In many real life applications it is very crucial to monitor the temperature and ensure that its maintained within the optimum range. Many alert systems generate and share an alert only when the temperature has crossed the set limit. By that time, there is very little time left for any corrective measure and eventually it leads to damage of goods, money, effort and time.
Therefore in addition to monitoring the temperature real time, It is very essential to build some intelligence within the system, so that any anomaly can be detected immediately and
pre-configured alerts to be sent to the stack-holders. Thus, early detection provides maximum time for any corrective measures to take place.
TempGuard is a system with built-in artificial intelligence to monitor real-time temperature and detect any anomalous trend and share immediate alert to the pre-configured stack-holders.
Making hardware connection
Capture Temperature Data.
- The system has been built using the BoltIot platform. The Bolt wifi module is connected with the LM35 temperature sensor which in turn sends the data to the Bolt Cloud.
Sharing the captured data to the system.
- To build the anomaly detection system we are using DigitalOcean cloud. The linux machine in digital ocean is installed with all necessary os and software components such as python, Bolt library, json, math, statistics libraries.
The alert sending mechanism
- The system sends alert via sms. for this I am using Twilio trial account. Just register on Twilio. Get your SSID and Auth token and its done. Pass on the the details as given below: sms = Sms(conf.SSID, conf.AUTH_TOKEN, conf.TO_NUMBER, conf.FROM_NUMBER) and its done.
Anomaly Detection
- The system considers 10 recent readings of temperature as sent by the LM35 sensor via the bolt module using Z score method. Z score method gives an outer and inner bound and every time the temperature goes beyond the boundaries, the system generates the alert and share it through any of the configured platform. in this case the alert was sent via sms, however, any other platform such as telegram, email can also be easily configured. Please refer to the code file attached with the project to know how the z-score method has been implemented
Comments