MOTOR: A Multimodal Dataset for Two-Wheeler Rider Behavior Understanding

2026-05-21 · Source: Computer Vision and Pattern Recognition · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Expert, quick

Summary

The MOtorized TwO-wheeler Rider (MOTOR) dataset is introduced as the first large-scale, multi-view, multimodal resource specifically for two-wheeler rider behavior in dense, unstructured traffic, addressing a critical research gap in the Global South. Comprising 1,629 sequences and over 25 hours of video data from 16 riders, MOTOR integrates synchronized front, rear, and helmet videos, rider eye-gaze, on-road audio, and telemetry including GPS, accelerometer, and gyroscope data. The dataset features rich annotations covering traffic context, rider state, 12 distinct riding maneuvers, and legality labels (Legal, Illegal, Unspecified). Initial benchmarking with CNN and Transformer-based video action recognition backbones, enhanced by multimodal fusion, demonstrates that combining RGB, gaze, and telemetry consistently achieves superior performance. This dataset provides a unique foundation for advancing safety-critical understanding of two-wheeler riding and serves as a benchmark for developing models in behavior analysis, legality-aware prediction, and intelligent transportation systems.

Key takeaway

For Computer Vision Engineers and Research Scientists developing Advanced Driver Assistance Systems for two-wheelers, the MOTOR dataset offers an unprecedented resource. You should utilize its multimodal data, including RGB, gaze, and telemetry, to train and evaluate models for rider behavior recognition and legality classification. This enables more accurate and safety-critical predictions in dense, unstructured traffic environments, accelerating progress in a historically underserved domain.

Key insights

The MOTOR dataset enables advanced two-wheeler rider behavior analysis through its multimodal, multi-view data and comprehensive annotations.

Principles

• Multimodal data fusion improves behavior recognition.
• Eye-gaze and telemetry are crucial for rider state.
• Unstructured traffic requires diverse behavior modeling.

Method

The MOTOR dataset collection involved 16 riders, capturing synchronized front, rear, helmet videos, eye-gaze, audio, and telemetry. Annotations include traffic context, rider state, 12 maneuvers, and legality labels.

In practice

• Develop models for two-wheeler ADAS.
• Evaluate legality-aware prediction algorithms.
• Analyze rider behavior in dense traffic.

Topics

• MOTOR Dataset
• Two-Wheeler Safety
• Rider Behavior Analysis
• Multimodal Data Fusion
• Advanced Driver Assistance Systems
• Video Action Recognition

Best for: AI Scientist, Computer Vision Engineer, Research Scientist

Related on AIssential

• Multi-View In-Cabin Monitoring System for Public Transport Vehicles — A new multi-modal dataset and tools enable advanced in-cabin monitoring for public transport, supporting 3D perception model development.
• DUALVISION: RGB-Infrared Multimodal Large Language Models for Robust Visual Reasoning — Integrating infrared and RGB data via localized cross-attention significantly boosts MLLM robustness in degraded visual conditions.
• Video2GUI: Synthesizing Large-Scale Interaction Trajectories for Generalized GUI Agent Pretraining — Automated extraction of GUI interaction data from unlabeled videos significantly improves GUI agent pre-training.
• NERVE: A Neuromorphic Vision and Radar Ensemble for Multi-Sensor Fusion Research — NERVE is a multi-sensor dataset bridging neuromorphic vision and radar for robust perception research.
• ChartNet: A Million-Scale, High-Quality Multimodal Dataset for Robust Chart Understanding — Code-guided synthetic data generation significantly improves VLM chart understanding and reasoning capabilities.
• OR-Action: Multi-Role Video Understanding with Fine-Grained Actions — Fine-grained OR action understanding requires explicit temporal modeling and multi-view feature alignment beyond scene graphs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computer Vision and Pattern Recognition.