(Type) Approval Future and Current Developments INTRODUCTION 4 july 2018 Partner in Mobiliteit
Cora van Nieuwenhuizen, minister of Infrastructure and Water Management at the Intertraffic 2018 Amsterdam: I m going to create legal framework for automated driving. Laying down requirements for reliability and safety that cars must meet before they can hit the road. A driving license for self-driving cars, if you like. Not for the driver but for the car itself! Strong support in de Dutch Government with high ambition! Partner On the in road, Mobiliteit safely and reliably
Partner Safe and in Mobiliteit Trusted on the road Gerard Doll - Director Vehicle, Regulation & Admission RDW
43 companies involved in functioning AV (07-2017) Partner On the in road, Mobiliteit safely and reliably
companies involved in future transportation Partner On the in road, Mobiliteit safely and reliably
How to assess safety? Partner On the in road, Mobiliteit safely and reliably
Partner in Mobiliteit Assessing Safety and Security
RDW Honeycomb of Current Complications System Design Architecture Agile/Fast SW Dev Process Millions of LOC High Verification (Testing) Cost Deep Learning Algorithms SW Quality Deteriorate Self Driving Vehicles Politics (Europe) Connectivity No Control over SW SW OTA Update Slow Regulation Black-hat Hackers Data Misuse and Ownership No clarity on Validation Illegal motor tuning SW Recalls Distributed Dev teams Bug fixes Priority Privacy & Autonomy Use of Commercial OTS Application & software Outdated Type Approval No cooperation in supply chain 8
Finding Input New Regulation BRIDGING THE GAP Learning Audit / Learning Experience - VSSF Vehicle Driving License - VDL Experimenting Law (January 2019) Partner On the in road, Mobiliteit safely and reliably
Overall Structure 11 Point Learning Framework Future Autonomy 9. Deep Learning and Artificial Intelligence 8. Ethics and Human Machine Interface 10. Advanced Perception, Planning & Control 11. Generic Aspects or Future Developments Learning Framework 1. Functional Safety, Cybersecurity & Privacy 2. Software Design & Development 3. Software Config Management Process Engineering Dynamic Operations 7. Software Updates, Patch 6. Data 5. Product Statistics 4. Verification and Validation Product Evaluation
Vehicle Safety and Security Framework (VSSF) Goals Functionality Safety Security Privacy Strategy Process Engineering Product Evaluation Lifecycle Dynamic Operations Development and In-use compliance Future Autonomy Functional Safety, Cybersecurity and Privacy Engineering Software Verification Data Ethics Learning Areas Software Requirements, Design and Development System Validation Software Updates and/or Patch Management Machine and Deep Learning Software Configuration Management Product Statistics Human Machine Interaction Advanced Perception, Planning and Control Quality Software Quality Assurance
Partner in Mobiliteit 12 On the road, safely and reliably
Basic Principles (Sample Question set) Partner On the in road, Mobiliteit safely and reliably 13
Basic Principles (Minimal learning set) Partner On the in road, Mobiliteit safely and reliably 14
Partner in Mobiliteit On the road, safely and reliably VSSF Next steps
VSSF Process Overview Assessment by competent authority Learning Journey Information Analysis Information Gathering Analysis Learning Framework Product Approval maintenance Process Approval maintenance Initial / Full Product Assessment Process Assessment Type Approval Requirements Development Best Practice Standards Regulations Program Development New program Adapt/ modify TA
Finding Input New Regulation VEHICLE DRIVING LICENCE What if technology takes over all driving tasks? Partner On the in road, Mobiliteit safely and reliably
Partner in Mobiliteit On the road, safely and reliably
Background: Finding Input New Regulation Dutch Vehicle Authority (RDW) has no knowledge of driving behavior; how to license self driving vehicles? RDW challenged Green Dino to develop a license for AI-drivers Green Dino started robottuner a new company for assessment and training of robot / AI-drivers. The challenge resulted in the Digital Driving License Project. A collaboration of stakeholders who want to attribute to an international standard for licensing of intelligent vehicle operating systems, human and AI. Initial Group Nvidia, AON Risk Solutions, Ricardo, HAN-Automotive Research, 2getthere, Roborace and initiators robottuner and RDW. robottuner Partner On the in road, Mobiliteit safely and reliably
A driving license for an automated vehicle NEW WORK ITEM PROPOSAL (ISO) Safety of Intelligent Vehicle Operating Systems (SIVOS) Proposed process of testing: Partner On the in road, Mobiliteit safely and reliably
VDL Next steps Partner in Mobiliteit On the road, safely and reliably
Next steps We start with pilot projects in the Netherlands. In these projects we develop and test the digital driving license methodology. -2018 Ommelander Hospital, NAVYA in cooperation with the Province of Groningen -2019 Rivium 2.0, 2getthere Partner On the in road, Mobiliteit safely and reliably
Partner in Mobiliteit On the road, safely and reliably