Skip to main content

BRidging gaps for the adoption of Automated VEhicles

Periodic Reporting for period 3 - BRAVE (BRidging gaps for the adoption of Automated VEhicles)

Reporting period: 2020-03-01 to 2021-02-28

Advances in vehicle automation with minimal human intervention are expected to e.g. increase road safety but bring new challenges that must be addressed to ensure adoption of level 3 automated vehicles (CAC).
The overall objective has been to perform multidisciplinary, user-centric research to ensure that needs of drivers, VRU, other road users and stakeholders are accounted for, resulting in products that are accepted.
Conclusions
1) Multidisciplinary studies of requirements and expectations of drivers, VRU, stakeholders, show a lack of trust in CAC and that communication with CAC desired, e.g. via eHMI.
2) System transparency on vehicle actions can enhance trust, using e.g. concepts for driver-vehicle interaction to assure safe vehicle handling with reduced driver attention. If the system is in control, drivers do not want to be disturbed. The amount of information provided should be adjustable to driver trust and system experience. Mainly use visual feedback.
3) Concepts for vehicle-environment interaction, enhancing current ADAS by inclusion of predictive capabilities can be used for better and faster ADAS reaction, by anticipating system that identifies vehicle cut-in situations and pedestrian intentions to cross road.
4) Validation requirements, acceptance and impact assessment using realistic user-centric testing under different scenario conditions show e.g. that for pedestrians the eHMI should be adequately timed with the vehicle deceleration.
5) Proposals for advancements on regulations and consumer assessment to enhance user acceptance have been developed and presented.
1. Data on road users’ needs and concerns regarding AV was obtained through literature review, stakeholder interviews, and road user focus groups. Based on this, two online surveys were undertaken - an expert survey (D2.2) and a population survey (D2.3). Three papers are prepared, focussing acceptance of CAC from the perspective of different road user groups, ethical aspects of the widespread introduction of CAC; international comparative aspects of the road safety implications of automated vehicles.
Two summaries of the Slovenian and Spanish respondents of the BRAVE population survey were compiled the and published in the member magazine of AMZS, and in media of ACASA.
2.Based on the user requirements identified in WP2 and WP5, FHG led production of a concept for interaction between driver and automated vehicle (D3.2). It targets both safety and acceptance, and was implemented in FHG's driving simulator (see D3.4) to demonstrate it, and to perform a summative evaluation. There the concept has been evaluated positively by end users (update of D5.2). HMI guidelines for trustworthy driver/vehicle interaction were produced and can be adapted to OEM's requirements.
A last-generation Driver Monitoring System (DMS) was developed and tested using videos taken at FHG driving simulator. The DMS classifies the driver's behaviour and status (attentive, questionable, drowsy, asleep), which meets the requirements of the interior HMI concept. The DMS developed is compliant with use cases defined in BRAVE to assure smooth transition between automated and manual driving in SAE 3 level.
Publications include a chapter in a book by sister project TrustVehicle, published in Springer series.
3. UAH has led the development of two prediction systems: for predicting vehicles trajectories and lane changes (D4.3) and for predicting VRUs intentions (D4.4). These have been trained on datasets containing thousands of labelled examples (PREDICTIVE dataset developed in BRAVE). The systems have been put to test in order to assess their validity to increase safety and comfort in ADAS and automated driving systems and, thus, to contribute to enhance user acceptance of automated cars. Quantitative results have been attained, showing a state-of-the art performance. Discussions with OEM about use of the systems (TLR6).
4. The DMS and the predictive systems have been tested in realistic scenarios. A summative evaluation of the HMI concept developed in the BRAVE project (D5.2). (see point 2 above).
• DMS has been tested on several individuals at UAH and at the FHG driving simulator.
• Predictive systems have been tested on 5 use cases, following EuroNCAP or new test protocol at UTAC, using UAH's DRIVERTIVE prototype vehicle. A baseline has been established in these experiments with a view to measure the value that these predictive systems add to current ADAS in terms of safety and anticipation in critical situations. (D5.3). BRAVE’s predictive system anticipates lane changes with 400ms of anticipation with respect to the average human reaction time.
• Tests in VTI’s VR pedestrian simulator evaluated VRUs requirements identified in T4.4 (D4.5) and point out the importance of integrating eHMI messages and vehicle dynamic cues.
• A fully developed concept of the autonomous vehicle and the HMI was demonstrated at UAH premises to media representatives who were interviewed after the demonstration. (D5.4)
5. AD and HMI related test protocols of regulatory and EuroNCAP working groups were evaluated under UTAC leadership. Predictive systems developed in BRAVE were tested at UTAC with a view to provide faster and safer ADAS reactions. Proposals for how to evaluate acceptance enhancing concepts were developed. The tests gave essential input to dialogue with relevant regulatory and EuroNCAP working groups.
UTAC has presented BRAVE proposals to WGs for AD regulation and to EuroNCAP WGs(D6.3 D6.6):
• HMI and driver monitoring, with proposals and recommendations from BRAVE developments by FHG (WP3) and tests in VTI’s driving simulator (WP5),
• vehicle anticipation in critical scenarios, proposing modificatitions of official scenarios, with 4 BRAVE proposals resulting from tests in UTAC (WP5) with BRAVE predictive prototype vehicle built by UAH (WP4).
Presentations will continue throughout 2021.
6. Driver training programmes at AMZS and ACASA are being updated with BRAVE results, to also include ADAS aspects that will increase understanding and adoption.
Communication of BRAVE results to the public has been made via AMZS and ACASA, targeting their large memberships. Furthermore results are disseminated to Mov'eo's network of industries.
The main objective of BRAVE has been to ensure safe vehicles handling with reduced driver attention for level 3 automated driving systems through an effective interaction between the driver and the vehicle.
BRAVE has aimed at enhancing adoption of HMI and ADAS by bringing the users' and stakeholders’ expectations and needs into the design process of a Driver Monitoring System, guidelines for HMI, and systems for prediction of other road users, including VRU.
Through a user-centric approach with several iterations, users have tested automated concepts developed within BRAVE. Testing protocols have been evaluated wit respect to ability to recognise the concepts developed in BRAVE, resulting in proposals for advancements on the regulation and consumer assessment procedures.

BRAVE has developed two predictive ADAS systems and HMI concepts for increased user adoption. Additionally, development of tests protocols that recognise these systems has resulted in proposals for inclusion in official test procedures (regulations, EuroNCAP). The BRAVE actions thus completed contribute to Vision Zero by providing user accepted, more natural and safer human-vehicle interaction.
first subjective and multicultural tests in France in June 2018