Inferring the Driver's Lane Change Intention through LiDAR-Based Environment Analysis Using Convolutional Neural Networks.

Alberto Díaz-Álvarez,Miguel Clavijo,Felipe Jiménez,Francisco Serradilla

doi:10.3390/s21020475

Alberto Díaz-Álvarez, Miguel Clavijo + Show 2 more

Open Access

https://doi.org/10.3390/s21020475

Copy DOI

Abstract

Most of the tactic manoeuvres during driving require a certain understanding of the surrounding environment from which to devise our future behaviour. In this paper, a Convolutional Neural Network (CNN) approach is used to model the lane change behaviour to identify when a driver is going to perform this manoeuvre. To that end, a slightly modified CNN architecture adapted to both spatial (i.e., surrounding environment) and non-spatial (i.e., rest of variables such as relative speed to the front vehicle) input variables. Anticipating a driver’s lane change intention means it is possible to use this information as a new source of data in wide range of different scenarios. One example of such scenarios might be the decision making process support for human drivers through Advanced Driver Assistance Systems (ADAS) fed with the data of the surrounding cars in an inter-vehicular network. Another example might even be its use in autonomous vehicles by using the data of a specific driver profile to make automated driving more human-like. Several CNN architectures have been tested on a simulation environment to assess their performance. Results show that the selected architecture provides a higher degree of accuracy than random guessing (i.e., assigning a class randomly for each observation in the data set), and it can capture subtle differences in behaviour between different driving profiles.

Highlights

Identifying a driver’s lane change intention (i.e., “when” as opposed to lane change execution, which deals with “how”), is of special interest in the area of Intelligent Transportation Systems (ITS) from the safety and prevention standpoint
The name of the architectures is coded as follows: cF-W-H is a convolutional layer composed of F filters of width W and height H, whereas dN corresponds to a dense layer of N neurons
It should be mentioned that in the evaluation of misclassified cases, some of the wrong classifications occurred between correct Left-Right classifications (e.g., Left-Left-Right-Left during a left lane change), suggesting that the model could be significantly improved by providing feedback on its previous output

Summary

Introduction

Identifying a driver’s lane change intention (i.e., “when” as opposed to lane change execution, which deals with “how”), is of special interest in the area of Intelligent Transportation Systems (ITS) from the safety and prevention standpoint. It allows us to forecast the future status of the modelled object which offers the opportunity to anticipate possible scenarios and be prepared. The possibility of modelling a driver’s lane change intention profile opens several possibilities for the development of advanced systems (both hardware and software) that make use of the data these models infer. One example is the use of the inferred information in a Vehicle-to-Everything (V2X). If these models’ predictions attain enough confidence, the ego vehicle intentions can be predicted and disseminated along an inter-vehicular network, in which ADAS devices would use that information to improve their outputs (e.g., to offer information in embedded devices or to take decisions before risk situations)

Objectives

Methods

Discussion

Conclusion