Planning the Best Path with LSTM Based on the PM2.5 Sensors

[[abstract]]空氣汙染問題漸漸受到民眾關注，民眾開始注重今天空氣品質是否適合出門。隨著物聯網(Internet of Things,IoT)技術快速發展，生活中大量地出現物聯網相關的應用。本研究開發以物聯網技術建構以空氣品質感測器連接手機應用程式的應用，主要分為兩個步驟，第一個步驟為建立預測模型，本研究使用LSTM(Long Short Term Memory)方法建立預測模型，為了建立適合的預測模型，設計了六組實驗，且每組實驗有五種不同的時間序列，探討環保署監測站資料與微型監測站資料與時間序列的改變在預測模型上的表現。第二個步驟為開發系統，以Android為作業系統開發手機應用程式，使用者選擇起點與終點，系統將感測器所測量到的值帶入預測模型，算出範圍內的空氣品質並透過Google Direction API規畫最佳路徑給使用者做為參考。[[abstract]]Air pollution issue are gradually receiving people's attention, people began to pay attention to whether today's air quality is suitable for going outside. With the rapid development of the Internet of Things(IoT) technology, a large of Internet related applications have appeared in life. This research develops the application of Internet of Things based technology to connect mobile phone with air quality sensors. It is mainly divided into two steps. The first step is to establish a predictive model. This study uses LSTM (Long Short Term Memory) method to establish predictions. Models, in order to establish a suitable predictive model, six sets of experiments were designed, and each set of experiments had five different time series to explore the performance of the EPA monitoring station data and micro-monitoring data and time series changes on the predictive model. The second step is to develop a system to develop a mobile application for Android. The user selects the start point and the end point. The system takes the value measured by the sensor into the prediction model, calculates the air quality within the range and passes Google Direction API maps produce the best path for the user to refer.[[note]]碩