The overall goal of SALSA is to increase the safety and acceptance of automated vehicles in mixed traffic. Both types of vehicles, manually controlled vehicles and automated or autonomously driving vehicles, will be operated in parallel in automobile traffic in Germany for the foreseeable future. This will create new, previously unknown and possibly unclear situations.
The aim of this topic in SALSA is to increase acceptance and the sense of security through situationally correct driving behavior in mixed traffic. This means that the passengers of a vehicle should always experience a comfortable, predictable and safe driving behavior when automated driving functions (SAE levels 1 to 3) or autonomous vehicles (SAE levels 4 and 5) are in operation, and other road users should also feel safe due to confident and predictable driving behavior.
To achieve these goals, increase the acceptance and sense of safety of automated and autonomous vehicles, the project partners involved are planning theoretical analyses together with practical studies in driving simulators with test subjects. With this approach, user experiences and user perceptions in different situations are examined and insights for the design of driving behavior are derived.
Several studies on various topics are planned. The driving simulator studies enable passenger to experience automated driving functions and autonomous driving in reproducible situations. Test persons of different ages and with different levels of previous experience drive predefined situations.
Internal acceptance of automated vehicles in mixed traffic
Some users deliberately switch off assistance functions and drive without their help because they suspect that the assistance functions could intervene unintentionally. In situations in which the assistance function would be helpful, they can no longer intervene. In order to prevent the assistance functions from being switched off due to unwanted intervention, the characteristics and driving behavior on the basis of which the driver’s intention can be recognized are being investigated. For example, intervention should be prevented when the driver consciously avoids an obstacle and the assistance system should not be activated in this situation. To implement such a function, the driver’s intention must be provided to the assistance system as an input variable.
A study is first being conducted to determine how the driver’s intention can be determined. Can this be recognized from the driver’s inputs, such as the steering angle or pedal positions? Are there other parameters that allow conclusions to be drawn that a driver is consciously leaving his lane? A study in which the behavior of several test subjects is to be examined is intended to shed more light on these questions. The results of this study will then be used to derive requirements for upgrading an assistance system and for better recognizing the driver’s intention.
Internal acceptance of autonomous vehicles in normal operation
Another block of the “Acceptance” topic deals with the question of how passengers in autonomous vehicles (SAE levels 4 and 5) want to be driven. In the past, the desired driving behavior of a vehicle for the driver has often been investigated. However, if the drivers are no longer responsible for driving the vehicle and are instead passengers, the demands on driving behavior change. In the predecessor project RUMBA, investigations have already been carried out in this area. In SALSA, these findings are built upon and the desired driving behavior is examined in more detail.
Various driving dynamics variants are compared for acceleration and deceleration, as well as for cornering. With the help of the vehicle passengers’ perceptions, recommendations for adapted driving behavior in normal driving situations are derived that come close to the wishes and ideas of different people.
Internal acceptance of autonomous vehicles in minimal risk maneuvers
When driving with autonomous vehicles, there is a small possibility that the intended operating range will be left and an passenger will be unable to take over the driving, e.g. because they are sleeping. In this case, a so-called minimal risk maneuver is necessary. This means that the vehicle should autonomously bring itself into a situation with the lowest possible risk for passengers, without endangering third parties. There are both hard, legal requirements, e.g. how quickly a safe state should be achieved, and soft requirements, i.e. the “path” by which the safe state should be achieved. A minimal risk maneuver should be designed in such a way that the passengers are given a high sense of security and a high level of trust in the system. SALSA is also investigating how this can be implemented in terms of driving dynamics.
Rule-compliant driving behavior of automated vehicles in mixed traffic
In road traffic, there are clear rules that are intended to create order. These rules are defined by signs, markings and laws. However, not all road users always follow them. Often, rules are broken in order to get ahead faster, because one overestimates one’s own driving ability, or for personal gain. Sometimes, ignoring certain rules – such as coming to a complete stop at a stop line – is even considered normal.
However, automated vehicles will strictly adhere to these rules. This serves both as legal protection for the manufacturer and for the safety of the passengers. However, this raises the question of the extent to which the rule-abiding driving behavior of automated vehicles could lead to a conflict with other road users. Bystanders and passengers could be disturbed by the vehicle strictly adhering to the traffic rules. The strictly correct driving behavior could lead to traffic congestion or tempt other road users to carry out dangerous overtaking maneuvers. The project identifies such scenarios that could potentially hinder acceptance or even be critical. These situations are then simulated in a driving simulator study. The experiment focuses on the acceptance of rule-abiding driving behavior and on suitable solution strategies to increase the acceptance and safety of all road users.
In this area, the Federal Highway Research Institute (BASt), INVENSITY GmbH, Spiegel Institut Mannheim GmbH, the Würzburg Institute for Traffic Sciences (WIVW), the Stuttgart University of Applied Sciences (HdM), Valeo Schalter und Sensoren GmbH and the Research Institute of Automotive Engineering and Vehicle Engines Stuttgart (FKFS) are working together on solutions to the question of how to increase the internal acceptance and sense of security of automated and autonomous vehicles.
We thank all partners for their support and look forward to the next steps in the SALSA project!
