Sie können Operatoren mit Ihrer Suchanfrage kombinieren, um diese noch präziser einzugrenzen. Klicken Sie auf den Suchoperator, um eine Erklärung seiner Funktionsweise anzuzeigen.
Findet Dokumente, in denen beide Begriffe in beliebiger Reihenfolge innerhalb von maximal n Worten zueinander stehen. Empfehlung: Wählen Sie zwischen 15 und 30 als maximale Wortanzahl (z.B. NEAR(hybrid, antrieb, 20)).
Findet Dokumente, in denen der Begriff in Wortvarianten vorkommt, wobei diese VOR, HINTER oder VOR und HINTER dem Suchbegriff anschließen können (z.B., leichtbau*, *leichtbau, *leichtbau*).
Das Kapitel stellt den Fahrsimulator DriSMi vor, eine hochmoderne Einrichtung am Politecnico di Milano, die den Einsatz von vernetzten, kooperativen und automatisierten Mobilitätssystemen (CCAM) und Fahrerassistenzsystemen (ADAS) bewerten soll. Die einzigartige seilgetriebene Betätigung und immersive Umgebung des Simulators ermöglichen realistische Simulationen verschiedener Fahrszenarien und ermöglichen die Bewertung objektiver und subjektiver Leistungskennzahlen. Der Text hebt die Fähigkeit des Simulators hervor, sich in den Mikrosimulator für den SUMO-Verkehr zu integrieren, wodurch eine virtuelle Umgebung entsteht, in der mehrere Fahrer interagieren, einschließlich automatisierter Fahrzeuge, die von kooperativer künstlicher Intelligenz gesteuert werden. Ein Schwerpunkt des Kapitels ist eine Studie zur Akzeptanz der Adaptiven Geschwindigkeitsregelung (ACC). An der Studie nahmen 33 Freiwillige teil, die das ACC-System in verschiedenen Szenarien testeten, einschließlich software- und menschengetriebener Fahrzeuge. Die Ergebnisse zeigten, dass die subjektiven Bewertungen von Komfort und Sicherheit der Fahrer stark mit objektiven Indizes wie Mindestabstand zum vorausfahrenden Fahrzeug und Kollisionszeit korrelierten. Die Einbeziehung eines zweiten menschlichen Fahrers in die Szenarien ermöglichte aggressivere Manöver, wodurch die Sicherheitsleistung des ACC besser beurteilt werden konnte. Das Kapitel schließt mit der Betonung der Bedeutung von Fahrsimulatoren wie DriSMi für die Untersuchung der Nutzerakzeptanz und -leistung autonomer Fahrsysteme, die eine sichere, kontrollierbare und ressourceneffiziente Testumgebung bieten.
KI-Generiert
Diese Zusammenfassung des Fachinhalts wurde mit Hilfe von KI generiert.
Abstract
The safe deployment of Connected, Cooperative and Automated Mobility (CCAM) systems needs to take into account advanced human-machine interaction. In fact, CCAM is going to change both the cooperation between the vehicle and the human driver and the interaction with road users. This paper presents DriSMi, an advanced cable-driven dynamic driving simulator that enables safe, affordable and reliable CCAM testing. DriSMi can be used with the objectives of (1) modelling and testing the technology and (2) characterizing and modelling the driver, in particular analysing and modelling the driver’s behaviour, ergonomics and safety in different infrastructure/weather/traffic scenarios. After describing the features of DriSMi, a CCAM scenario and a procedure for assessing the acceptability of Adaptive Cruise Control are presented.
1 Introduction
The safe deployment of Connected, Cooperative and Automated Mobility (CCAM) systems needs to take into account advanced human-machine interaction. In fact, CCAM is going to change the interaction between the vehicle and the human driver. CCAM is also redefining the interaction of vehicles with other road users.
Safe and affordable testing is enabled by indoor testing facilities, namely dynamic driving simulators. Testing must also be trustable, which is currently a major concern [1]. A compromise is needed among the potentially conflicting requirements of safety, affordability and trustable testing [1].
Anzeige
Driving simulators are more and more used to cope with the mentioned requirements [2]. Actually, traffic simulations at the micro-scale are now reproducible within a virtual reality environment [3]. This enables to check the performance of CCAM and, more simply, of Advanced Driver Assistance & Autonomous Driving Systems (ADAS/AD), as they are perceived by vehicle occupants. Vehicle occupants, i.e. drivers and passengers, are the actual actors and judges of CCAM performance. CCAM performance has to be assessed both subjectively and objectively. The objective performance is related to the Safety of the Intended Function (SOTIF): in [4] the SOTIF is defined as “the absence of unreasonable risk due to hazards resulting from functional insufficiencies of the intended functionality or by reasonably foreseeable misuse by persons”. The subjective evaluation of CCAM performance is more tricky to be defined and actually there is a lack of standardization in this area. Nonetheless the perceived safety and comfort are crucial to the deployment of CCAM [5].
In this paper a new generation driving simulator is presented. Then a brief example of how CCAM performance can be assessed is shown. Finally, a methodology for assessing the acceptability of ADAS by the driver is described and applied to the case of Adaptive Cruise Control (ACC).
2 Features of the DriSMi Driving Simulator
DriSMi is a new generation dynamic driving simulator installed at Politecnico di Milano in 2021, able to realistically simulate a driving vehicle (Ego vehicle, Fig. 1) in many different scenarios. First simulator in the world equipped with a new technology of cable actuation, it is characterised by realism and immersivity. Its mid-size allows keeping the latency below the threshold that is perceived negatively by the driver.
The simulator structure comprises six electric actuators (hexalift) that can move the driver’s compartment in six ways (movements along the x, y and z axis, yaw rotation, pitch and roll). The hexalift system is mounted on three pneumatic pads for moving the driver’s compartment on a plane without friction. In addition to five projectors and a 270° screen, a seat with active cushions and tension belts is used to improve the driver’s immersiveness during the test. These systems allow long lateral and longitudinal accelerations to be felt. For the purpose of studying comfort and driver behavior, 8 shakers on board reproduce the vibrations that come from the engine and road irregularities. There are an instrumented steering wheel able to measure the forces and moments exerted by the driver, an eye-tracking system for studying the driver’s behaviour while driving, and a bio-telemetry system that measures the electrocardiogram and the skin potential resistance during driving. Lastly, there is a cap to make the electroencephalogram to the driver during manoeuvres.
The motion of surrounding vehicles can be managed by:
1.
a traffic engine to simulate different traffic conditions based on AI algorithms;
2.
directly imposing the trajectory of the vehicles to perform some specific manoeuvres (lane crossing, lane change, emergency braking, etc.);
3.
additional human drivers to perform random and/or risky/aggressive manoeuvres (multi-user driving simulation mode).
To enable multi-driver simulations, the DriSMi control room hosts a control station including a high-end gaming steering wheel (provided with a feedback torque device) with gear paddles and pedals and 24’’ size LCD-display showing the visual field of the vehicle controlled by a second human driver (Second-Drive, Fig. 2).
Fig. 2.
Experimental set-up for the dynamic driving simulator with multi-drivers.
Copyright Politecnico di Milano, reproduced with permission.
DriSMi can be used with the objectives of i) modelling and testing active controls, advanced driver-assistance systems, and (cooperative) connected and autonomous driving, and ii) characterizing and modelling the driver, in particular analyzing and modelling the driver’s behavior, ergonomics and safety in different infrastructure/weather/traffic scenarios, following the introduction of CCAM systems.
A virtual simulation environment like DriSMi represents a key facility to test the deployment of CCAM/ADAS systems and to assess the driver behaviour and the interaction with the other road users (including vulnerable users such as pedestrians and/or bicyclists). In fact, whereas some systems are already at an advanced stage and available on the market (e.g. lane keeping and adaptive cruise control), open issues are still related to user acceptance. Since these systems substitute the human driver in controlling the vehicle, it is essential that their behaviour is perceived as safe and comfortable by users [7, 8]. It is therefore clear that driving simulator like DriSMi give the opportunity to deeply investigate all these aspects.
3 Multi-Vehicle Interaction Using the SUMO Micro-Simulator
By integrating the SUMO traffic micro-simulator [6], DriSMi provides a virtual environment in which many different drivers interact. SUMO offers the possibility to include driver models such as IDM and Wiedemann. Automated cars can be also included. For example, a CCAM environment was fully reproduced in a case study of the Horizon 2020 project AI@EDGE, which aims to introduce enhanced 5G technology, based on edge computing and Artificial Intelligence [3]. The traffic into a three-leg roundabout was reconstructed, simulating human-driven cars interacting with automated cars driven by an edge controller based on a cooperative artificial intelligence policy.
In the virtual environment, represented in Fig. 2, automated vehicles (AVs) share information with the MEC-EDGE node at the center of the roundabout. The human-driven vehicle is sharing information but not receiving any data. SUMO is used both to drive AVs and to provide the training data to define the policy of the reinforcement learning algorithm, based on a cooperative approach (Fig. 3).
Fig. 3.
CCAM environment reproduced at the DriSMi.
Copyright Politecnico di Milano, reproduced with permission.
Enabling the reproduction of real-world driving experiences under safe, controllable, and time- and resource-saving conditions, driving simulators such as DriSMi offer the possibility to investigate the acceptance of ADAS. In particular, the Adaptive Cruise Control technology is already at an advanced stage in terms of stability and safety performances, but there are still open questions related to user acceptance, in particular in terms of perception of safety and comfort. Clearly, to gain meaningful data, it is essential to consider a variety of scenarios involving other vehicles.
As an example, a first study was conducted to setup a methodology capable of assessing the acceptability of an ACC. As known, ACC maintains the user-defined speed set-point when the road ahead is clear, switching to distance control whenever obstacles are detected, thanks to the information retrieved from sensors (cameras and radars). Therefore, ACC substitutes the human driver in controlling the longitudinal dynamics of the vehicle [7, 8].
The study was carried out involving 33 volunteers (23 males and 10 females) ranging from 19 to 66 years old (34 years as an average). Among the participants, 14 had previous experience with a driving simulator and 10 had previous experience with ACC systems on public roads. Each participant was required to drive the dynamic driving simulator vehicle (Ego), which was equipped with an ACC, while the other vehicles involved in the scenario were either software-driven or driven by a second human driver (Second-Drive). The scenario consisted of a three-lane motorway with speed limit at 120km/h. During the test, the participant had to follow a reference vehicle, with a third vehicle (lane-changing vehicle) performing several maneuvers such as lane insertion, lane crossing (see Fig. 4) and sudden braking. Each participant performed four runs:
RUN 1: ACC system OFF and software-driven lane-changing vehicle;
RUN 2: ACC system ON and software-driven lane-changing vehicle;
RUN 3: ACC system OFF and Second-Drive lane-changing vehicle;
RUN 4: ACC system ON and Second-Drive lane-changing vehicle.
At the end of the test, participants were asked to answer a survey about the driving simulator realism and the ACC system performance in terms of safety and comfort.
A multinomial regression model was trained for correlating comfort and safety subjective ratings provided by the participants and quantitative objective indexes measured during the tests, namely the minimum distance from the preceding vehicle, the time-headway to the vehicle ahead, the time-to-collision, the peak braking acceleration and the rms longitudinal acceleration during the manoeuvre.
The main results of the analysis can be summarized as follows [9, 10]:
time-headway is the index most correlated with the driving style: drivers adopting a short time-headway when the ACC is OFF are likely to prefer a more aggressive tuning of the ACC control logic;
both comfort and safety subjective ratings strongly correlate with the same indexes: minimum distance from the vehicle ahead and time-to-collision. Contrarily to the expectation, the correlation is weak between comfort ratings and longitudinal accelerations perceived by the drivers (both peak and rms values). This means that drivers judge vehicles that are considered safer as more comfortable. As a consequence, making a clear distinction between comfort and safety ratings appears challenging;
results are consistent both considering human-driven (Second-Drive) or software-driven surrounding vehicles;
including a Second-Driver into the scenario, however, allows for more aggressive and risky manoeuvres, which better highlight the performance of the ACC in terms of safety. In fact, safety ratings for manoeuvres performed in RUN 4 (Second-Drive and ACC ON) are higher than the ones in RUN 2 (software-driven vehicles and ACC ON).
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Mastinu, G., Previati, G., Campi, E., Gobbi, M.: Traffic simulation with human in the loop: roundabout scenario with reinforcement learning, edge computing and 5G connection. In: DSC Congress, Antibes (2023)
4.
ISO/PAS 21448, Road vehicles – Safety of the intended functionality
5.
De Winter, J.C.F., Happee, R., Martens, M.H., Stanton, N.A.: Effects of adaptive cruise control and highly automated driving on workload and situation awareness: a review of the empirical evidence. Transport. Res. Part F: Traff. Psychol. Behav. 27(PB), 196–217 (2014)
6.
Previati, G., et al.: SUMO roundabout simulation with human in the loop. In: SUMO User Conference, Berlin (2023)
7.
Xiao, L., Gao, F.: A comprehensive review of the development of adaptive cruise control systems. Veh. Syst. Dyn. 48(10), 1167–1192 (2010)CrossRef
8.
Bengtsson J.: Adaptive cruise control and driver modeling. Department of Automatic Control, Lund Institute of Technology (2001)
9.
Asperti, M., Francesconi, A., Sabbioni E.: Development and testing of an adaptive cruise control system using a dynamic driving simulator with multi-drivers. In: Proceedings of the 22nd DSC, Antibes (2023)
10.
Asperti, M., Francesconi, A., Sabbioni, E.: Development acceptance assessment of an adaptive cruise control system using a multi-driver dynamic driving simulator. In: Proceedings of 28th IAVSD Symposium, Ottawa (2023)
