factor matrix; S133. Construct a multi-dimensional clothing feature map based on the compressed feature map.

8. The method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to claim 1, characterized in that: The method of generating a clothes drying strategy based on a multi-dimensional clothing feature map using an ant colony algorithm comprises the following steps: S21. Construct a clothes drying strategy search space based on a multi-dimensional clothing feature map; S22. Based on the pheromone mechanism, the path selection is continuously optimized to explore the optimal combination of clothes drying strategies in the search space; S23. Utilize the real-time feedback information of each clothes-drying position to dynamically adjust the optimal clothes-drying strategy combination.

9. The method for controlling an intelligent plastic clothes drying rack based on artificial intelligence according to claim 8, characterized in that: The method of continuously optimizing path selection based on the pheromone mechanism and exploring the optimal combination of clothes drying strategies in the search space includes the following steps: S211, generating an initial clothes-drying strategy set based on the initial pheromone distribution and heuristic information; S212, using the strategy evaluation function to obtain the quality score of each clothes drying strategy; S213. Based on the strategy quality scoring results, update the pheromone distribution and generate a new set of optimized clothes-drying strategies.

10. An artificial intelligence-based intelligent plastic clothes drying rack control system, used to implement the artificial intelligence-based intelligent plastic clothes drying rack control method according to any one of claims 1 to 9, characterized in that: The artificial intelligence-based intelligent plastic clothes drying rack control system includes: The clothing feature construction module is used to collect the original data of clothes drying and construct a multi-dimensional clothing feature map using tensor decomposition technology; A clothes drying strategy generation module is used to generate clothes drying strategies based on a multi-dimensional clothing feature map using an ant colony algorithm; The control scheme optimization module is used to determine the specific control scheme of the intelligent plastic clothes drying rack based on the clothes drying strategy and using a multi-objective optimization algorithm.

Images