Driver support in congestion. An assessment of user needs and impacts on driver and traffic flow

Mobility is a key factor for modern societies. However, it also brings about problems, such as congestion, accidents and pollution. High expectations rest on in-vehicle systems to contribute to solving these problems. These so-called driver support systems use advanced information and communication technology to assist the driver in performing elements of the driving task, such as maintaining a proper speed or avoiding an accident. A variety of systems is under investigation or already commercially available. Most current systems are autonomous systems that do not communicate with other vehicles or the infrastructure. Recently, the development of driver support systems is more and more directed at cooperative systems that do communicate and therefore extend the driver’s horizon. Despite the research and development efforts, the market introduction of driver support systems finds itself in an early stage. Car manufacturers employ a rather conservative strategy, because they are uncertain about the financial risks and the usability of these systems. Governments and road operators are uncertain about the actual impacts of driver support systems on traffic safety and traffic efficiency, which makes them hesitant to take measures to facilitate, stimulate or regulate the introduction of these systems. This thesis aims at reducing the above uncertainties by improving the knowledge of user needs for driver support systems and the impacts of one of such systems, the so-called Congestion Assistant, on the driver and the traffic flow. The success of driver support systems is highly dependent on the willingness of the end users, that are the drivers, to have and use these systems. So it is essential to know to what extent drivers would like to be assisted by their cars when driving. Therefore, a user needs survey was conducted to investigate the perceived needs for driver assistance. This survey focused on support with several driving tasks and situations, which is in contrast to earlier research that generally concentrated on ‘ready to use’ systems, such as Adaptive Cruise Control. A total of 1049 Dutch car drivers completed the survey on the Internet. It appeared that warnings for downstream traffic conditions and warnings for traffic in blind spots were favoured. Apparently, drivers appreciate being well informed when driving. Automatic actions from the car were not rated highly, except for taking over the driving task in congestion. Furthermore, the respondents preferred the ideal system to give support in critical situations, such as an imminent crash or reduced visibility. These needs have implications for the design of driver support systems. One can expect that the integration of functions is sensible, for example by exchanging information between vehicles and using one user interface. The results from the user needs survey also revealed a significant need for several forms of congestion assistance. Based on these preferences, the Congestion Assistant was developed. This in-vehicle system consists of a mix of informing, assisting and controlling functions and supports the driver during congested traffic situations on motorways: • Warning & Information: the driver receives warnings about a traffic jam ahead and information about the length of the jam when driving in it. • Active pedal: the driver feels a counterforce of the gas pedal when approaching the jam at too high speed. • Stop & Go: the system takes over the longitudinal driving task from the driver when driving in the jam. Changes in driving behaviour due to driver support systems determine how useful and effective these systems are. So it is necessary to know to what extent drivers are able and willing to interact with a system. Therefore, a driving simulator experiment was conducted to investigate the impacts of the Congestion Assistant on the driver. A total of 37 participants gained experience with the Congestion Assistant in the driving simulator during normal view and fog conditions. These participants were selected from the respondents to the user needs survey. Their acceptance of the Congestion Assistant appeared to be related to their perceived needs for congestion assistance. It was therefore concluded that the user needs survey can be seen as a valid method for the indication of driver needs for congestion assistance. The assessment of the Congestion Assistant in the driving simulator experiment focused on driving behaviour, mental workload and acceptance. The Warning function was not found to affect driving behaviour. The Active pedal caused earlier speed adaptations and safer carfollowing behaviour when approaching the traffic jam, which shows indications of an improved traffic safety. The Stop & Go resulted in ‘smoother driving’ with smaller time headways in the traffic jam, which is expected to enhance traffic efficiency. The participants experienced a lower mental workload with the Congestion Assistant, but only when driving in fog. The mental workload was higher when one approached the traffic jam with the Active pedal. This could be due to an increase in the driver’s attention to the upcoming jam or to the pedal itself giving a ‘sudden’ counterforce. Driving with the Stop & Go resulted in a lower mental workload. This might decrease the driver’s alertness. In general, the participants stated that they appreciated the Congestion Assistant and were willing to buy the system. Particularly the Warning & Information and the Stop & Go were favoured. The acceptance of the Stop & Go significantly increased after having gained experience with it. The participants were less enthusiastic about the Active pedal. Individual driving behaviour determines to a large extent how efficient and safe the traffic flow behaves. So it is important to understand the significance of a change in driving behaviour of individual drivers due to driver support systems in relation to the performance of a whole traffic flow. Therefore, a microscopic traffic simulation study was conducted to investigate the impacts of the Congestion Assistant on the traffic flow. The Congestion Assistant in this study included either the Active pedal or the Stop & Go or a combination of both functions. The traffic flow impacts of six variants of the system were analysed at two equipment rates. The simulated road consisted of a four-lane motorway segment with a left lane drop that caused congestion. The traffic flow model applied in this research was extended to include vehicles equipped with the Congestion Assistant. Data collected on the Dutch A12 motorway were used to validate and calibrate the reference situation in which no vehicles were equipped with the system. The simulation results showed a satisfactory resemblance with respect to the congestion build-up. The calibration process led to more insight into the trade-off between the parameter settings on the one hand and the onset and course of congestion on the other hand. The assessment of the Congestion Assistant in the traffic simulation study focused on traffic efficiency and traffic safety. All variants of the Congestion Assistant resulted in less congestion in comparison with the reference situation. The higher the equipment rate of the Congestion Assistant, the larger were these positive effects. The Active pedal smoothed the traffic flow when approaching the traffic jam by inducing better anticipation behaviour of the driver compared to unsupported drivers. This had a small effect on the dissipation of congestion, rather it affected traffic safety by a safer approach to the jam. Vehicles equipped with the Stop & Go followed other vehicles more efficiently than non-equipped vehicles when driving in and leaving a jam by maintaining smaller headways and eliminating the reaction time of drivers. This reduced the amount of congestion significantly. At the same time, this function also increased the amount of hard braking. Adapting the acceleration algorithm of the Stop & Go will presumably compensate for this effect. The Active pedal showed no added value with respect to traffic efficiency when it was combined with the Stop & Go. But the combination of these two functions decreased the percentages of hard braking and small Time-To-Collision* values, although these percentages were lowest for the Congestion Assistant consisting of only the Active pedal. In conclusion, this thesis provides more insight into the user needs for driver assistance and the impacts of the Congestion Assistant on the driver and the traffic flow. The promising results found in this research project give rise to speeding up the further development of the Congestion Assistant. The Warning function and the Active pedal are assumed to have knowledge of what is happening further down the road. Such cooperative applications will probably become available after 2010. Until then, the efforts should also be concentrated on autonomous applications, such as the Stop & Go. For the automotive industry, it is relevant to know that people are willing to hand over the driving task in congestion to their cars. For public authorities, it is important to realize that the Stop & Go has promising impacts on the dissipation of traffic jams. In view of the severe congestion problems in Europe, it is recommended that both parties work together and in the short run come to a system that serves all interests, including those of the driver, best. * Time-To-Collision: the time required for two vehicles to collide if they continue at their present speed and on the same path.