INTENT: An LSTM Framework for Vehicle Intention Prediction in Intersection Scenarios with Comprehensive Ablation Analysis

Jul 16, 2026

Upload by Aurelia

Vehicle intention prediction is essential for autonomous driving safety and agility, particularly in complex intersection interactions such as roundabouts, sudden stops, and emergency maneuvers where human-like interpretation is required. The INTENT framework introduces an LSTM-based model to forecast vehicle intention 2 seconds before event occurrence, classifying straight, left turn, and right turn behaviors. Experiments and comprehensive ablation studies on the InD dataset achieve 99.71% accuracy.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages6

Reading Time15 min

Introduction and Related Work

INTENT Pipeline Implementation

INTENT: An LSTM Framework for Vehicle Intention Prediction in Intersection Scenarios with Comprehensive Ablation Analysis Logine M. Zaki 1 ,2 and Catherine M. Elias 1 ,2 , arXiv :2607 .083 16v 1 [ cs .AI] 9 Jul 2026 Abstract—Vehicle intention prediction is a pivotal aspect in the agility and safety of autonomous vehicles in all driving scenarios; if genuine enhancement of autonomous vehicles are required, we need to make them adopt human interpretation of driver’s intention especially in cases that require a lot of human interaction as well as complex driving behaviors like the ones at intersections, roundabouts and emergency cases such as sudden stops where vehicle intention prediction helps in taking the correct evasive action within a real time period where every second of action makes an impact and can prevent a catastrophe from taking place. In the worst case, it helps minimize the damage and make safety a priority. Intention prediction can also be used to enhance trajectory prediction (intention conditioned trajectory prediction). In this study, The INTENT framework is proposed using LSTM model to predict the vehicle’s intention at intersections 2 seconds ahead of the event occurrence to predict whether the cars in intersections are going straight, turning left, or turning right. Various model experiments and ablation study are thoroughly tested on InD dataset achieving 99.71% accuracy. Index Terms—Autonomous Vehicles, Intelligent vehicles, Intention prediction, Intersections, Behaviour prediction, LSTM, deep learning models, dataset labeling. I. INTRODUCTION AND RELATED WORK In continuous efforts of enhancing road safety to reduce accidents , autonomous vehicles could be a leading participant if it is able to avoid main causes of accidents like human error, having a misleading intention , taking the wrong action , unexpected lane changes, vague driving behaviors at intersections and sometimes not seeing vehicles if they are occluded.Understanding the scene and even predicting the intention of the target vehicle could alert the driver if the driven car is not fully autonomous or even better, if it is fully autonomous, would take the safest evasive action to ensure optimal safety. Intention prediction is vital for all road users for example: pedestrians,cyclists and vehicles. For each road user, there are even various traffic and road scenarios like pedestrian intention prediction at intersectionsand highways .In this paper ,vehicle intention prediction is going to be our main focus. Car accidents at intersections suffer fatality and severity for few reasons like:vehicles could be at high speed, the collision angle could be deadly ( 90 degrees ) ; imagine collision of 2 cars one that is crossing from South to North and another one from East to West) *This work was not supported by any organization 1C-DRiVeS Lab: Cognitive Driving Research in Vehicular Systems, Cairo, Egypt [cdrives.researchlab@gmail.com](cdrives.researchlab@gmail.com) [2](2 Computer Science and Engineering Department - Faculty of Media)[ Computer Science and Engineering Department - Faculty of Media](2 Computer Science and Engineering Department - Faculty of Media)[ ](2 Computer Science and Engineering Department - Faculty of Media)Engineering and Technology -German University in Cairo, Egypt [logine.elkelani@student.guc.edu.eg](logine.elkelani@student.guc.edu.eg) , [catherine.elias@ieee.org](catherine.elias@ieee.org) which reinforces the vitality of addressing the challenge of vehicle intention prediction at intersections which could exceedingly improve road safety and accelerate the development process of autonomous vehicle. This is why vehicle intention prediction at intersections is our paper motivation. Vehicle intention prediction is a dominant factor in the agility and safety of autonomous vehicles .Intention prediction is a classification problem thus diverse classifiers throughout the years have been exploited to carry out

FAQs

What problem does the INTENT framework address?

It targets vehicle intention prediction at intersections to support timely and safe decisions for autonomous vehicles, reducing risks caused by misleading or unexpected driving behaviors.

How does INTENT predict intention and what are the output classes?

INTENT uses an LSTM model to predict vehicle intention 2 seconds ahead of the event, classifying whether the vehicle goes straight, turns left, or turns right.

What dataset and performance results are reported?

Model experiments and ablation studies are conducted on the InD dataset, reaching 99.71% accuracy.

Jul 16, 2026

Upload by Aurelia