Four years of vigorous research has borne fruit for these PhD students of NIT Rourkela and, as a result, the smart city goals of Bhubaneswar might be fulfilled as well. eWARNSYSTEM is the name of their start-up and is the name of the first device developed by K Vinod Kiran and Harshit Srivastava. This IoT-enabled integrated platform is an AI-based surveillance system that comes with an app and with the scope to be a whole lot more.
The idea is to allow the device to tap into your surveillance cameras after which it will use AI-based models to detect any anomalies in the footage in real-time. So instead of a post-incident fix, where the police or security company checks the footage after the crime has occurred, it becomes a preventive measure. "The AI system will send an alert message on our app, which we have also developed, and one can prevent the incident immediately," says Vinod. Currently, the team of 12 is working on increasing accuracy of the device. "The accuracy rate is at 80 per cent currently and we want to bring it between 92 to 100 per cent. The issue is in being able to tell that it is okay for a policeman carrying a gun to enter the house, but not a layman," he says.
Installing new cameras might not even be necessary for eWARNSYSTEM as it is capable enough to tap into the existing system. For now, they are developing the device for retail spaces, like shops and malls where cases of shoplifting can be detected. The device complies with all the guidelines laid down by the Regional Transport Office, Government of India, and a few others.
The team has also been working on other projects, like a pollution monitoring system, traffic density and accident detection system as well. The idea is to equip the device with hybrid configuration so that its various sensors capture data, like live video streams and pollution levels, and then the information is processed, analysed, sent to cloud servers which send it back to the users in a way they would understand.
