Path Estimation Strategies for Mid Vehicle Collision Detection and Avoidance

Abstract

The Major cause of mid vehicle collision is due to the distraction of drivers in the newlineFront and rear-end vehicle witnessed in dense traffic and high speed road conditions. newlineTo sidestep such imminent collision, the conventional path strategy is fashioned in this thesis. This research work focuses on an adaptive path estimation scheme for Mid vehicle Collision Detection and Avoidance System (MCDAS) to avoid such imminent newlinecollision. Consequently, four adaptive path estimation algorithms coined for MCDAS newlineis introduced in this thesis for evading the possible collision at both ends of the mid newline(host) vehicle. The adaptive path trajectories involve two motions, namely longitudinal and lateral (curvilinear) motions for intelligent mid vehicle traction. The longitudinal path estimation is preferred to maintain the distance between the front and rear vehicles. newlineHence, adaptive (offset-based) curvilinear path estimation is preferred where longitudinal motion appears to collide with preceding vehicles. The novel offset-based curvilinear motion is adopted in this proposed method for intelligent mid vehicle trajectory. newlineTherefore, threat-free motion is adopted for intelligent transportation system, namely newlineMCDAS at diverse road conditions in real-time scenarios. Extending such curvilinear newlinemotion with possible reverse direction is adopted for parallel parking application. newlineThese adaptive path trajectories for MCDAS are modelled using Crisp (Actual), Fuzzy, newlineMARS and Fuzzy Regression models. Accordingly, the proposed models are evaluated using the real trajectory from Next Generation Simulation (NGSIM) Interstate 80 (I- 80) dataset. The MSE analysis investigates the efficiency of the proposed models for newlineMCDAS. newline