In-vehicle air conditioner regulation control method, device and equipment and storage medium

Abstract

The invention provides an in-vehicle air conditioner adjusting control method, device and equipment and a storage medium, and the method comprises the steps that in the vehicle running process, in-vehicle environment parameters and in-vehicle environment parameters are collected in real time, and an environment parameter set is generated; and calculating a human body heat balance model of the passenger according to the parameter set, and generating a multi-region thermal comfort index set. And based on the environment parameter set and the thermal comfort index set, the operation state of the air conditioner system is evaluated, and an air conditioner state evaluation result is generated. And multi-strategy parallel reasoning is carried out through a preset deep reinforcement learning model, and an optimal control strategy set of the air conditioning system is generated. And according to the strategy set, distributed cooperative control is conducted on an air conditioner executing mechanism, the adjusting effect is fed back in real time, a dynamic adjusting parameter set is generated, and fine adjustment is conducted on the control strategy. According to the method, through multi-dimensional environment parameter collection and model reasoning, the air conditioner control strategy is dynamically adjusted according to the real-time environment and the thermal comfort degree, more efficient air conditioner system adjustment is achieved, and the comfort of passengers is improved.

Description

Technical Field

[0001] The present invention relates to the field of air conditioning regulation, and in particular to a method, device, equipment and storage medium for controlling in-vehicle air conditioning regulation. Background Art

[0002] With the continuous development of automobile intelligence and comfort technology, the in-car air conditioning system has become an important factor in improving the riding experience. However, most traditional in-car air conditioning adjustment systems use simple temperature settings and air volume adjustment methods, which cannot be adjusted intelligently according to changes in the in-car and out-of-car environment and the personalized needs of passengers. This fixed adjustment method is often difficult to cope with the dynamic changes in the in-car and out-of-car environment under different working conditions (such as high-speed driving, congested driving, etc.), making it difficult for the air conditioning system to maintain passenger comfort in various complex situations.

[0003] Existing air-conditioning systems mostly control the temperature inside the vehicle based on a preset temperature range, and do not fully consider the thermal comfort and individual differences of the passengers. They lack coordinated control with multi-dimensional factors such as the vehicle's operating status, the external environment, and the passenger's body thermal balance, resulting in high energy consumption of the air-conditioning system, low adjustment efficiency, and unsatisfactory comfort. Summary of the invention

[0004] The main purpose of the present invention is to solve the technical problems that the existing in-vehicle air conditioning adjustment method lacks coordinated control of multi-dimensional factors, resulting in high energy consumption of the air conditioning system, low adjustment efficiency, and unsatisfactory comfort; A first aspect of the present invention provides a method for adjusting and controlling an in-vehicle air conditioner, the method comprising: During the operation of the target vehicle, multi-dimensional real-time collection and processing of the environmental parameters inside and outside the vehicle are performed to obtain the environmental parameter set of the target vehicle; Performing multi-region thermal comfort calculation on a human body thermal balance model of passengers in the target vehicle according to the environmental parameter set to obtain a thermal comfort index set for the passengers; Performing a multi-parameter evaluation on the operating state of the air conditioning system in the target vehicle according to the environmental parameter set and the thermal comfort index set to obtain an air conditioning state evaluation result; Performing multi-strategy parallel reasoning based on the air-conditioning state evaluation result through a preset deep reinforcement learning model to obtain an optimal control strategy set for the air-conditioning system; According to the optimal control strategy set, distributed collaborative control is performed on the actuators of the air-conditioning system, and real-time feedback evaluation is performed on the control effect to obtain a dynamic adjustment parameter set. Through the dynamic adjustment parameter set, the optimal control strategy set is fine-tuned in real time.

[0005] Optionally, in a first implementation of the first aspect of the present invention, performing multi-region thermal comfort calculation on a human thermal balance model of a passenger in the target vehicle according to the environmental parameter set to obtain a thermal comfort index set for the passenger includes: Performing spatial interpolation processing on the temperature data, humidity data and airflow velocity data in the environmental parameter set to obtain local environmental parameters of a plurality of predefined human body regions of the passenger; Calibrate the parameters of the preset first human body thermal balance model according to the local environmental parameters and the physiological characteristic parameters of the occupant in the environmental parameter set to obtain a personalized thermal balance equation group; Numerically solving the thermal balance equations to obtain the passenger's human thermal state parameters, and performing multi-region parallel calculations on a preset second human thermal balance model according to the human thermal state parameters to obtain local thermal sensation values ​​of multiple predefined human regions; The local thermal sensation values ​​are subjected to weighted averaging and nonlinear mapping processing to obtain a thermal comfort index of the passenger.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the first human body thermal balance model is a Fanger PMV model; The parameter calibration of the preset first human body thermal balance model according to the local environmental parameters and the physiological characteristic parameters of the occupant in the environmental parameter set to obtain a personalized thermal balance equation group includes: Performing a multi-factor correlation analysis on the local environmental parameters and the physiological characteristic parameters of the passengers to obtain a weight matrix of thermal balance influencing factors; Dynamically correcting the metabolic rate and clothing thermal resistance parameters in the Fanger PMV model according to the weight matrix of the thermal balance influencing factors to obtain personalized human body heat generation and heat dissipation parameters, and substituting the personalized human body heat generation and heat dissipation parameters into the basic equation of the Fanger PMV model to obtain a corrected local PMV calculation equation; Performing regional division and parameter configuration on the modified local PMV calculation equation to obtain local heat balance equations for multiple predefined areas in the vehicle; The local heat balance equations of the plurality of predefined regions are combined and coupled to obtain a personalized heat balance equation group.

[0007] Optionally, in a third implementation of the first aspect of the present invention, performing a multi-parameter evaluation on the operating state of the air-conditioning system in the target vehicle according to the environmental parameter set and the thermal comfort index set to obtain an air-conditioning state evaluation result includes: Performing time series analysis and trend prediction on the environmental parameter set and thermal comfort index set to obtain in-vehicle thermal environment change trend data; Performing a dynamic response analysis on the operating parameters of each component of the air-conditioning system in the target vehicle according to the in-vehicle thermal environment change trend data to obtain an air-conditioning system state parameter set; Calculating the deviation between the state parameter set of the air conditioning system and the preset optimal operating parameter range to obtain the performance efficiency index of each component in the air conditioning system; According to the performance efficiency index and the thermal comfort index set, a multi-objective optimization analysis is performed on the air conditioning system to obtain a comprehensive performance evaluation index; The comprehensive performance evaluation index and the air conditioning system state parameter set are subjected to weighted fusion processing and quantitative grading processing to obtain an air conditioning state evaluation result.

[0008] Optionally, in a fourth implementation of the first aspect of the present invention, performing multi-strategy parallel reasoning according to the air-conditioning state evaluation result through a preset deep reinforcement learning model to obtain an optimal control strategy set for the air-conditioning system includes: Performing feature extraction and normalization processing on the air conditioning state evaluation result to obtain a state vector suitable for a deep reinforcement learning model; Performing forward propagation calculation on a policy network of a preset deep reinforcement learning model according to the state vector to obtain multiple candidate control actions; Performing parallel simulation and evaluation on the multiple candidate control actions to obtain an expected reward value and a state transition probability corresponding to each action; The candidate control actions are sorted and screened according to the expected reward value and the state transition probability to obtain an optimal control strategy set, wherein the optimal control strategy set includes the compressor speed adjustment amount, the blower air volume adjustment amount of each area, and the air outlet mode selection.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, the parallel simulation and evaluation of the multiple candidate control actions to obtain the expected reward value and state transition probability corresponding to each action includes: Combining the multiple candidate control actions and the state vector to obtain multiple groups of state-action pairs; Solving the preset in-vehicle thermal environment dynamics model in parallel according to the multiple groups of state-action pairs to obtain a predicted state set at the next moment; Performing thermal comfort evaluation and energy consumption calculation on the predicted next-moment state set to obtain an instant reward value corresponding to each candidate action; Perform forward propagation calculation on a preset value network according to the instant reward value and the predicted next moment state to obtain a long-term value estimate for each candidate action; The instant reward value and the long-term value estimate are weightedly combined, and combined with the uncertainty analysis of the state transition, to obtain the expected reward value and state transition probability corresponding to each action.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, the distributed collaborative control of the actuators of the air-conditioning system is performed according to the optimal control strategy set, and real-time feedback evaluation of the control effect is performed to obtain a dynamic adjustment parameter set, and the real-time fine-tuning of the optimal control strategy set by the dynamic adjustment parameter set includes: Decomposing and prioritizing the optimal control strategy set to obtain a distributed control instruction sequence for the compressor, the expansion valve and the blower; Coordinated control of each actuator is performed according to the distributed control instruction sequence, and feedback data during the execution process is collected in real time to obtain an original feedback data set including actual output parameters and system response characteristics; Performing noise filtering and feature extraction on the original feedback data set to obtain key performance indicators reflecting the control effect; Deviation analysis and trend prediction are performed based on the key performance indicators and expected control effects to obtain a dynamic adjustment parameter set, and the dynamic adjustment parameter set and the optimal control strategy set are fused and optimized to obtain a fine-tuned optimal control strategy set.

[0011] A second aspect of the present invention provides an in-vehicle air conditioning control device, the in-vehicle air conditioning control device comprising: A data acquisition module is used to perform multi-dimensional real-time acquisition and processing of the environmental parameters inside and outside the vehicle during the operation of the target vehicle to obtain an environmental parameter set of the target vehicle; a thermal comfort calculation module, configured to perform multi-region thermal comfort calculation on a human body thermal balance model of a passenger in the target vehicle according to the environmental parameter set, and obtain a thermal comfort index set for the passenger; An air conditioning state calculation module, used to perform a multi-parameter evaluation on the operating state of the air conditioning system in the target vehicle according to the environmental parameter set and the thermal comfort index set to obtain an air conditioning state evaluation result; A strategy reasoning module, used to perform multi-strategy parallel reasoning according to the air-conditioning state evaluation result through a preset deep reinforcement learning model to obtain an optimal control strategy set for the air-conditioning system; The collaborative control module is used to perform distributed collaborative control on the actuators of the air-conditioning system according to the optimal control strategy set, and to perform real-time feedback evaluation on the control effect to obtain a dynamic adjustment parameter set, and to perform real-time fine-tuning on the optimal control strategy set through the dynamic adjustment parameter set.

[0012] The third aspect of the present invention provides an in-vehicle air conditioning adjustment and control device, comprising: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected via lines; the at least one processor calls the instructions in the memory so that the in-vehicle air conditioning adjustment and control device executes the steps of the above-mentioned in-vehicle air conditioning adjustment and control method.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the steps of the above-mentioned in-vehicle air conditioning control method.

[0014] The above-mentioned in-vehicle air conditioning adjustment and control method, device, equipment and storage medium generate an environmental parameter set by collecting the environmental parameters inside and outside the vehicle in real time during the operation of the vehicle. The passenger's human thermal balance model is calculated based on the parameter set to generate a multi-region thermal comfort index set. Based on the environmental parameter set and the thermal comfort index set, the operating status of the air-conditioning system is evaluated to generate an air-conditioning status evaluation result. Through the preset deep reinforcement learning model, multi-strategy parallel reasoning is performed to generate the optimal control strategy set of the air-conditioning system. According to the strategy set, the air-conditioning actuator is distributedly coordinated and controlled, and the adjustment effect is fed back in real time to generate a dynamic adjustment parameter set to fine-tune the control strategy. This method dynamically adjusts the air-conditioning control strategy according to the real-time environment and thermal comfort through multi-dimensional environmental parameter collection and model reasoning, so as to achieve more efficient air-conditioning system adjustment and improve passenger comfort.

[0015] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0016] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of a first embodiment of a method for controlling in-vehicle air conditioning in an embodiment of the present invention; Figure 2 A schematic diagram of an embodiment of an in-vehicle air conditioning control device according to an embodiment of the present invention; Figure 3 Schematic diagram of an embodiment of an in-vehicle air conditioning control device in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device end including a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or device ends.

[0020] To facilitate understanding of this embodiment, a vehicle air conditioning control method disclosed in an embodiment of the present invention is first described in detail. Figure 1 As shown, the method comprises the following steps: 101. During the operation of the target vehicle, the environmental parameters inside and outside the vehicle are collected and processed in real time in multiple dimensions to obtain an environmental parameter set of the target vehicle; In one embodiment of the present invention, a variety of sensors are installed at strategic positions of the vehicle. Temperature sensors, humidity sensors and solar radiation intensity sensors are installed outside the vehicle, respectively located on the roof, above the front windshield and on the side. A more complex sensor network is installed inside the vehicle, including temperature sensors, humidity sensors and air flow velocity sensors distributed on the driver's head, chest, legs, and corresponding positions of the co-pilot and back seats. In addition, temperature and air flow velocity sensors are installed at the air outlet of the air conditioner, and a temperature sensor is installed on the surface of the evaporator. These sensors transmit data to the central control unit in real time through the on-board CAN bus, and the sampling frequency is set to once per second to capture instantaneous changes in the environment. Next, the data collected by the external environmental sensors are subjected to multi-source fusion processing to obtain external environmental parameters including external temperature, humidity and solar radiation intensity. Then, the data collected by the multi-region sensor network in the vehicle are subjected to adaptive filtering and spatial interpolation processing to obtain internal environmental parameters including temperature, humidity and air flow velocity at the corresponding positions of the driver's head, chest, legs, and co-pilot and back seats. At the same time, the system obtains the current location, time and other information through the vehicle GPS and clock, and combines the pre-installed geographic information database to perform spatiotemporal correlation analysis on these data to infer the current driving environment type, such as city, highway, tunnel, etc. According to the current driving environment type parameters obtained, the system selectively collects data from the vehicle infotainment system to obtain vehicle operating status parameters including vehicle speed and engine load. In order to achieve personalized control, the system also needs to obtain occupant information. By extracting and fusing multimodal features from the data collected by the seat weight sensor and the in-vehicle camera, the physiological characteristic parameters of the occupants including occupant height, weight, age and gender are obtained. These parameters will be used for the subsequent personalized adjustment of the thermal comfort model. Finally, all the collected and processed data, including the external environment parameters, the internal environment parameters, the driving environment type, the vehicle operating status parameters and the physiological characteristic parameters of the occupants, are integrated and standardized to form a complete set of environmental parameters.

[0021] 102. Perform multi-region thermal comfort calculation on a human body thermal balance model of passengers in the target vehicle according to the environmental parameter set to obtain a thermal comfort index set of passengers; In one embodiment of the present invention, performing multi-region thermal comfort calculation on the human body thermal balance model of the passengers in the target vehicle according to the environmental parameter set to obtain the thermal comfort index set of the passengers includes: performing spatial interpolation processing on the temperature data, humidity data and airflow velocity data in the environmental parameter set to obtain local environmental parameters of multiple predefined human body regions of the passengers; performing parameter calibration on a preset first human body thermal balance model according to the local environmental parameters and physiological characteristic parameters of the occupants in the environmental parameter set to obtain a personalized thermal balance equation group; numerically solving the thermal balance equation group to obtain human body thermal state parameters of the passengers, and performing multi-region parallel calculation on a preset second human body thermal balance model according to the human body thermal state parameters to obtain local thermal sensation values ​​of multiple predefined human body regions; and performing weighted averaging and nonlinear mapping processing on the local thermal sensation values ​​to obtain the thermal comfort index of the passengers.

[0022] Specifically, the system performs spatial interpolation processing on the temperature data, humidity data and airflow velocity data in the environmental parameter set. This step uses the Kriging interpolation method. This method takes into account the spatial relationship between the sensor position and various regions of the human body. By establishing a spatial correlation model, the environmental parameter values ​​of the points that are not directly measured in the space around the passenger are estimated. In this way, the system obtains a high-resolution environmental parameter distribution map covering the entire body of the passenger, including local environmental parameters of multiple predefined human body regions such as the head, neck, trunk, upper arm and foot. Next, the system calibrates the parameters of the preset first human thermal balance model based on the local environmental parameters and the physiological characteristic parameters of the occupants in the environmental parameter set to obtain a personalized thermal balance equation set. This step uses the Fanger PMV model as the basis, and adjusts the model by considering individual differences to more accurately reflect the thermal perception of specific passengers. Subsequently, the system numerically solves the obtained personalized thermal balance equation set. Through the solution, the system obtains human thermal state parameters including local skin temperature, average skin temperature and skin temperature change rate. After obtaining the human thermal state parameters, the system inputs these parameters into the preset second human thermal balance model, namely the Zhang local thermal comfort model. This model is specially designed to deal with local thermal comfort problems in non-uniform thermal environments, and is particularly suitable for complex thermal environments in cars. The system performs parallel calculations on 15 predefined human body regions, and the calculation process of each region takes into account local environmental parameters, thermal state parameters, and region-specific physiological characteristics. Through this step, the system obtains the local thermal sensation value (LTS) of each predefined human body region. Finally, the system performs weighted averaging and nonlinear mapping on the obtained local thermal sensation values. The weighted averaging process takes into account the different contributions of various body parts to the overall thermal comfort. For example, the trunk and head usually have a greater impact on the overall thermal sensation than the limbs. The weight coefficient is derived through a machine learning algorithm based on a large amount of experimental data. After weighted averaging, the system uses a sigmoid function for nonlinear mapping to convert the results to a standard 7-point thermal comfort scale ranging from -3 (cold) to +3 (hot), with 0 representing a thermal neutral (most comfortable) state. In addition to the overall thermal sensation (OTS), the system also calculates the predicted percentage dissatisfaction (PPD). PPD is obtained by nonlinearly transforming the OTS value and reflects the proportion of people who may feel uncomfortable in the current thermal environment. In addition, the system introduces a dynamic adaptation factor that takes into account the gradual adaptation process of passengers to the environment. This factor is derived based on time series analysis and personal historical data. Finally, the system combines OTS, PPD and dynamic adaptation factors to calculate a comprehensive thermal comfort index (CTC). CTC not only takes into account the instantaneous thermal sensation, but also includes the prediction of future thermal comfort changes. This index is calculated using a fuzzy logic algorithm and can more comprehensively reflect the overall thermal comfort status of passengers.

[0023] Furthermore, the first human thermal balance model is a Fanger PMV model; the parameters of the preset first human thermal balance model are calibrated according to the local environmental parameters and the physiological characteristic parameters of the occupants in the environmental parameter set to obtain a personalized thermal balance equation group, including: performing a multi-factor correlation analysis on the local environmental parameters and the physiological characteristic parameters of the passengers to obtain a weight matrix of thermal balance influencing factors; dynamically correcting the metabolic rate and clothing thermal resistance parameters in the Fanger PMV model according to the thermal balance influencing factor weight matrix to obtain personalized human heat generation and heat dissipation parameters, and substituting the personalized human heat generation and heat dissipation parameters into the basic equation of the Fanger PMV model to obtain a corrected local PMV calculation equation; performing regional division and parameter configuration on the corrected local PMV calculation equation to obtain local thermal balance equations for multiple predefined areas in the vehicle; and combining and coupling the local thermal balance equations of the multiple predefined areas to obtain a personalized thermal balance equation group.

[0024] Specifically, a multi-factor correlation analysis is first performed on the local environmental parameters and the physiological characteristic parameters of the passengers. This step uses a combination of principal component analysis (PCA) and partial least squares regression (PLS-R). PCA is used to reduce the parameter dimension and extract the main influencing factors, while PLS-R is used to analyze the relationship between these factors and thermal comfort. The analysis process takes into account local environmental parameters such as temperature, humidity, air flow velocity, and physiological characteristics of passengers such as age, gender, height, and weight. Through this analysis, the system obtains a thermal balance influencing factor weight matrix, which quantifies the degree of influence of each factor on thermal balance. Next, the system dynamically corrects the key parameters in the Fanger PMV model based on the thermal balance influencing factor weight matrix. The focus of the correction is on the two parameters of metabolic rate and clothing thermal resistance, because they directly affect the heat generation and heat dissipation process of the human body. The correction process uses an adaptive neural fuzzy inference system (ANFIS), which combines the learning ability of neural networks and the reasoning ability of fuzzy logic. ANFIS dynamically adjusts the calculation formulas of metabolic rate and clothing thermal resistance based on the input of the weight matrix to make them more suitable for individual characteristics. For example, for older passengers, the system will appropriately reduce the estimated metabolic rate; for larger passengers, the calculation method of clothing thermal resistance will be adjusted. In this way, the system obtains personalized human body heat generation and heat dissipation parameters. Then, the system substitutes these personalized human body heat generation and heat dissipation parameters into the basic equation of the Fanger PMV model. The basic equation of the Fanger PMV model is a complex heat balance equation that includes multiple components such as human body heat generation, convection heat dissipation, radiation heat dissipation, and evaporative heat dissipation. By substituting personalized parameters into this equation, the system obtains a revised local PMV calculation equation. This revised equation is no longer a general PMV calculation formula, but a personalized PMV calculation method for specific passengers in specific environments. After obtaining the revised local PMV calculation equation, the system performs regional division and parameter configuration. The interior space is divided into multiple predefined areas, each of which is configured with a specific set of parameters, including the average temperature, humidity, air flow velocity, and special parameters related to that part of the human body (such as the head is more sensitive to temperature changes). This regional division and parameter configuration enable the system to more accurately calculate the local thermal comfort of each body part. Finally, the system combines and couples the local thermal balance equations of multiple predefined areas. This step uses multi-scale modeling technology to link the microscopic local thermal balance with the macroscopic overall thermal comfort. Specifically, the system establishes a hierarchical set of equations, with the bottom layer being the local thermal balance equations of each area, the middle layer being the thermal balance equations of the main parts of the body (such as the trunk and limbs), and the top layer being the overall thermal balance equation. These equations are coupled to each other through factors such as heat flow and blood circulation to form a complete set of equations.The nonlinear dynamic system theory is used in the coupling process, taking into account the heat exchange and mutual influence between various parts of the human body.

[0025] 103. Perform a multi-parameter evaluation on the operating state of the air-conditioning system in the target vehicle according to the environmental parameter set and the thermal comfort index set to obtain an air-conditioning state evaluation result; In one embodiment of the present invention, the multi-parameter evaluation of the operating status of the air-conditioning system in the target vehicle is performed according to the environmental parameter set and the thermal comfort index set to obtain the air-conditioning status evaluation result, including: performing time series analysis and trend prediction on the environmental parameter set and the thermal comfort index set to obtain the in-vehicle thermal environment change trend data; performing dynamic response analysis on the operating parameters of each component of the air-conditioning system in the target vehicle according to the in-vehicle thermal environment change trend data to obtain the air-conditioning system status parameter set; performing deviation calculation between the air-conditioning system status parameter set and a preset optimal operating parameter range to obtain the performance efficiency index of each component in the air-conditioning system; performing multi-objective optimization analysis on the air-conditioning system according to the performance efficiency index and the thermal comfort index set to obtain a comprehensive performance evaluation index; performing weighted fusion processing and quantitative grading processing on the comprehensive performance evaluation index and the air-conditioning system status parameter set to obtain the air-conditioning status evaluation result.

[0026] Specifically, first, the environmental parameter set and thermal comfort index set are analyzed and trend predicted. This step uses time series analysis technology, specifically the ARIMA (autoregressive integrated moving average) model combined with the Kalman filter. The ARIMA model is used to capture the time dependence of the data, while the Kalman filter is used to deal with measurement noise and system dynamics. The system analyzes the environmental parameters and thermal comfort indicators over a period of time (usually 30 minutes) in the past, and identifies the change patterns of key indicators such as temperature, humidity, and thermal comfort. Then, based on these patterns, the system uses the long short-term memory network (LSTM) for short-term prediction to predict the trend of the in-vehicle thermal environment changes within the next 15 minutes. This hybrid approach combining statistical methods and deep learning can effectively handle the complex dynamic characteristics of the in-vehicle environment and obtain accurate in-vehicle thermal environment change trend data. Next, the system performs a dynamic response analysis on the operating parameters of each component of the air-conditioning system based on the obtained in-vehicle thermal environment change trend data. This step uses state space models and system identification technology. First, a state space model of the air-conditioning system is established, including key components such as the compressor, condenser, evaporator, and expansion valve. Then, the model parameters are estimated in real time using the recursive least squares (RLS) method to adapt to the dynamic characteristics of the system. Based on this dynamic model, the system predicts how the operating parameters of each component (such as compressor speed, condenser and evaporator temperature, refrigerant pressure, expansion valve opening, blower speed, etc.) will change under the predicted trend of thermal environment changes. This prediction takes into account the interaction between components and thermodynamic constraints, thereby obtaining a comprehensive set of air-conditioning system state parameters. Subsequently, the system calculates the deviation between the obtained air-conditioning system state parameter set and the preset optimal operating parameter range. The optimal operating parameter range is predetermined based on a large amount of experimental data and theoretical analysis, and represents the ideal operating state of the air-conditioning system under different working conditions. The system uses fuzzy mathematics methods to calculate the deviation between the actual parameters and the optimal range, which can handle the fuzziness and uncertainty of the parameter range. Through this deviation calculation, the system obtains the performance efficiency indicators of each component of the air-conditioning system, which reflect the gap between the current air-conditioning system operating efficiency and the ideal state. Then, the system performs a multi-objective optimization analysis of the air-conditioning system based on the obtained performance efficiency indicators and the previously calculated thermal comfort indicator set. This step uses a multi-objective genetic algorithm (NSGA-II) to simultaneously consider multiple objectives such as thermal comfort, energy efficiency, and system stability. The algorithm generates a series of non-dominated solutions through iterative optimization, each of which represents a possible air conditioning control strategy. The system then uses the TOPSIS (top-up to ideal solution ranking method) method to select the best solution from these non-dominated solutions. The result of this multi-objective optimization analysis is a set of comprehensive performance evaluation indicators, including the estimated time to reach the target comfort level, control efficiency score, system stability evaluation, and energy consumption level evaluation.Finally, the system performs weighted fusion processing and quantitative grading processing on the obtained comprehensive performance evaluation indicators and air conditioning system status parameter sets. The weighted fusion process uses the analytic hierarchy process (AHP) to determine the weights of each indicator, taking into account the relative importance of different indicators to the overall performance. Then, the system uses the fuzzy comprehensive evaluation method to integrate these weighted indicators into a multidimensional performance vector, where each dimension represents a key performance indicator, such as thermal comfort, energy efficiency, system response speed, stability, etc. Each element in this multidimensional vector is a continuous value, reflecting the performance level of the corresponding indicator. For quantitative grading processing, the system uses the Gaussian mixture model (GMM) to estimate the probability density of this multidimensional vector. GMM is able to capture the complex distribution of data and assign a probability distribution to each data point instead of a simple category label.

[0027] 104. Perform multi-strategy parallel reasoning based on the air-conditioning state evaluation results through the preset deep reinforcement learning model to obtain the optimal control strategy set for the air-conditioning system; In one embodiment of the present invention, the preset deep reinforcement learning model performs multi-strategy parallel reasoning according to the air-conditioning state evaluation result to obtain the optimal control strategy set for the air-conditioning system, including: extracting features and normalizing the air-conditioning state evaluation result to obtain a state vector suitable for the deep reinforcement learning model; performing forward propagation calculation on the policy network of the preset deep reinforcement learning model according to the state vector to obtain multiple candidate control actions; performing parallel simulation and evaluation on the multiple candidate control actions to obtain the expected reward value and state transition probability corresponding to each action; sorting and screening the candidate control actions according to the expected reward value and state transition probability to obtain the optimal control strategy set, wherein the optimal control strategy set includes compressor speed adjustment, blower air volume adjustment for each zone, and air outlet mode selection.

[0028] Specifically, the air conditioner state evaluation results are first feature extracted and normalized. This step uses the principal component analysis (PCA) method for feature extraction to compress the high-dimensional air conditioner state evaluation results to a lower dimension while retaining key information. Then, the extracted features are normalized using the Z-score normalization method to convert features of different scales to the same scale. The state vector obtained after this processing is suitable for the input of the deep reinforcement learning model, ensuring the stability and effect of model training. Next, the system performs forward propagation calculations on the policy network of the preset deep reinforcement learning model based on the obtained state vector. The deep reinforcement learning model used here is based on the PPO (Proximal Policy Optimization) algorithm, which performs well in continuous control tasks. The policy network adopts a multi-layer perceptron (MLP) structure, which contains multiple fully connected layers and ReLU activation functions. During the forward propagation process, the state vector passes through these layers in turn and finally outputs an action distribution. In order to increase exploration and adaptability, the system adds Gaussian noise to the output layer so that the model can generate multiple candidate control actions. These candidate actions include parameters such as the compressor speed adjustment, the blower air volume adjustment in each area, and the air outlet mode selection. Then, the system simulates and evaluates these multiple candidate control actions in parallel to obtain the expected reward value and state transition probability corresponding to each action. This step utilizes the air conditioning system dynamics model and thermal comfort evaluation model established previously, and improves processing efficiency through parallel computing technology. Subsequently, the system sorts and screens the candidate control actions according to the expected reward value and state transition probability obtained. The sorting process adopts a multi-objective sorting algorithm, while considering the uncertainty of the expected reward value and state transition probability. Specifically, the system uses the expected improvement (EI) criterion to balance exploration and utilization. The EI criterion not only considers the size of the expected reward, but also the uncertainty of the prediction, which helps to strike a balance between the known optimal solution and the potential better solution. After sorting, the system selects the top-K actions as candidate optimal control strategies, where the K value is dynamically adjusted according to the real-time computing resources and control accuracy requirements. Finally, the system further refines and locally searches these K candidate optimal control strategies. This step uses model-based optimization methods, such as Gaussian Process Regression (GPR). The GPR model can interpolate and extrapolate the entire decision space based on limited sampling points, so as to find better control parameters near the candidate strategy. In this way, the system can achieve local fine-tuning while ensuring global exploration. After this series of processing, the system finally obtains the optimal control strategy set. This strategy set includes precise compressor speed adjustment, blower air volume adjustment in each area, and air outlet mode selection.The compressor speed adjustment is in RPM (revolutions per minute), accurate to the single digit; the blower air volume adjustment for each area is in cubic meters per hour (m³ / h). Taking into account the different needs of different areas in the car, adjustments are given for the driver area, co-pilot area and rear area respectively; the air outlet mode selection includes a combination of multiple options such as upper air supply, lower air supply, defrost mode, etc.

[0029] Furthermore, the parallel simulation and evaluation of the multiple candidate control actions to obtain the expected reward value and state transition probability corresponding to each action includes: combining the multiple candidate control actions and the state vector to obtain multiple groups of state-action pairs; solving the preset in-vehicle thermal environment dynamics model in parallel according to the multiple groups of state-action pairs to obtain a set of predicted states at the next moment; performing thermal comfort evaluation and energy consumption calculation on the predicted set of states at the next moment to obtain an immediate reward value corresponding to each candidate action; performing forward propagation calculation on a preset value network according to the immediate reward value and the predicted state at the next moment to obtain a long-term value estimate for each candidate action; performing weighted combination of the immediate reward value and the long-term value estimate, and combining with uncertainty analysis of state transition, to obtain the expected reward value and state transition probability corresponding to each action.

[0030] Specifically, multiple candidate control actions and state vectors are first combined to obtain multiple sets of state-action pairs. This step uses the Cartesian product operation to pair each candidate control action with the current state vector. Specifically, the system combines each control action (including the compressor speed adjustment, the blower air volume adjustment in each area, and the air outlet mode selection) with a vector representing the current in-vehicle environment and the state of the air conditioning system. This combination ensures a comprehensive evaluation of the effect of each possible control decision in the current state. Next, the system solves the preset in-vehicle thermal environment dynamics model in parallel based on the multiple sets of state-action pairs obtained to obtain the predicted state set at the next moment. The in-vehicle thermal environment dynamics model used here is a hybrid model based on physical principles and data-driven. The model contains thermodynamic equations, fluid mechanics equations, and nonlinear mapping functions trained by machine learning. The system uses the RungeKutta method to numerically solve this complex nonlinear model and predict the changes in parameters such as in-vehicle temperature, humidity, and airflow distribution in the future in a short period of time (such as 5 minutes) under each control action. In order to improve computing efficiency, the system uses GPU acceleration technology and the CUDA parallel computing framework to simultaneously process the model solution of multiple sets of state-action pairs. Then, the system performs thermal comfort evaluation and energy consumption calculation on the predicted next-moment state set to obtain the immediate reward value corresponding to each candidate action. The thermal comfort evaluation uses an improved PMV-PPD model, which takes into account the non-uniformity and individual differences of the in-vehicle environment. The energy consumption calculation is based on the thermodynamic model of the air-conditioning system, taking into account factors such as compressor power consumption and blower power consumption. The system combines the thermal comfort index and energy consumption index into an immediate reward value by weighted summation. The weight coefficient is determined by a multi-objective optimization method to balance comfort and energy efficiency. Then, the system performs forward propagation calculation on the preset value network based on the immediate reward value and the predicted next-moment state to obtain the long-term value estimate of each candidate action. The value network adopts a deep neural network structure, which contains multiple fully connected layers and LSTM layers, and can capture the time dependency of the state sequence. The network input is the current state, the predicted next-moment state and the immediate reward, and the output is the estimated long-term cumulative reward. This design allows the system to consider the long-term impact of control actions, not just the immediate effects. Finally, the system performs a weighted combination of the immediate reward value and the long-term value estimate, and combines it with the uncertainty analysis of the state transition to obtain the expected reward value and state transition probability corresponding to each action. The weighted combination uses the idea of ​​temporal difference learning (TD-learning) to combine the immediate reward and the discounted long-term value estimate. The uncertainty analysis of state transition uses the Monte Carlo dropout method, which calculates the mean and variance of the state prediction through multiple forward propagation calculations.Based on this information, the system uses Gaussian process regression (GPR) to build a probability model that not only gives a point estimate of the expected reward value, but also provides a confidence interval that reflects the uncertainty of the prediction. At the same time, by analyzing the probability distribution of state predictions, the system obtains the transition probability from the current state to the predicted state.

[0031] 105. Distributed collaborative control is performed on the actuators of the air-conditioning system according to the optimal control strategy set, and real-time feedback evaluation is performed on the control effect to obtain a dynamic adjustment parameter set. Through the dynamic adjustment parameter set, the optimal control strategy set is fine-tuned in real time.

[0032] In one embodiment of the present invention, distributed collaborative control is performed on the actuators of the air-conditioning system according to the optimal control strategy set, and real-time feedback evaluation is performed on the control effect to obtain a dynamic adjustment parameter set. The optimal control strategy set is fine-tuned in real time through the dynamic adjustment parameter set, including: task decomposition and priority sorting of the optimal control strategy set to obtain a distributed control instruction sequence for the compressor, expansion valve and blower; collaborative control of each actuator according to the distributed control instruction sequence, and real-time collection of feedback data during the execution process to obtain an original feedback data set including actual output parameters and system response characteristics; noise filtering and feature extraction are performed on the original feedback data set to obtain key performance indicators reflecting the control effect; deviation analysis and trend prediction are performed based on the key performance indicators and the expected control effect to obtain a dynamic adjustment parameter set, and the dynamic adjustment parameter set and the optimal control strategy set are fused and optimized to obtain a fine-tuned optimal control strategy set.

[0033] Specifically, the optimal control strategy set is firstly decomposed and prioritized. This step uses a combination of the analytic hierarchy process (AHP) and the task dependency graph to decompose the overall control strategy into control instructions for specific actuators such as compressors, expansion valves and blowers. The AHP method is used to determine the relative importance of the control tasks of each actuator, while the task dependency graph is used to analyze the logical relationship and execution order between the control instructions. Through this decomposition and sorting, the system obtains an ordered distributed control instruction sequence, ensuring that each actuator can coordinate and execute the optimal control strategy. Next, the system performs collaborative control on each actuator according to the distributed control instruction sequence, and collects feedback data in real time during the execution process. The collaborative control process adopts the distributed model predictive control (DMPC) framework. Each actuator is equipped with a local controller, which exchanges information through a communication network to achieve global coordination. At the same time, the system collects the operating data of each actuator in real time through the high-speed data acquisition module, including the actual speed and power consumption of the compressor, the opening and flow of the expansion valve, the air volume and outlet temperature of the blower in each area, and the dynamic response characteristics of these parameters over time. These data constitute the original feedback data set, which provides a basis for the subsequent control effect evaluation. Then, the system performs noise filtering and feature extraction on the collected original feedback data set. The noise filtering adopts the wavelet transform method, which can effectively remove high-frequency noise while retaining the important features of the signal. The feature extraction uses the combination of principal component analysis (PCA) and independent component analysis (ICA) to extract the key features that best reflect the control effect from the multidimensional data. Through these processes, the system obtains a set of key performance indicators that reflect the control effect, including system response speed, steady-state error, energy efficiency, temperature uniformity, etc. Then, the system performs deviation analysis and trend prediction based on the obtained key performance indicators and the expected control effect. The deviation analysis uses the fuzzy comprehensive evaluation method to compare the actual value of each indicator with the expected value in multiple dimensions to obtain a comprehensive deviation score. Trend prediction uses a long short-term memory network (LSTM) to predict the changing trends of various indicators in the future based on historical data and current status. Through these analyses and predictions, the system generates a dynamic adjustment parameter set, which contains the direction and amplitude of adjustment of each control parameter. Finally, the system integrates and optimizes the dynamic adjustment parameter set and the current optimal control strategy set to obtain the optimal control strategy set after fine-tuning. This step uses the Rolling Horizon Optimization method to make local adjustments to specific control parameters while maintaining the overall framework of the original control strategy.The optimization process takes into account multiple objectives, including improving control accuracy, reducing energy consumption, and improving passenger comfort, and uses a multi-objective particle swarm optimization algorithm (MOPSO) to balance these potentially conflicting objectives.

[0034] In this embodiment, during the operation of the vehicle, the environmental parameters inside and outside the vehicle are collected in real time to generate an environmental parameter set. The passenger's human thermal balance model is calculated based on the parameter set to generate a multi-zone thermal comfort index set. Based on the environmental parameter set and the thermal comfort index set, the operating status of the air-conditioning system is evaluated to generate an air-conditioning status evaluation result. Through the preset deep reinforcement learning model, multi-strategy parallel reasoning is performed to generate an optimal control strategy set for the air-conditioning system. According to the strategy set, the air-conditioning actuator is distributedly coordinated and controlled, and the adjustment effect is fed back in real time to generate a dynamic adjustment parameter set to fine-tune the control strategy. This method dynamically adjusts the air-conditioning control strategy according to the real-time environment and thermal comfort through multi-dimensional environmental parameter collection and model reasoning, so as to achieve more efficient air-conditioning system adjustment and improve passenger comfort.

[0035] The above describes the in-vehicle air conditioning control method according to the embodiment of the present invention. The following describes the in-vehicle air conditioning control device according to the embodiment of the present invention. Figure 2 In one embodiment of the present invention, an in-vehicle air conditioning control device includes: The data acquisition module 201 is used to perform multi-dimensional real-time acquisition and processing of the environmental parameters inside and outside the vehicle during the operation of the target vehicle to obtain the environmental parameter set of the target vehicle; A thermal comfort calculation module 202 is used to perform multi-region thermal comfort calculation on a human body thermal balance model of a passenger in the target vehicle according to the environmental parameter set to obtain a thermal comfort index set of the passenger; Air conditioning state calculation module 203, for performing a multi-parameter evaluation on the operating state of the air conditioning system in the target vehicle according to the environmental parameter set and the thermal comfort index set to obtain an air conditioning state evaluation result; A strategy reasoning module 204 is used to perform multi-strategy parallel reasoning according to the air-conditioning state evaluation result through a preset deep reinforcement learning model to obtain an optimal control strategy set for the air-conditioning system; The collaborative control module 205 is used to perform distributed collaborative control on the actuators of the air-conditioning system according to the optimal control strategy set, and to perform real-time feedback evaluation on the control effect to obtain a dynamic adjustment parameter set, and to perform real-time fine-tuning on the optimal control strategy set through the dynamic adjustment parameter set.

[0036] In an embodiment of the present invention, the in-vehicle air conditioning adjustment and control device runs the in-vehicle air conditioning adjustment and control method. The in-vehicle air conditioning adjustment and control device generates an environmental parameter set by collecting the environmental parameters inside and outside the vehicle in real time during the operation of the vehicle. The passenger's human thermal balance model is calculated based on the parameter set to generate a multi-region thermal comfort index set. Based on the environmental parameter set and the thermal comfort index set, the operating state of the air-conditioning system is evaluated to generate an air-conditioning state evaluation result. Through a preset deep reinforcement learning model, multi-strategy parallel reasoning is performed to generate an optimal control strategy set for the air-conditioning system. According to the strategy set, the air-conditioning actuator is distributedly coordinated and controlled, and the adjustment effect is fed back in real time to generate a dynamic adjustment parameter set to fine-tune the control strategy. This method dynamically adjusts the air-conditioning control strategy according to the real-time environment and thermal comfort through multi-dimensional environmental parameter collection and model reasoning, so as to achieve more efficient air-conditioning system adjustment and improve passenger comfort.

[0037] above Figure 2 The in-vehicle air conditioning adjustment control device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The in-vehicle air conditioning adjustment control device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0038] Figure 3 3 is a structural diagram of an in-vehicle air conditioning control device provided by an embodiment of the present invention. The in-vehicle air conditioning control device 300 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage device terminals) storing application programs 333 or data 332. Among them, the memory 320 and the storage medium 330 may be temporary storage or permanent storage. The program stored in the storage medium 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the in-vehicle air conditioning control device 300. Furthermore, the processor 310 may be configured to communicate with the storage medium 330, and execute a series of instruction operations in the storage medium 330 on the in-vehicle air conditioning control device 300 to implement the steps of the in-vehicle air conditioning control method.

[0039] The in-vehicle air conditioning control device 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that Figure 3 The illustrated structure of the in-vehicle air conditioning control device does not constitute a limitation on the in-vehicle air conditioning control device provided by the present invention, and may include more or fewer components than illustrated, or a combination of certain components, or a different arrangement of components.

[0040] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the in-vehicle air conditioning control method.

[0041] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0042] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0043] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for controlling in-vehicle air conditioning, characterized in that: The in-vehicle air conditioning adjustment and control method comprises: During the operation of the target vehicle, multi-dimensional real-time collection and processing of the environmental parameters inside and outside the vehicle are performed to obtain the environmental parameter set of the target vehicle; Performing multi-region thermal comfort calculation on a human body thermal balance model of passengers in the target vehicle according to the environmental parameter set to obtain a thermal comfort index set for the passengers; Performing a multi-parameter evaluation on the operating state of the air conditioning system in the target vehicle according to the environmental parameter set and the thermal comfort index set to obtain an air conditioning state evaluation result; Performing multi-strategy parallel reasoning based on the air-conditioning state evaluation result through a preset deep reinforcement learning model to obtain an optimal control strategy set for the air-conditioning system; According to the optimal control strategy set, distributed collaborative control is performed on the actuators of the air-conditioning system, and real-time feedback evaluation is performed on the control effect to obtain a dynamic adjustment parameter set. Through the dynamic adjustment parameter set, the optimal control strategy set is fine-tuned in real time.

2. The in-vehicle air conditioning control method according to claim 1, characterized in that: The multi-region thermal comfort calculation of the human body thermal balance model of the passenger in the target vehicle according to the environmental parameter set to obtain the thermal comfort index set of the passenger includes: Performing spatial interpolation processing on the temperature data, humidity data and airflow velocity data in the environmental parameter set to obtain local environmental parameters of a plurality of predefined human body regions of the passenger; Calibrate the parameters of the preset first human body thermal balance model according to the local environmental parameters and the physiological characteristic parameters of the occupant in the environmental parameter set to obtain a personalized thermal balance equation group; Numerically solving the thermal balance equations to obtain the passenger's human thermal state parameters, and performing multi-region parallel calculations on a preset second human thermal balance model according to the human thermal state parameters to obtain local thermal sensation values ​​of multiple predefined human regions; The local thermal sensation values ​​are subjected to weighted averaging and nonlinear mapping processing to obtain a thermal comfort index of the passenger.

3. The in-vehicle air conditioning control method according to claim 2, characterized in that: The first human body thermal balance model is the Fanger PMV model; The parameter calibration of the preset first human body thermal balance model according to the local environmental parameters and the physiological characteristic parameters of the occupant in the environmental parameter set to obtain a personalized thermal balance equation group includes: Performing a multi-factor correlation analysis on the local environmental parameters and the physiological characteristic parameters of the passengers to obtain a weight matrix of thermal balance influencing factors; Dynamically correcting the metabolic rate and clothing thermal resistance parameters in the Fanger PMV model according to the weight matrix of the thermal balance influencing factors to obtain personalized human body heat generation and heat dissipation parameters, and substituting the personalized human body heat generation and heat dissipation parameters into the basic equation of the Fanger PMV model to obtain a corrected local PMV calculation equation; Performing regional division and parameter configuration on the modified local PMV calculation equation to obtain local heat balance equations for multiple predefined areas in the vehicle; The local heat balance equations of the plurality of predefined regions are combined and coupled to obtain a personalized heat balance equation group.

4. The in-vehicle air conditioning control method according to claim 1, characterized in that: The multi-parameter evaluation of the operating state of the air conditioning system in the target vehicle according to the environmental parameter set and the thermal comfort index set to obtain the air conditioning state evaluation result includes: Performing time series analysis and trend prediction on the environmental parameter set and thermal comfort index set to obtain in-vehicle thermal environment change trend data; Performing a dynamic response analysis on the operating parameters of each component of the air-conditioning system in the target vehicle according to the in-vehicle thermal environment change trend data to obtain an air-conditioning system state parameter set; Calculating the deviation between the state parameter set of the air conditioning system and the preset optimal operating parameter range to obtain the performance efficiency index of each component in the air conditioning system; According to the performance efficiency index and the thermal comfort index set, a multi-objective optimization analysis is performed on the air conditioning system to obtain a comprehensive performance evaluation index; The comprehensive performance evaluation index and the air conditioning system state parameter set are subjected to weighted fusion processing and quantitative grading processing to obtain an air conditioning state evaluation result.

5. The in-vehicle air conditioning control method according to claim 1, characterized in that: The method of performing multi-strategy parallel reasoning according to the air-conditioning state evaluation result through a preset deep reinforcement learning model to obtain an optimal control strategy set for the air-conditioning system includes: Performing feature extraction and normalization processing on the air conditioning state evaluation result to obtain a state vector suitable for a deep reinforcement learning model; Performing forward propagation calculation on a policy network of a preset deep reinforcement learning model according to the state vector to obtain multiple candidate control actions; Performing parallel simulation and evaluation on the multiple candidate control actions to obtain an expected reward value and a state transition probability corresponding to each action; The candidate control actions are sorted and screened according to the expected reward value and the state transition probability to obtain an optimal control strategy set, wherein the optimal control strategy set includes the compressor speed adjustment amount, the blower air volume adjustment amount of each area, and the air outlet mode selection.

6. The in-vehicle air conditioning control method according to claim 5, characterized in that: The parallel simulation and evaluation of the multiple candidate control actions to obtain the expected reward value and state transition probability corresponding to each action includes: Combining the multiple candidate control actions and the state vector to obtain multiple groups of state-action pairs; Solving the preset in-vehicle thermal environment dynamics model in parallel according to the multiple groups of state-action pairs to obtain a predicted state set at the next moment; Performing thermal comfort evaluation and energy consumption calculation on the predicted next-moment state set to obtain an instant reward value corresponding to each candidate action; Perform forward propagation calculation on a preset value network according to the instant reward value and the predicted next moment state to obtain a long-term value estimate for each candidate action; The instant reward value and the long-term value estimate are weightedly combined, and combined with the uncertainty analysis of the state transition, to obtain the expected reward value and state transition probability corresponding to each action.

7. The in-vehicle air conditioning control method according to claim 1, characterized in that: The distributed collaborative control of the actuators of the air-conditioning system according to the optimal control strategy set and the real-time feedback evaluation of the control effect are performed to obtain a dynamic adjustment parameter set. The real-time fine-tuning of the optimal control strategy set by the dynamic adjustment parameter set includes: Decomposing and prioritizing the optimal control strategy set to obtain a distributed control instruction sequence for the compressor, the expansion valve and the blower; Coordinated control of each actuator is performed according to the distributed control instruction sequence, and feedback data during the execution process is collected in real time to obtain an original feedback data set including actual output parameters and system response characteristics; Performing noise filtering and feature extraction on the original feedback data set to obtain key performance indicators reflecting the control effect; Deviation analysis and trend prediction are performed based on the key performance indicators and expected control effects to obtain a dynamic adjustment parameter set, and the dynamic adjustment parameter set and the optimal control strategy set are fused and optimized to obtain a fine-tuned optimal control strategy set.

8. An in-vehicle air conditioning control device, characterized in that: The in-vehicle air conditioning control device comprises: A data acquisition module is used to perform multi-dimensional real-time acquisition and processing of the environmental parameters inside and outside the vehicle during the operation of the target vehicle to obtain an environmental parameter set of the target vehicle; a thermal comfort calculation module, configured to perform multi-region thermal comfort calculation on a human body thermal balance model of a passenger in the target vehicle according to the environmental parameter set, and obtain a thermal comfort index set for the passenger; An air conditioning state calculation module, used to perform a multi-parameter evaluation on the operating state of the air conditioning system in the target vehicle according to the environmental parameter set and the thermal comfort index set to obtain an air conditioning state evaluation result; A strategy reasoning module, used to perform multi-strategy parallel reasoning according to the air-conditioning state evaluation result through a preset deep reinforcement learning model to obtain an optimal control strategy set for the air-conditioning system; The collaborative control module is used to perform distributed collaborative control on the actuators of the air-conditioning system according to the optimal control strategy set, and to perform real-time feedback evaluation on the control effect to obtain a dynamic adjustment parameter set, and to perform real-time fine-tuning on the optimal control strategy set through the dynamic adjustment parameter set.

9. An in-vehicle air conditioning control device, characterized in that: The in-vehicle air conditioning control device comprises: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory to enable the in-vehicle air conditioning adjustment and control device to execute the steps of the in-vehicle air conditioning adjustment and control method as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the in-vehicle air conditioning adjustment and control method as described in any one of claims 1-7 are implemented.

Images