Artificial Intelligence for Vehicle Behavior Anticipation: Hybrid Approach Based on Maneuver Classification and Trajectory Prediction

Abdelmoudjib Benterki,Moussa Boukhnifer,Vincent Judalet,Choubeila Maaoui

doi:10.1109/access.2020.2982170

Abstract

Innovative technologies and naturalistic driving data sources provide a great potential to develop reliable autonomous driving systems. Understanding the behaviors of surrounding vehicles is essential for improving safety and mobility of autonomous vehicles. Onboard sensors like Radar, Lidar and Camera are able to track surrounding vehicles motion and to get different features like position, velocity and yaw. This paper proposes a hybrid approach to integrate maneuver classification using neural networks and trajectory prediction using Long Short-term Memory (LSTM) networks to get the future positions of adjacent vehicles. In this study we use the Next Generation Simulation (NGSIM) public dataset that provides a real driving data. The proposed approach is validated experimentally using VEDECOM demonstrator data. The results demonstrate that the proposed approach is able to predict driver intention to change lanes on average 2.2 seconds in advance. The Root Mean Square (RMS) errors of lateral and longitudinal positions are 0.30 m and 3.1 m respectively. The results demonstrate a high performance compared to various existing methods.

Highlights

The technologies used in intelligent transport systems, especially in autonomous vehicles, are today at the heart of the research and innovation activities of many research teams
SYSTEM OVERVIEW This work aims at developing a framework for trajectory prediction based on Artificial Neural Network (ANN) and deep Long Short-term Memory (LSTM) Recurrent Neural Networks from an autonomous vehicle (Fig. 2)
EXPERIMENTAL SETUP In this part, we present the architectures of the ANN and LSTM models used in this paper

Summary

Introduction

The technologies used in intelligent transport systems, especially in autonomous vehicles, are today at the heart of the research and innovation activities of many research teams. Perception, sensor data fusion, motion planning and control are the main technical challenges to ensure secure and safe autonomous driving. Motion planning uses sensors data fusion such as the location of obstacles, road signs and marking to. Sensors reaction and data fusion time constitute the processing delay in the autonomous system. In order to anticipate the motion of vehicles and increase the level of safety, many solutions are proposed to predict the intention of target vehicles. To address this issue, many studies have attempted to incorporate different models to identify vehicle future motion.

Objectives

Methods

Results

Conclusion