Berthold Färber

Three Decades of Driver Assistance Systems: Review and Future Perspectives

This contribution provides a review of fundamental goals, development and future perspectives of driver assistance systems. Mobility is a fundamental desire of mankind. Virtually any society strives for safe and efficient mobility at low ecological and economic costs. Nevertheless, its technical implementation significantly differs among societies, depending on their culture and their degree of industrialization. A potential evolutionary roadmap for driver assistance systems is discussed. Emerging from systems based on proprioceptive sensors, such as ABS or ESC, we review the progress incented by the use of exteroceptive sensors such as radar, video, or lidar. While the ultimate goal of automated and cooperative traffic still remains a vision of the future, intermediate steps towards that aim can be realized through systems that mitigate or avoid collisions in selected driving situations. Research extends the state-of-the-art in automated driving in urban traffic and in cooperative driving, the latter addressing communication and collaboration between different vehicles, as well as cooperative vehicle operation by its driver and its machine intelligence. These steps are considered important for the interim period, until reliable unsupervised automated driving for all conceivable traffic situations becomes available. The prospective evolution of driver assistance systems will be stimulated by several technological, societal and market trends. The paper closes with a view on current research fields.

Online Driver Distraction Detection Using Long Short-Term Memory

Lane-keeping assistance systems for vehicles may be more acceptable to users if the assistance was adaptive to the drivers state. To adapt systems in this way, a method for detection of driver distraction is needed. Thus, we propose a novel technique for online detection of drivers distraction, modeling the long-range temporal context of driving and head tracking data. We show that long short-term memory (LSTM) recurrent neural networks enable a reliable subject-independent detection of inattention with an accuracy of up to 96.6%. Thereby, our LSTM framework significantly outperforms conventional approaches such as support vector machines (SVMs).

Driving Behavior and Simulator Sickness While Driving the Vehicle in the Loop: Validation of Longitudinal Driving Behavior

The Vehicle in the Loop (VIL) is a simulator, which combines real driving experience with the replicability and safety of simulators. In the VIL test setup a real test vehicle is combined with a virtual testing environment which is displayed to the user via a head-mounted display (HMD). In theory, this simulation concept renders the VIL uniquely suited to the development and evaluation of numerous automotive applications, including driver assistance systems. Aiming to assess the extent to which the VIL elicits realistic driving responses, a validation study was performed. In this first validation study the focus was on longitudinal driving behavior. 44 participants performed five common traffic maneuvers in reality and the VIL setup. Simulator sickness was assessed with the simulator sickness questionnaire (SSQ). Descriptive and inferential analyses of the data showed that the VIL achieves relative validity concerning brake pressure and reaction times and absolute validity concerning the steering angle. However, subjects showed longer reaction times and accelerated more smoothly while driving the VIL. One possible explanation for these results could be the presence of simulator sickness. Overall, the study indicates that the VIL represents a suitable testing method for the evaluation and development of driver assistance systems. The study also provided clear indications for the future development of the VIL.

Position and force augmentation in a telepresence system and their effects on perceived realism

Haptic assistance functions for a telepresence system are presented and assessed. These assistance functions are based on the augmentation of exchanged position and force data, and they are intended to increase the transparency of the telepresence system while maintaining stability. We present the concept and implementation of different assistance functions. Furthermore, we show the setup and results of a psychophysical experiment, which was designed to evaluate the effects of the assistance functions on perceived realism. As a result, the position assistance can increase stability and safety without negatively affecting transparency, and the force assistance can even increase the feeling of presence under certain conditions.

A Meta-Analysis of the Effects of Haptic Interfaces on Task Performance with Teleoperation Systems

Human task performance with teleoperation systems is characterized by long task completion times, handling errors, and excessive force application to objects in the remote environment. Haptic interfaces promise to address these challenges by providing the human user with sensory feedback from the remote environment that would otherwise be lacking. Until now, only few attempts have been made to present current research efforts from a broader, more integrative perspective. To address this need, several meta-analyses were conducted, which aimed at establishing the overall effectiveness of haptic interfaces in improving the critical performance aspects in teleoperation systems. In this context, the influence of potential moderator variables (i.e., virtual versus real teleoperation setup; vibrotactile versus kinaesthetic force feedback) as well as outcome-specific effects (i.e., force regulation ability; task completion time; performance errors) were investigated.

An experimental study of lossy compression in a real telepresence and teleaction system

High network requirements of multiple degrees of freedom haptic data exchange represent a challenge in modern telepresence and teleaction systems. This study presents a systematic evaluation of a psychophysics-based approach to lossy haptic data compression proposed in the literature: the deadband approach and its extension which introduces additional signal model based prediction. In an experimental study, the effects of these approaches on packet rate reduction, perceived interface quality, and relevant task performance criteria were investigated in a three degrees of freedom telepresence and teleaction system. The results indicate that the extended approach did not lead to significant data reduction and adversely affected perceived interface quality as well as task performance. Without prediction, the deadband approach showed excellent rate reduction performance without adversely affecting perceived interface quality and most task performance criteria. We conclude from our study that the combined deadband and prediction approach is not practical for a telepresence and teleaction system with the used control structure, while the deadband compression approach alone exceeded expectations.

Negative Behavioral Adaptation to Lane-Keeping Assistance Systems

In addition to the improvement of driving comfort, modern advanced driver assistance systems also aim to substantially increase traffic safety. Meanwhile, initial optimism with respect to potential safety gains has given way to a more critical view. There is, for example, a danger that continued use is associated with drivers systematically adapting their behavior to the new systems in a negative manner. According to theoretical considerations, excessive trust in such systems may lead to the development of a tendency for the driver to delegate safety-relevant aspects of the driving task to the system, which, in cases where system limits are reached, can result in the driver and other road users being endangered. The present study examined this phenomenon in the context of lateral control assistance. In a field experiment, it was specifically investigated whether drivers who had become familiar with a heading control system developed excessive trust in the system and misjudged its limits. This question was addressed in a driving trial in which, unbeknown to the driver, the heading control system was repeatedly deactivated and the drivers behavior and lane control were observed. Analysis of subjective and objective data clearly demonstrated that drivers did not develop excessive trust in the system and accordingly regulated lateral control to a sufficient degree even after prolonged exposure to the system. Negative behavioral adaptation to the heading control system was thus not observed.

Real-time framework for multimodal human-robot interaction

This paper presents a new framework for multimodal data processing in real-time. This framework comprises modules for different input and output signals and was designed for human-human or human-robot interaction scenarios. Single modules for the recording of selected channels like speech, gestures or mimics can be combined with different output options (i.e. robot reactions) in a highly flexible manner. Depending on the included modules, online as well as offline data processing is possible. This framework was used to analyze human-human interaction to gain insights on important factors and their dynamics. Recorded data comprises speech, facial expressions, gestures and physiological data. This naturally produced data was annotated and labeled in order to train recognition modules which will be integrated into the existing framework. The overall aim is to create a system that is able to recognize and react to those parameters that humans take into account during interaction. In this paper, the technical implementation and application in a human-human and a human-robot interaction scenario is presented.

Implementation and evaluation of a gesture-based input method in robotic surgery

The introduction of robotic master-slave systems for minimally invasive surgery has created new opportunities in assisting surgeons with partial or fully autonomous functions. While autonomy is an ongoing field of research, the question of how the growing number of offered features can be triggered in a time-saving manner at the master console is not well investigated. We have implemented a gesture-based user interface, whereas the haptic input devices that are commonly used to control the surgical instruments, are used to trigger actions. Intuitive and customizable gestures are learned by the system once, linked to a certain command, and recalled during operation as the gesture is presented by the surgeon. Experimental user studies with 24 participants have been conducted to evaluate the efficiency, accuracy and user experience of this input method compared to a traditional menu. The results have shown the potential of gesture-based input, especially in terms of time savings and enhanced user experience.

Automatic Emergency Steering with Distracted Drivers: Effects of Intervention Design

Driver inattention is reported to be one of the most prominent contributing factors to crashes. Modern vehicles feature sensor equipment able to detect an imminent collision, potentially permitting advanced driver assistance systems (ADAS) to cope for such human error. Steering interventions, however, make high demands on the human-machine-interaction. Unlike in autonomous emergency braking, conflicting driver input cannot be omitted. Three different ADAS configurations for an automatic emergency steering intervention with small lateral offset were tested against an unassisted baseline condition in a driving experiment with distracted drivers. The results suggest an influence of feedback modalities and actuator choice.