| 1. |
- Innamaa, Satu, et al.
(författare)
-
Methodological challenges related to real-world automated driving pilots
- 2019
-
Ingår i: ITS World Congress Proceedings.
-
Konferensbidrag (refereegranskat)abstract
- This paper discusses the methodological challenges related to automated driving (AD) pilots in the realworld, providing an overview of some of the solutions offered by the L3Pilot project. Although theoverall methodology defined for Field Operational Tests (FOTs) has been developed quite extensivelyfor driver support systems, our efforts in the L3Pilot project show that the evaluation process can beadapted to suit the needs of AD pilot projects, as long as some caveats related to the pilot nature of ADstudies are acknowledged. The AD pilots currently in place around the world provide important insightsinto t he impacts of AD on their us ers, other road us ers and the societ y at large. However, as thes e s yst emsmature, large-scale FOTs will be needed as (closer to) ex-post evaluation, to verify the assessed impacts.This paper outlines the challenges and offers some solutions for working towards that goal.
|
|
| 2. |
|
|
| 3. |
|
|
| 4. |
- Rösener, Christian, et al.
(författare)
-
A Comprehensive Evaluation Approach for Highly Automated Driving
- 2017
-
Ingår i: Enhanced Safety of Vehicles. ; , s. 1-13
-
Konferensbidrag (refereegranskat)abstract
- Since the last decade, development efforts by academia and industry for automated driving functions have increased significantly. Also, the European research project AdaptIVe is looking into this topic. Due to the large operation spaces and various complex situations that are covered by these functions, efforts for evaluation increase also significantly. Within AdaptIVe, a comprehensive evaluation approach for automated driving functions ranging from SAE level 2-4 has been developed [1]. The approach splits the evaluation into technical, user-related, in-traffic and impact assessment addressing safety and environmental effects of automated driving. For each evaluation type appropriate test tools and methods are selected e.g. field test for technical assessment, trials on test track and in real traffic for the user-related assessments and simulations for the in-traffic and impact assessment. Next to the assessment type also the characteristics of the function must be considered when deciding for specific test tools. Hence, besides to the level of automation [8] the automated driving functions are classified into continuous and event-based operating functions. Whereas event-based operating functions are only operating for a short period in time (e.g. automated parking), continuous operating functions are, once they are active, operating for longer time periods (e.g. highway automation). Based on the classification the aspects to be evaluated and test methods are selected for all assessment types. The developed methodology has been applied to several automated driving functions developed within AdaptIVe. As an example, for the technical assessment of continuous operating functions it has been assessed whether the driving behavior of the developed functions is similar to human driving behavior and therefore not disturbing human traffic. In the user-related assessment, issues related to driver behavior, understanding of automation, trust, mental workload, resuming control, vigilance, usability and acceptance has been looked at. In this paper the key aspects of the AdaptIVe evaluation methodology for technical, user-related, in-traffic and impact assessment are presented as well as the key results of the application of this methodology on the within AdaptIVe developed automated driving functions.
|
|
| 5. |
- Rösener, Christian, et al.
(författare)
-
Application of AdaptIVe Evaluation Methodology.
- 2017
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- Since the last decade, development efforts by academia and industry for automated driving functions have increased significantly. Also, the European research project AdaptIVe is looking into this topic. Within AdaptIVe, 21 different automated driving functions for different speed ranges and target areas have been developed. They have been developed in three sub projects (SPs), addressing different automation scenarios:Sub project 4: Automation in close-distance scenarios: Addresses manoeuvres at low speed (below 30 km/h) that are characterised by the presence of close obstacles, such as in parking manoeuvres.Sub project 5: Automation in urban scenarios: Deals with driving scenarios in urban environments that are characterised by an average speed range of 0 to 70 km/h.Sub project 6: Automation in highway scenarios: Addresses motorway scenarios (or motorway similar roads) considering velocities up to 130 km/h. Due to the large operation spaces and various complex situations that are covered by these functions, efforts for evaluation are expected to increase significantly. In order to enable an efficient assessment of automated driving functions, within the subproject 7 a comprehensive evaluation methodology addressing this challenge has been developed. Technical Assessment: Evaluates the performance of the developed automated driving functions with respect to a defined baseline.User-related Assessment: Analyses the interaction between the function and the user, trust, usability as well as acceptance of the developed functions.In-Traffic Assessment: Focuses on the effects of the surrounding traffic on the automated driving function as well as the effects of the automated driving function on the surrounding non-users.Impact Assessment: Determines the potential effects of the function with respect to safety and environmental aspects (e.g. fuel consumption, traffic efficiency).The evaluation methodologies developed in previous research projects dealt mainly with active safety functions, for which the assessment focused mainly on testing of functions’ use cases. For automated driving the assessment approach needs to be extended in order to ensure that the whole situation space which is addressed by the functions is covered. Therefore, in the developed evaluation approach the test resources are allocated based on the functions’ classification in order to enable a holistic and efficient assessment. Hence, the automated driving functions are classified based on their automation level and their operation time in two different function types: •Functions that operate only for a short period of time (seconds up to few minutes). Typical examples are automated parking functions and the minimum risk manoeuvre function. These functions are called "event based”.•Functions that, once they are active, can operate over a longer period of time (minutes up to hours). A typical example of this type of function is a “highway pilot”. They are called "continuously operating" functions.
|
|
