Building Secure and Scalable dApps: Challenges and Best Practices (Blockchain-based Apps)
4 min. reading
Fatigue behind the wheel is a serious problem that can be dangerous for both the driver and those around them. Several signs may indicate that a driver should take a break:
The National Highway Safety Administration in the United States estimates that about 20% of fatal accidents are related to driver fatigue (about 100,000 accidents in which about a thousand people die each year). There have also been studies in other countries with similar findings.
Especially at risk are truck drivers, because after four hours spent behind the wheel without a break – the reaction speed of the driver drops by half, after eight hours – by six times.
Some countries include fatigue at the wheel as a factor in police reports, but this may still underestimate the actual number of fatigue-related incidents, as it often remains unaddressed or overlooked in investigations.
To reduce this sad statistic, many leading automakers started developing their driver fatigue tracking systems back in the late 90s. The pioneer in this area was the IT department of Volvo, which presented in the late 90s a simple by today’s standards system for tracking the position of the steering wheel relative to the road markings.
Later, Mercedes introduced a new version of the system with a large number of parameters, including recording “ideal driving” comparing it with the current one, and calculating the difference.
There are several technologies and systems that automakers are implementing to determine driver fatigue:
At IT-Dimension, we have introduced an efficient and autonomous system that, unlike many products from automakers, is not tied to the specific equipment of certain premium cars manufacturers. The setup uses cameras and sensors to track driver behavior, including the analysis of specific points on the face and their position.
The system is based on the use of the Dlib library, which helps to collect the facial landmark points – a 68-point set. The app detects:
Therefore, if the system notices abnormal behavior or signs of fatigue, it can alert the driver with a sharp “beep”. Threshold values for aspect ratio were taken from researched publications on driver drowsiness. Empirics can be configured:
We also have another version of such a system. It is based on the use of convolutional neural network (CNN), where only the eyes state is detected. The audible signal sounds according to the same algorithm as in the 1st app.
Contact us if you are interested in the development of a similar solution or if you need consultation. Our team has extensive experience working with monitoring systems, neural networks, and solutions for security or transportation in general. Let’s build safer and more efficient environment together!