The invention discloses an intelligent automobile air conditioner comfort control method based on deep learning and reinforcement learning, and relates to the technical field of non-contact intelligent vehicle-mounted air conditioners. According to the technical concept, an air conditioner model is trained on a simulation platform through deep learning and reinforcement learning, control over a real vehicle air conditioner system is optimized, the system can directly evaluate the sendible temperature of a passenger through the non-contact face thermal imaging technology instead of depending on a temperature sensor in a passenger compartment, and the system can directly evaluate the sendible temperature of the passenger through the non-contact face thermal imaging technology. The method more accurately reflects the actual comfort degree of passengers, the compressor rotating speed and the air blower flow are controlled through the reinforcement learning model, the system can achieve more accurate and dynamic control so as to meet the real-time comfort degree requirement of the passengers, the deep learning model is used for predicting the comfort degree of the human body, and the comfort degree of the passengers is improved. The human body comfort is predicted by using the ResNet50 deep learning network of the enhanced local features, and the higher the prediction speed of the ResNet50 deep learning network of the enhanced local features is, the higher the accuracy is.

    本发明公开了基于深度学习和强化学习的智能汽车空调舒适性控制方法,涉及非接触式智能车载空调技术领域。本发明技术构思是在仿真平台对深度学习和强化学习对空调模型进行训练,实现对实车空调系统控制的优化,这种方法通过非接触式人脸热成像技术,系统能够直接评估乘员的体感温度,而不是依赖于乘员舱内的温度传感器,这种方法更准确地反映了乘员的实际舒适度,且利用强化学习模型控制压缩机转速和鼓风机流量,系统能够实现更加精确和动态的控制,以适应乘员的实时舒适度需求,且用深度学习模型来预测人体的舒适性,利用增强局部特征的ResNet50的深度学习网络对人体舒适性进行预测,增强局部特征的ResNet50的深度学习网络预测的速度越快,准确率越高。


    Access

    Download


    Export, share and cite



    Title :

    Intelligent automobile air conditioner comfort control method based on deep learning and reinforcement learning


    Additional title:

    基于深度学习和强化学习的智能汽车空调舒适性控制方法


    Contributors:
    ZHANG YAN (author) / GUO CONG (author) / QI CHUANG (author) / HE LIANGE (author) / WU LIMIN (author) / WU QINGLONG (author) / YANG KANGWEI (author)

    Publication date :

    2025-04-15


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / B60H ARRANGEMENTS OR ADAPTATIONS OF HEATING, COOLING, VENTILATING, OR OTHER AIR-TREATING DEVICES SPECIALLY FOR PASSENGER OR GOODS SPACES OF VEHICLES , Anordnung oder Einbau der Heizung, Kühlung, Lüftung oder anderer Luftbehandlungsvorrichtungen für die Fahrzeugräume für Reisende oder Fracht



    New energy automobile charging intelligent scheduling method based on deep reinforcement learning

    DENG QINGXU / HAO SHUJING / XING WENXUAN | European Patent Office | 2024

    Free access

    Ride Comfort Enhancement Using Deep Reinforcement Learning

    Khandavalli, Guru Bhargava / Eckstein, Lutz | TIBKAT | 2020


    Intelligent automobile decision-making method based on driving intention and deep reinforcement learning

    PEI XIAOFEI / LU SONGXIN / YANG BO | European Patent Office | 2024

    Free access

    Automobile thermal management intelligent control method based on confrontation reinforcement learning

    LIU YANG / LIU JIAN / OU JUNFENG et al. | European Patent Office | 2023

    Free access

    Automobile emergency collision avoidance control method based on DQN deep reinforcement learning

    LU XIAOHUI / ZHENG XINYI / LYU XINZHAN et al. | European Patent Office | 2023

    Free access