Naturalistic Driving Data Analytics

The 9th International Workshop on Naturalistic Driving Data Analytics (NDDA)

Sponsored by IEEE ITSS Technical Committee on Data Analytics and Intelligent Systems for Advanced Driving and Mobility (DAISY)

IEEE Intelligent Vehicles Symposium - IV 2022

June 5, 2022 | Aachen, Germany

Theme: Emerging Opportunities in NDDA

Important Dates

Submission Deadline: Mar. 15, 2022

Acceptance Notification: Apr. 22, 2022

Camera-Ready Copy Due: May 1, 2022

Meeting Room

This year's workshop will be held at the Eurogress Aachen Conference Center ( with hybrid format.

For in-person participants: Room K8, Eurogress Aachen

For online participants: Zoom virtual meeting room

Meeting-ID: 916 3533 7255

Code: 334914

Note: The demonstration during the break will only be available to in-person participants. It will not be broadcasted online.


Understanding driving characteristics, driver behaviors, vehicle performance characteristics, traffic environment and modalities in real world driving context are important for the development of future mobility applications for Intelligent Vehicles. Driver engagement with the vehicle operation, driver-vehicle capabilities for handling demanding traffic situations, traffic management protocols and fuel efficiency improvements are some of the key research topics to address using naturalistic driving data.

Naturalistic Driving data collected from various onboard sensors, infrastructure sensors, and other emerging data sources provide a wealth of information pertaining to a snapshot of real-world driving context. However, these data streams are inherently heterogeneous due to multimodal nature of sensor suites and data collection platforms used. Therefore, our intention is to investigate intelligent data analytic approaches to produce meaningful inferences from real-world driving data for the safe deployment of intervening technologies for future mobility applications.


  • Data collection

      • Data needs and specifications for Naturalistic Driving data collection

      • Robust data compression and annotation techniques

      • Driver, driving, and environment data collection platforms including in-vehicle physiology sensing

  • Data analytics

      • Advanced and automated driving systems data analytics

      • Understanding/ interpretation of naturalistic driver behavior and modeling

      • Driver state, collaborative driving algorithms

  • Naturalistic Driving applications

      • Driver-Vehicle Interaction, Driver- ADAS/ADS Interaction

      • Crash risk analysis, reconstruction, modeling, and intervention using naturalistic driving data

      • Integration of naturalistic driving models for scenario generation for the safety evaluation of Automated Driving systems

      • Assessment protocols for driver health and wellness using naturalistic driving data

      • Fuel efficiency

      • Collaborative, connected, and shared traffic management systems


  • Understand driver behaviors in various levels of automation from naturalistic driving data.

  • Analyze driving conditions and driver-vehicle performance for safety applications and future mobility solutions.

  • Explore driving performance metrics for early diagnosis of health issues.

  • Requirement analysis for future applications in automated-connected driving, traffic management, and infrastructure design for future mobility.


Duration: 8:30 AM - 12:30 PM Local Time (German Time)

Prof. Mohan Trivedi,

Chair, Electrical and Computer Engineering

Laboratory for Intelligent and Safe Automobiles (LISA)

University of California San Diego, USA.

Vision and Machine Learning for Safe Autonomous Driving with Naturalistic Studies

In this presentation we emphasize the centrality of novel Naturalistic Driving Studies (NDS) in the advancement of Safe Autonomous Driving. Over the years, naturalistic driving experiments mainly developed by the human factors community, have contributed to improving of vehicle safety, traffic flow efficiency, roadway and infrastructure designs, and advanced driver assistance systems. Rapid technological advancements as well as popularity of vehicles with highly automated operational modes, the earlier NDS approaches need to be revamped to address the demands of new technologies at the core of the growth of highly automated vehicles. These advancements require innovative data-driven algorithmic solutions for capture, analysis, prediction, planning and evaluation of states and behaviors of intelligent agents in a complex and dynamic environment, with assured safe operation. We will discuss how novel naturalistic driving studies support development of curated datasets, metrics and algorithms for situational criticality assessment, in-cabin activity monitoring, driver readiness and take-over time predication.

Prof. Dr.-Ing. Dr. med. Dr. h.c. Steffen Leonhardt Steffen Leonhardt

Chair for Medical Information Technology

Helmholtz Institute for Biomedical Engineering

RWTH Aachen University, Germany.

Unobtrusive Vital Sign Monitoring in Automotive Environments

Human vital signs, i.e. heart/pulse rate, breathing/respiratory rate, body temperature and blood pressure, provide essential information about the individual health status. With regard to an automotive environment, the formerly diagnostic intentions of classical methods for vital signs monitoring offer the opportunity to extend their medical application to a traffic safety-related topic: driver state monitoring. This specific area of automotive applications is dedicated to assessing the driver’s attention or alertness and mental state. Therefore, the detection of stress and drowsiness while driving is of particular interest. For many years, researchers engage in integrating vital sign monitoring techniques to automotive setups for this purpose. However, since classical wired solutions are inappropriate in a driving scenario, there is extensive need for unobtrusive methods implementing monitoring techniques. There exist a variety of different methods to unobtrusively obtain vital parameters, such as the capacitive electrocardiography (cECG), ballistocardiography (BCG), reflective photoplethysmography (rPPG), magnetic induction, radar, ultrasound, or camera-based methods, which comprise RGB, near infrared (NIR), or infrared thermography (IRT) imaging, all of which are currently being considered for a driver state monitoring application. This talk provides an overview about these sensing techniques and discusses challenges and opportunities for future directions of assessing physiological quantities in the automotive environment.

He Zhang

Tongji University, China

Risk Assessment of Highly Automated Vehicles with Naturalistic Driving Data: A Surrogate-based Optimization Method

One essential goal for Highly Automated Vehicles (HAVs) safety test is to assess their risk rate in naturalistic driving environment, and to compare their performance with human drivers. The probability of exposure to risk events is generally low, making the test process extremely time-consuming. To address this, we proposed a surrogate-based method in scenario-based simulation test to expediate the assessment of the risk rate of HAVs. HighD data were used to fit the naturalistic distribution and to estimate the probability of each concrete scenario. Machine learning model-based surrogates were proposed to quickly approximate the test result of each concrete scenario. Considering the different capabilities and domains of various surrogate models, we applied six surrogate models to search for two types of targeted scenarios with different risk levels and rarity levels. We proved that the performances of different surrogate models greatly distinguish from each other when the target scenarios are extremely rare.Inverse Distance Weighted (IDW) was the most efficient surrogate model, which could achieve risk rate assessment with only 2.5% test resources. The required CPU runtime of IDW was 2% of that required by Kriging. The proposed method has great potential in accelerating the risk assessment of HAVs.

Claudia Goldman

General Motors, Israel Technical Center

Trusting Explainable Autonomous Driving: Simulated Studies

Automated behaviors, resulting from deploying AI planning systems in real life applications, can be concerning for their end users. This is due to possible mismatches between what the system considers optimal actions and what the user expects as their subjective optimal action. This problem becomes even more challenging when considering AI decision making algorithms, controlling the autonomous driving behaviors, that are complex and affected by uncertainty in the environment and in the system own sensors. Therefore, to improve human trust in AI deployed planning systems, we need to solve two problems: understanding users’ needs for explanations, and computing those explanations according to the context of the AI system.

This paper presents results from two large user studies in simulated autonomous driving scenarios, assessing users’ trust when provided different explanations. Then, we present a data driven solution to infer probabilistically an explanation that is the most suitable for a driving context and users’ group according to the data analysis and trust measures.

Daniel Omeiza

Oxford Robotics Institute

Dept. of Engineering, Science,

University of Oxford, UK.

From Spoken Thoughts to Automated Driving Commentary: Predicting and Explaining Intelligent Vehicles’ Actions

Commentary driving is a technique in which drivers verbalise their observations, assessments, and intentions. By speaking out their thoughts, both learning and expert drivers are able to create better understanding and awareness of their surroundings. In the intelligent vehicles’ context, automated driving commentary can provide intelligible explanations about driving actions, and thereby assist a driver or an end-user during driving operations in challenging and safety-critical scenarios. In this paper, we conducted a field study in which we deployed a research vehicle in an urban environment.

While collecting sensor data of the vehicle’s surroundings, we also recorded a driving instructor using the think-aloud methodology to verbalize their thoughts while driving. We analyzed the collected data to uncover necessary requirements for effective explainability in intelligent vehicles. We show how intelligible natural language explanations that fulfil some of the key elicited requirements can be automatically generated based on observed driving data using a simple tree-based approach. Finally, we discuss how our approach can be built on in the future to realize more robust and effective explainability for driver assistance as well as partial and conditional automation of driving functions.

Demonstration: 4xU Sensor – Monitoring the driver’s vital signs

Steffen Leonhardt, Marian Walter, Onno Linschmann, Markus Lueken

In this demonstration, we will present our recently developed version of the 4xU sensor, which incorporates four different modalities of unobtrusive sensing technologies: capacitive electrocardiography (cECG), reflective photoplethysmography (rPPG), seismocardiography (SCG) and magnetic induction measurement (MIM). The sensor setup was integrated in a seat cover to provide a flexible measurement basis. It will be presented in a car-driving scenario and participants will have the opportunity to test the system.


Pujitha Gunaratne, Ph.D.

Principal ScientistToyota Motor North AmericaAnn Arbor, MI, USA

Steffen Leonhardt, Dipl-Ing., Dr.-Ing., Dr. Med.

Chair, Medical Information Technology, Helmholtz-Institute for Biomedical Engineering, RWTH Aachen University, Aachen

Udara Manawadu, Ph.D.

Senior Research EngineerWoven Planet Holdings Inc, Tokyo, Japan