Final Project Results: Deliverable D1.5

First concrete ideas of automated driving were not presented until in the 1960’s. More realistic steps were taken under the EUREKA Prometheus Project from 1986 to 1994 when experiments on technologies for automated driving were carried out intensively – however, the technologies were not yet mature for automation at that time. More recently, the US DARPA Grand Challenge prize competition for American autonomous vehicles, Waymo robotaxis and Tesla have created hype and raised a lot of expectations for an early launch of driving automation.
The Hi-Drive main objective was to take intelligent vehicles technology towards high automation in scenarios neither extensively tested nor demonstrated earlier in European and overseas traffic. The project demonstrated robust High automation functions and interaction with other traffic participants enabling the industry to proceed towards market deployment of these systems in the early 2030.
The consortium, coordinated by VW, was large with 53 partners and comprised the necessary value chain to bring automated vehicles to the deployment phase. The consortium comprises of OEMs, suppliers and engineering companies responsible for enabling technologies and sub-systems to be integrated into vehicles. Moreover, research, consulting and academic partners supported the work in numerous specific questions related to enabler functionality, HMI, cyber-security, further development, testing and impacts on the user and societal levels.
The project defined numerous research questions for the work. The technical research questions were related to defining and adapting technology enablers to perception, decision making and actuation. The aim of enablers was to enable a robust and resilient automated driving functions (ADF) perception-decision-action. The challenges for the technology enablers were related to vehicle automation performance in extended and variable ODDs. The technology enablers were integrated into vehicles for automated functions. Technical performance and effects on users were tested in different traffic conditions. The collected data were evaluated to draw conclusions from the impacts of high automated driving. The work detailed each technology enabler thematic area, connectivity, high precision positioning, V2X cybersecurity and machine learning, the implementation of the first release of the enablers, the technical research questions of the technological solutions and the related KPIs and logging requirements.
Hi-Drive methodology was built on both the enablers defined and the previous work of the projects testing ADAS and SAE L3 automated driving functions in road conditions. The work updated the existing methodology in view of higher levels automation with special attention to (i) extension of ODD (ii) user aspects and (iii) transport system level impacts. The project listed and defined six high-level research areas covering the project goals. For each research area, research questions were derived and starting from high-level questions related to separate research topics to low-level questions that can be linked to single studies. For the automated driving functions design and tests the goal was to identify the use cases (UC) and scenarios which combine and allow to demonstrate the vehicles’ functions, the enablers’ impact and driver behaviour. The test scenarios defined showed how the advanced AD functions can extend the ODD compared to the state-of-the-art. To achieve this, a set of use cases for testing ODD including urban, motorway, motorway-to-urban transition, cross-border, and parking driving scenarios were defined. The experimental procedure was an essential part of the methodology. It ensured that the data for the evaluation were collected in the operations by observing the experimental design and procedures. By test design and adhering to it, it was possible draw causal conclusions from the impacts of the automated functions. The project collected both objective and subjective data in variable conditions to find out the impacts and functionality of automated systems approaching high automation on SAE L4. The aim was to understand first user interaction with the system, user opinions and preferences and then to demonstrate the extended ODD compared to SAE L3 Conditional automation.
The project tested various automated driving functions in diverse scenarios covering motorways, urban areas and even arctic conditions in Finnish Lapland. The trials on public road network were mainly carried out in European TEN-T corridors and urban nodes with a specific attention to demanding, error-prone conditions. For safety reasons, a lot of tests were carried out in closed areas.