DriveDNA: A Large-Scale Multimodal Naturalistic Driving Dataset and Benchmark for Driving Style Identification

Hugging Face Daily Papers Papers
driving-style dataset benchmark multimodal naturalistic-driving few-shot time-series

Summary

Introduces DriveDNA, a large-scale multimodal naturalistic driving dataset with 4,121 drives from 465 drivers across 115 vehicle models, and a benchmark for driving style identification through tasks like few-shot driver re-identification and personalized behavior prediction.

Driving style captures stable, driver-specific patterns in how a vehicle is driven. In naturalistic data, however, this signal is hard to isolate because drivers are observed in different vehicles, on different roads, and under different conditions, so models may mistake vehicle- or situation-specific regularities for driver-specific style. We introduce DriveDNA, a large-scale naturalistic dataset and benchmark for personalized driving-style modeling, comprising 4,121 drives from 465 drivers across 115 vehicle models and totaling 975 hours of human-controlled driving at 10 Hz with forward video, collected from community drivers in everyday use. DriveDNA defines driving style as a consistent, driver-specific behavioral pattern in how a vehicle moves under similar conditions. The benchmark evaluates this signal through three core tasks: few-shot driver re-identification, personalized behavior prediction, and condition-matched comparison, and provides behavioral annotations plus 276,248 rule-generated maneuver events across six classes with large-scale human auditing. We evaluate baselines spanning classical descriptors, supervised and self-supervised time-series encoders, multimodal fusion, probabilistic prediction, and zero-shot foundation models under a fixed multi-seed protocol. Learned representations substantially outperform classical descriptors on unseen drivers (AUROC .935 vs. .707) and retain driver-specific information under matched driving conditions, while descriptor performance approaches chance. Video-only models achieve comparable re-identification accuracy but exhibit severe route leakage, showing that strong recognition may arise from contextual shortcuts rather than driving behavior. These findings show that reliable driving-style evaluation must assess both the behavioral value of learned representations and their robustness to vehicle, drive, and condition confounds.
Original Article
View Cached Full Text

Cached at: 07/28/26, 06:34 AM

Paper page - DriveDNA: A Large-Scale Multimodal Naturalistic Driving Dataset and Benchmark for Driving Style Identification

Source: https://huggingface.co/papers/2607.23822 Published on Jul 26

·

Submitted byhttps://huggingface.co/HenryYHW

Henryon Jul 28

Abstract

Drivingstylecapturesstable,driver-specificpatternsinhowavehicleisdriven.Innaturalisticdata,however,thissignalishardtoisolatebecausedriversareobservedindifferentvehicles,ondifferentroads,andunderdifferentconditions,somodelsmaymistakevehicle-orsituation-specificregularitiesfordriver-specificstyle.WeintroduceDriveDNA,alarge-scalenaturalisticdatasetandbenchmarkforpersonalizeddriving-stylemodeling,comprising4,121drivesfrom465driversacross115vehiclemodelsandtotaling975hoursofhuman-controlleddrivingat10Hzwithforwardvideo,collectedfromcommunitydriversineverydayuse.DriveDNAdefinesdrivingstyleasaconsistent,driver-specificbehavioralpatterninhowavehiclemovesundersimilarconditions.Thebenchmarkevaluatesthissignalthroughthreecoretasks:few-shotdriverre-identification,personalizedbehaviorprediction,andcondition-matchedcomparison,andprovidesbehavioralannotationsplus276,248rule-generatedmaneuvereventsacrosssixclasseswithlarge-scalehumanauditing.Weevaluatebaselinesspanningclassicaldescriptors,supervisedandself-supervisedtime-seriesencoders,multimodalfusion,probabilisticprediction,andzero-shotfoundationmodelsunderafixedmulti-seedprotocol.Learnedrepresentationssubstantiallyoutperformclassicaldescriptorsonunseendrivers(AUROC.935vs..707)andretaindriver-specificinformationundermatcheddrivingconditions,whiledescriptorperformanceapproacheschance.Video-onlymodelsachievecomparablere-identificationaccuracybutexhibitsevererouteleakage,showingthatstrongrecognitionmayarisefromcontextualshortcutsratherthandrivingbehavior.Thesefindingsshowthatreliabledriving-styleevaluationmustassessboththebehavioralvalueoflearnedrepresentationsandtheirrobustnesstovehicle,drive,andconditionconfounds.

View arXiv pageView PDFGitHub0Add to collection

Get this paper in your agent:

hf papers read 2607\.23822

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2607.23822 in a model README.md to link it from this page.

Datasets citing this paper1

#### HenryYHW/DriveDNA Updatedabout 4 hours ago • 19

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.23822 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

DriveZero: End-to-End Driving Beyond Human Demonstrations

Hugging Face Daily Papers

DriveZero is an end-to-end autonomous driving system that combines a vision foundation model for perception with reinforcement learning for action, achieving driving behaviors beyond human demonstrations and state-of-the-art performance on benchmarks like nuPlan and NAVSIM.