Deep Learning-Based Drowsiness Detection System Using IoT

The proposed system is based on the Internet of Things (IoT). We proposed a Drowsiness detection system with Deep Learning using the internet of things. The system's goal is to prevent vehicle accidents caused by drowsy drivers. Millions of people have lost their lives globally as a result of drowsy driving incidents involving fast motorized vehicles. Research by the Central Road Research Institute (CRRI) found that almost 40% of all road accidents are triggered by drivers who are asleep at the wheel. The study's goal is to create a smart warning system that can recognize and discourage driving while fatigued. There were numerous approaches to resolving this problem. Here are a few examples: Support Vector Machine and Histogram of Gradients are used to create a system that does not require a big amount of data and is very sensitive when a picture is rotated. An application is created with the help of sensors; however, these sensors are only effective in certain environments. The surroundings are also working efficiently by counting the blinks of the eyes. The web camera is also inefficient because the camera's bandwidth varies from location to location owing to intensity. Now, a solution is developed that takes into account all of the disadvantages of the Raspberry Pi camera while still being efficient and portable. The Raspberry Pi board functions as a mini-computer that can be carried around. Using the OpenCV algorithm and OpenCV function, the Raspberry Pi camera translates the driver's video stream to a frame and grayscale image. The eye pairs take into account face landmarks, and the Eye Aspect Ratio is derived using Euclidean distances between the locations of the eyes. When the driver's condition is determined to be drowsy, an alert alarm will play, causing the driver to wake up. In 2022, 1,55,622 people died in road accidents in India. According to the National Safety Council, drowsy driving causes around 100,000 accidents, 71,000 injuries, and 1,550 fatalities each year (NSC).