Reducing Cost and Time-to-Market on Supporting Driver Assistance Systems to Avoid Rear-end Collisions in Vehicles Traffic

The number of traffic accidents worldwide is a worrisome problem, and its leading causes are related to lack of attention and non-observance of safety distance. Rear-end collisions are the primary type of accident in this context. To mitigate this, we can apply intelligent technologies integrated into vehicles to reduce the probability of an accident. There are research initiatives that propose alternatives for vehicular safety, but the existing solutions are either very expensive or need long-term implementation, so depending on the installation of standardized systems with DSRC (Dedicated Short-range Communications) support. Considering this background, this article proposes BrakeAlert as a driver assistance model designed to avoid rear-end collisions derived from the non-observance of the safety distance of the vehicle immediately ahead. The main contribution resides on providing a framework that takes profit from the own drivers' smartphone, so addressing both the low-cost feature and short-term applicability. The developed prototype was tested in experiments with satisfactory results, obtaining an average response time between a braking event and its notification of 1.13 seconds. This time is acceptable when considering the perception and conductors reaction, showing the effectiveness of the model on reducing the number of collisions in traffic.