3. Process Description 3.1 Temperature sensor collecting in-car temperature every 10 seconds and send it to EC2 server.
3.2 Microphone recording in-car sound every 10 seconds, generate a 5- second sound clip. Then, Raspberry Pi performs feature extraction, send it to AWS machine learning model and get the ML result. Next, the result will be sent to the EC2 Server.
3.3 If it is a baby crying sound (we simulate it by playing baby crying sound we recorded before), then the ML module outcome is "1" (represents there is a baby crying in the car). When EC2 server receives this result, it will send the notification containing the danger and current temperature to the APP.
3.4 Alert message with current temperature received in the APP
3.5 Click the "confirm" button and confirmed message shown on the APP
3.5 Fan Opened after receiving the command sent by the APP (Temperature now is 16.38, below 20℃)
3.6 Fan opened at 28℃,greater than 20℃
3.7 To double guarantee that users can receive alarm message even though the internet signal is not well, we use AWS SNS, it sends both email and SMS