Artificial Neural Network Model Analysis Based on the Rules that unless Overtaking otherwise Driving on the Right

Abstract

In this article, I analyzing the real problem and abstract the mathematical model to save it step by step. First I consider the mathematical model underlying one lane, in the hypothesis of even flow and single kind, I analyze the relationships between flow, velocity and density; I figure out the linear relationship of velocity and density, including both high-density and low-density conditions; the parabolic relationship flow and density follow so that I gain the best density when flow gets its maximum. Then I extend it to two-lane ( slow lane and fast lane that is equal to the overtaking lane and normal lane ) model: I expand and modify the above relationship in changing system to include the problem of road congestion, as well as the changing lane condition and the density affection from its downstream neighbor . Besides, three scenarios were discussed to control the situation of changing lanes. After this, I use a safety multi-lane highway overtaking control model based on BP artificial neural network to reduce the complexity of problem solving, and discuss the algorithm of the BP artificial neural network, and give examples and preliminary results. Short circuit in a multi-vehicle lane driving, driving behavior mainly consists of acceleration, deceleration and lane changing , the reality of traffic environment, the driver's decision-making are affected by many factors, such as driving rules , vehicle types and other basic physical attributes , characters and vehicle drivers’ plans , driving behavior control strategies and so on. I focus on the main factors to form of some the rules of changing lanes.