An Intelligent Overtaking Assistant System for Autonomous Vehicles

Abstract

Autonomously overtaking cars is one of the most challenging functionality when developing self-driving vehicles due to its safe crucial nature. This study presents a novel intelligent overtaking assistant based on fuzzy logic and its deployment into TORCS environment. For achieving this goal, the developed intelligent overtaking assistant decides when to overtake, and calculates the desired position and speed values to accomplish the overtaking action by using the relative speed and position information. Thus, to estimate the positions and speed of the other vehicles, we designed and deployed a multi object-tracking algorithm. Intelligent overtaking assistant consists of four fuzzy logic systems that have modular cascade structure. The fuzzy logic based overtaking assistant decides “when to start or end the overtaking action” and “whether speed adjustment is needed or not”, and calculates “the desired lateral position on the road” and “how much speed increase is needed during overtaking action”. The efficiency of the intelligent overtaking assistant is examined in TORCS environment and the presented results clearly demonstrate that its performance is highly efficient in various road structures.