An early warning model of traffic accidents based on fractal theory

Traffic accidents occur with tremendous increase over the years. We propose an early warning method combine the factors from both the drivers and their vehicles. We extract data of physiological parameters and vehicle parameters and analysis them with fractal theory. We use Hurst index to verify that parameters of driver and vehicle have a characteristic of fractal. The simulation demonstrates that when the average box dimension value of the heart rate volatility, rotational speed volatility, and the acceleration is greater than 1.7, accident will probably occur so that we should warn the driver that he should have a rational decision.