非侵入性人體熱感覺預測系統

Non-invasive Human Thermal Sensation Predict System

陳冠瑜(Chen Kuan-Yu)

指導教授 ： 劉寅春

淡江大學/工學院/電機工程學系碩士班/碩士(2022年)

https://doi.org/10.6846/TKU.2022.00266

摘要

本論文開發了一個人化熱感覺預測系統，偵測並記錄人體臉部溫度數據結合機器學習達到預測人體熱感覺的目標。本論文建立了一測試環境，針對三位受試者進行臉部溫度資料蒐集，利用冷氣及暖氣使環境溫度在20至30度之間改變，受試者將在此環境中利用投票來表達當下之熱感覺，並且在投票的當下利用FLIR Lepton3.5紀錄受試者的和臉部鼻子、耳朵、臉頰、額頭四部位的溫度資料。臉部溫度資料將拆成80%訓練集及20%的測試集，將訓練集輸入支援向量回歸、隨機森林、人工神經網路三種機器學習方法進行模型訓練，最終將訓練好的三個機器學習模型與PMV針對測試集中的數據去做預測，再將四個模型的預測結果進行比較找出最佳模型。
實驗結果表明，三種機器學習方法之預測結果皆優於傳統之PMV模型，而支援向量回歸模型為最佳模型，具有良好的準確率穩定性，測試者三人的支援向量回歸平均MAE為0.359，平均Adjusted R-Square為0.886。
本論文成功開發出一個人化熱感覺預測系統，利用鼻子、耳朵、臉頰三處溫度來較準確的判斷人體之熱感覺，其成果若結合辦公室HVAC之調控，便能使環境中各辦公人員都能達到良好之熱舒適性，大幅改善辦公室之室內環境舒適度，提升辦公人員之辦公效率，也可用於偵測人在睡覺時的臉部溫度變化，預測人睡著時的熱感覺自動調控HVAC溫度，提高居住者的睡眠品質。

關鍵字

熱感覺 ； 熱舒適性 ； 機器學習 ； 紅外線熱顯像儀

並列摘要

This paper develops a humanized thermal sensation prediction system, which detects and records human face temperature data combined with machine learning to achieve the goal of predicting human thermal sensation. This paper establishes a test environment, collects facial temperature data for three subjects, uses air-conditioning and heating to change the ambient temperature between 20 and 30 degrees, and subjects will use voting in this environment to express the current situation. Subjects four parts temperature were recorded by FLIR Lepton 3.5 at the moment of voting. The facial temperature data will be split into 80% training set and 20% test set, and the training set will be input into three machine learning methods of support vector regression, random forest, and artificial neural network for model training. The model and PMV make predictions against the data in the test set, and then compare the prediction results of the four models to find the best model.The experimental results show that the prediction results of the three machine learning methods are better than the traditional PMV model, and the support vector regression model is the best model with good accuracy and stability. This paper has successfully developed a humanized thermal sensation prediction system, To achieve good thermal comfort, greatly improve the indoor environment comfort of the office, and improve the office efficiency of office workers. It can also be used to detect changes in the temperature of people's faces when they are sleeping, predict the thermal sensation of people when they are asleep, and automatically adjust the HVAC temperature, to improve the sleep quality of occupants.

並列關鍵字

Thermal sensation ； thermal comfort ； machine learning ； thermal imaging cameras

參考文獻

[1] P.O.Fanger "Study on indoor environmental comfort based on improved PMV index," 1998 3rd International Conference on Computational Intelligence Communication Technology (CICT), 1998.

[2] W. Hu, Y. Wen, K. Guan, G. Jin and K. J. Tseng, "iTCM: Toward Learning-Based Thermal Comfort Modeling via Pervasive Sensing for Smart Buildings," IEEE Internet of Things Journal, vol. 5, no. 5, 2006.

[3] Izhar, W. Xu, X. Wang, R. Wang and Y.-K. Lee, "Low-Cost Micro Predictive Mean Vote (PMV) Sensor and its Application for Smart Personal Ventilation Systems," 2019 IEEE 32nd International Conference on Micro Electro Mechanical Systems (MEMS), 2008.

[4] S. Bin and H. Ke, "Indoor Thermal Comfort PMV Index Prediction Based on Particle Swarm Algorithm and Least Square Support Vector Machine," 2010 International Conference on Intelligent System Design and Engineering Application, 2010.

[5] Q. Zhu, J. Yi, S. Sheng, C. Wen and H. Hu, "A Computer-Aided Modeling and Measurement System for Environmental Thermal Comfort Sensing," IEEE Transactions on Instrumentation and Measurement, vol. 64, no. 2, pp. 478-486, 2015.