Integrated 4D driver modelling under uncertainty

Risultati finali

Experimental data relevant to model development

Open-access GitHub repository containing data that have been measured in field experiments and have informed the development of the model. With the data, various hypotheses and assumptions are tested before they are integrated into the model. Linked to task 5.2

Experimental data relevant to model validation

After the model development, field experiments are performed to test the models in specific situations and compare i4Driving model predictions with measured driving behaviours. An Open-access GitHub repository will include data that are measured in field experiments and passed to the validation framework. Linked to task 5.3

Harmonized, annotated, and processed data in usable format

Open-access GitHub repository containing converted data from various sources, available in the OpenSCENARIO and the CommonRoad format. Linked to task 1.1.

Specification of the scenario description language, and conversion tool

This deliverable contains: 1) a documentation of a scenario description language specification, specifying the language concept, the language grammar, and the conversion approach to the ASAM OpenX format, 2) a documentation detailing the conversion toolchain development for converting the scenarios described in the scenario description language in point 1 to the ASAM OpenX standards. Linked to task 6.2

Human driver models as baseline for consumer testing campaigns of ADS

A document published in an open access GitHub repository will describe:a) the execution of concrete scenarios using, as ego vehicle, either the ADS and the human driver models in order to compare their behaviours. b) the execution of a set of scenarios using an integrated human driver model + ADS/ADAS unit as ego vehicle to demonstrate that the human driver models can be used for ADS/ADAS testing (not fully automated functions). Linked to task 6.4

Dissemination and exploitation plan update

Update/revised dissemination and exploitation plan, including communication activities). Linked to task 7.1

Critical review of state-of-the-art techniques to model drivers’ heterogeneity

A report published on an open access GitHub repository will detail the state-of-the-art techniques applied in Traffic Flow Theory and Human Factors literatures to model inter- and intra-driver heterogeneity. Linked to task 4.1

Research Data Management Plan 2

Second update of the research data management plan. Linked to task 8.4

Experimental setup for the driving simulator experiments

A report published on open access GitHub repository, describing the experimental design of the driving simulator experiments. This includes description of which use cases will be studied, in which driving simulator, and descriptions of the driving simulator scenarios. Linked to task 3.1

Methods to harmonize data on human driving performance from different datasets

A report uploaded on an open access GitHub repository detailing developed methods to translate data into different formats (for instance, map information of traffic scenes is stored in different formats, such as OpenDRIVE or lanelets). Converters will be able to convert all data into a single format for a streamlines evaluation. This is linked to task 1.1.

Project Quality Handbook

Develop a Project Quality Handbook (including risk assessment plan, ethics and GDPR) and annual quality reviews. This will be developed in M1, revised and updated in M36. Linked to task 8.3

Research Data Management Plan 1

Develop the Data Management Plan by M6 and revise in M18 and M36, which follows the FAIR data principles making project data/research outputs findable, accessible, interoperable and reusable. Linked to task 8.4

Report on dissemination activities, including cooperation with other projects

This report summarises the dissemination activities, stating the details of individual items, partners’ contribution, impact, and how they compare with the original plan. Linked to task 7.2, 7.3, and 7.4

White paper

A documentation describes the project development process within the project, and provides guidelines on how the work can be used in ADS development and policy making. Linked to task 7.5

Final booklet

Develop the final booklet summarising the main project results, outputs and tools in an easy-to-read and concise format. Linked to task 7.5

Causal relationships between human/external factors and human driving behaviors: modelling requirements & framework of testable hypotheses

A report published on an open-source GitHub repository detailing key requirements and testable hypotheses for the modelling framework developed in WP2 geared towards credible and plausible modelling outcomes from an HF perspective. Major categories of testable hypotheses examined are awareness, cognitive workload and tactical driving strategies. Linked to task 1.3

Dissemination and exploitation plan update final

Final update of the dissemination and exploitation plan. Linked to task 7.1

Project glossary

Develop a project glossary which contains a summary of all main acronyms, terms and definitions relevant for the project. This will be developed in M1, reviewed and updated in M24. Linked to task 8.2

i4Driving Framework design & modeling and coding design principles

A document and/or reference published on an open access GitHub repository that describes the conceptual and mathematical principles of the microsimulation software (OTS/Aimsun) used in the project in terms of scope and behavioral principles (key assumptions, abstractions and simplifications); numerical solutions (approximation/solution methods, vehicle & infra/graph representation, time/event-queue handling, etc.) and software design principles (naming and coding conventions, data/project management, version management, etc.). Linked to task 2.1

Methods to extract statistically significant relationships between human/external factors and driver behavioral mechanisms, in uncritical and critical situations

A report uploaded on an open access GitHub repository detailing the selection and use of specific machine learning techniques for identifying the statistically significant relationships between factors and behaviours. Linked to task 1.2.

Implementation framework of the scenario-based evaluation workflow

This deliverable is the documentation of the development of the scenario based evaluation framework, including all the sub-elements of the framework, and how the framework can be implemented and executed in simulation environment. This is linked to task 6.1.

Project Management Handbook

Develop an internal handbook to establish and define clear rules/ procedures for communication, contractual, financial and administrative management aspects. Linked to taska 8.1 and 8.2

Methodology and results: relevant use cases and safety-critical scenarios

This deliverable is to identify the relevant use cases and safety critical scenarios for evaluating the human driver model and the target applications. Using an hazard based testing approach, this deliverable will generate scenarios and identify scenario parameter ranges. They will be in the form of documentation of the scenario generation methodology, and the actual generated scenarios uploaded to the Safety PoolTM scenario database. Linked to task 1.4

Sensitivity Auditing

A technical self-sustaining report uploaded in an open access GitHub repository will describe the principles of sensitivity auditing, how these have been applied to the modelling activities of i4Driving, and relevant findings. Linked to task 2.5

Evaluation criteria and detailed description of field experiments

A document published in an open access GitHub repository describing field experiments, including the definition of key driving performance indicators and how they are measured in the field experiments, the characterization of scenarios, necessary equipment, how many participants are needed to realistically replicate the scenarios, and how many experiments are needed to generate statistically valid results. Linked to task 5.1

Human driver models as baseline for PEARS

Documentation on how the evaluation framework is used for the human driver model (from integrating with the model, to obtaining the scenarios, to establishing the evaluation criteria and executing the scenarios), it also contains the details of the evaluation results. This is linked to task 6.3

Project glossary update

Updated project glossary. Linked to task 8.2

Dissemination and exploitation plan (including communication activities)

This deliverable contains a detailed plan stating the dissemination and exploitation roadmap for the i4Driving project. It includes the responsibilities of each partner, the dissemination channels intend to use, together with a dissemination monitor/tracker. Dissemination and exploitation roadmap will include publications, presentations, inputs to standards and regulations. This will be done by M6 and revised in M18 and M36. Linked to task 7.1

Open-source library of validated probabilistic human driver behavioural models

The library of validated probabilistic human driver behavioural models will be published open source on a GitHub repository. Linked to tasks 4.3 and 4.4.

Open-source library of techniques to encode drivers’ heterogeneity into models

The software employed to estimate distributions and correlation structures of i4Driving model parameters will be published open-source on a GitHub repository. Linked to task 4.2

Validated ethics assessment plans and approval from Ethics committee

A document published in an open access GitHub repository describing the ethics considerations for the driving simulator and real worlds test from the different test facilities. It also includes the approvals from the different national ethic committees. Linked to task 3.1

Website and social networks profiles

Project website and social networks profiles designed and published online. Linked to task 7.2

Software for automatically increasing the criticality of scenarios

Open-source GitHub repository containing a software that will alter the states of vehicles (by using optimization techniques), such that the solution space of the ego vehicle is reduced, thus creating a more critical scenario for that vehicle. The input to the software will be a traffic scene in the CommonRoad format. Linked to task 3.1

Open-source library of data mining techniques

The software employed and adapted for the data mining research will be published on an open-source GitHub repository. Linked to task 1.2

Incremental versions of i4Driving / software library

Open-source GitHub repository containing software versions. A version represents a major new release of the simulation software (e.g. version 1.0 or 2.1) including release notes, and e.g. updated examples, demo’s, documentation and unit tests. Linked to task 2.2

Open-source GitHub evaluation software toolchain

OOpen source the language and conversion toolchain (including identifying the correct license) on Github, the deliverable will be a repository with required documentation. This is linked to task 6.1

Suite of unit tests for model development

A (growing!) catalog / library of annotated testcases (code snippets) designed to identify, locate, and reproduce semantic errors, published in an open-source GitHub repository. This library is a safeguard in preventing re-occuring (resolved) bugs and errors and in making sure new releases/versions are backward compatible. Linked to task 2.1

Validated ethics assessment plan and approval from Ethics committee

A document published in an open access GitHub repository describing the process of unification and validation of the ethics assessment plans from each partner executing experimental work in closed field tests. The document will also describe the comprehensive programme on Research Ethics at TUM (LfE), and the training possibly provided to the partners to support and ensure the familiarization process with conducting the type of experimental work required for this project. Linked to task 5.1

Pubblicazioni

An annotated timeline of sensitivity analysis

Autori: arantola S., F. Ferretti, S. Lo Piano, M. Kozlova, A. Lachi, R. Rosati, A. Puy, P. Roy, G. Vannucci, M. Kuc-Czarnecka and Saltelli, A.
Pubblicato in: Elsevier, in Environmental Modelling & Software 174, Numero Elsevier, in Environmental Modelling & Software 174, 2024, ISSN 1873-6726
Editore: Elsevier
DOI: 10.1016/j.envsoft.2024.105977

What is Post-normal Science? A post-normal encounter

Autori: Saltelli, A.
Pubblicato in: Springer, in Foundations of Science Journal, 2024, ISSN 1572-8471
Editore: Springer
DOI: 10.1007/s10699-023-09932-x

Unpacking uncertainty in the modelling process for energy policy making

Autori: Lo Piano, S., Lőrincz, M. J., Puy, A., Pye, S., Saltelli, A., Smith, S. T., & van der Sluijs, J. P.
Pubblicato in: Wiley Periodicals LLC on behalf of Society for Risk Analysis, 2023, ISSN 1539-6924
Editore: Wiley Periodicals LLC on behalf of Society for Risk Analysis
DOI: 10.1111/risa.14248

The challenge of quantification: an interdisciplinary reading

Autori: Di Fiore, M., Kuc-Czarnecka, M., Lo Piano, S., Puy, A. and Saltelli, A.
Pubblicato in: Minerva, 2023, ISSN 1573-1871
Editore: Minerva
DOI: 10.1007/s11024-022-09481-w

Impact assessment culture in the European Union. Time for something new?

Autori: Saltelli, A., Kuc-Czarnecka, M., Piano, S.L., Lőrincz, M.J., Olczyk, M., Puy, A., Reinert, E., Smith, S.T. and van Der Sluijs, J.P.
Pubblicato in: Elsevier - environmental science & policy, Numero Elsevier, in journal of Environmental Science & Policy, 142, pp.99-111, 2023, ISSN 1873-6416
Editore: Elsevier
DOI: 10.1016/j.envsci.2023.02.005

Discrepancy Measures for Global Sensitivity Analysis

Autori: Puy, Arnald; Roy, Pamphile T.; Saltelli, Andrea
Pubblicato in: Technometrics, 2024, ISSN 0040-1706
Editore: American Statistical Association
DOI: 10.48550/arxiv.2206.13470

Predicting Autonomous Driving Behavior through Human Factor Considerations in Safety-Critical Events

Autori: Jamal Raiyn; Galia Weidl
Pubblicato in: MDPI - Smart Cities journal, Numero MDPI, in Smart Cities journal, 7(1), 460-474, 2024, ISSN 2624-6511
Editore: MDPI
DOI: 10.3390/SMARTCITIES7010018

Risultati finali

Pubblicazioni

Condividi questa pagina Condividi questa pagina sui social network

Scarica Scarica il contenuto della pagina