To optimize thermal comfort for occupants’ wellbeing and health care, it's essential to adjust heating and cooling systems in real-time based on occupants' thermal preferences. For this, personal factors affect individual thermal comfort, such as metabolic rate and clothing insulation, should be estimated in real-time. The aim of this research is introducing an intelligent model capable of estimating metabolic rate and clothing insulation values from indoor images, suitable for both single and multi-occupant scenarios. Additionally, a control algorithm considering a real-time predicted mean vote (PMV), was developed using the proposed model, and its implications for thermal comfort and energy efficiency were investigated. Utilizing advanced computer vision methodologies, the model achieved a remarkable 95% training accuracy, and its reliability was further validated through experimentation. Evaluations of the PMV-based algorithm underscored its efficacy in enhancing thermal comfort relative to conventional methods in both individual and multi-occupant settings. Conversely, energy use was contingent upon the personal factors. In group settings, the mode values of metabolic rate and clothing insulation were effective for determining a representative PMV. In conclusion, the real-time PMV-based control represents a pioneering approach to augment thermal comfort using actual occupant data, paving the way for a synergistic balance between comfort augmentation and energy saving.

中文翻译：

---

使用估计的居住者个人代谢率和服装隔热因素，通过实时 PMV 进行热控制的影响

为了优化居住者的健康和医疗保健的热舒适度，必须根据居住者的热偏好实时调整供暖和制冷系统。为此，应实时估计影响个人热舒适度的个人因素，例如新陈代谢率和衣服隔热性。本研究的目的是引入一种智能模型，能够根据室内图像估计代谢率和服装隔热值，适用于单人和多人场景。此外，使用所提出的模型开发了一种考虑实时预测平均投票（PMV）的控制算法，并研究了其对热舒适性和能源效率的影响。利用先进的计算机视觉方法，该模型实现了高达 95% 的训练准确率，并通过实验进一步验证了其可靠性。对基于 PMV 的算法的评估强调了其在个人和多人环境中相对于传统方法提高热舒适性的功效。相反，能源使用取决于个人因素。在群体设置中，代谢率和服装隔热的众数对于确定代表性 PMV 是有效的。总之，基于 PMV 的实时控制代表了一种利用实际乘员数据增强热舒适度的开创性方法，为增强舒适度和节能之间的协同平衡铺平了道路。