Mark Brackstone

CAR-FOLLOWING: A HISTORICAL REVIEW

In recent years, the topic of car-following has become of increased importance in traffic engineering and safety research. Models of this phenomenon, which describe the interaction between (typically) adjacent vehicles in the same lane, now form the cornerstone for many important areas of research including (a) simulation modelling, where the car-following model (amongst others) controls the motion of the vehicles in the network, and (b) the functional definition of advanced vehicle control and safety systems (AVCSS), which are being introduced as a driver safety aid in an effort to mimic driver behaviour but remove human error. Despite the importance of this area however, no overview of the models availability and validity exists. It is the intent of this paper therefore to briefly assess the range of options available in the choice of car-following model, and assess just how far work has proceeded in our understanding of what, at times, would appear to be a simple process.

Towards an understanding of adaptive cruise control

Adaptive cruise control (ACC) provides assistance to the driver in the task of longitudinal control of their vehicle during motorway driving. The system controls the accelerator, engine powertrain and vehicle brakes to maintain a desired time-gap to the vehicle ahead. This research describes the results of a detailed microscopic simulation investigation into the potential impacts of ACC on motorway driving. In addition to simulation, real vehicle driving profiles, obtained from instrumented vehicle experiments in three European countries, have been used to compare real following behaviour with that of a simulated ACC equipped vehicle. This new approach has shown that following with an ACC system can provide considerable reductions in the variation of acceleration compared to manual driving. This indicates a potential comfort gain for the driver and environmental benefits. A number of critical situations in which ACC does not perform well have also been identified. The research also highlights the limitations of microscopic simulation in modelling the impacts of ACC because of the lack of understanding of the interaction between the driver and the ACC system relative to the traffic conditions.

Motorway driver behaviour: studies on car following

This paper will report findings of an instrumented vehicle study aimed at assessing one element of driver behaviour, that of car following, on UK motorways. The paper (re-) calibrates one of the most successful of such models—the Action Point model—using dynamic time series data acquired from field tests with an instrumented vehicle. Probability distributions for a number of parameters from the Action Point model are produced and a number of modifications made in order to enhance its value for use in traffic flow and simulation models. Lastly typical headways are compared with existing studies in the area, finding that current headways are far lower than believed. The rationale behind the adoption of such short headways is examined.

Fuzzy sets and systems for a motorway microscopic simulation model

Microscopic simulation modelling has recently become attractive to researchers in traffic engineering as it appears to offer a cost effective and ‘safe’ way of investigating intelligent vehicle-highway system (IVHS) at a fundamental level. However, the deterministic approach used assumes a consistency of behaviour which may severely detract from model validity. This may be overcome by using a ‘fuzzy’ approach to describe drivers’ decisions. This paper describes the development of the fuzzy logic motorway simulation model (FLOWSIM). Emphasis is placed on the research undertaken to establish fuzzy sets and systems for motorway driving behaviour models, the collection of data on appropriate motorway driving behaviour, fuzzy sets and systems calibration, and model validation. Model validation results have shown that the fuzzy model (FLOWSIM) can closely replicate real systems and in test cases have performed better than some common deterministic models such as the ‘GHR’, ‘Gipps’ and ‘MISSION’ models.

A collision model for safety evaluation of autonomous intelligent cruise control

This paper describes a general framework for safety evaluation of autonomous intelligent cruise control in rear-end collisions. Using data and specifications from prototype devices, two collision models are developed. One model considers a train of four cars, one of which is equipped with autonomous intelligent cruise control. This model considers the car in front and two cars following the equipped car. In the second model, none of the cars is equipped with the device. Each model can predict the possibility of rear-end collision between cars under various conditions by calculating the remaining distance between cars after the front car brakes. Comparing the two collision models allows one to evaluate the effectiveness of autonomous intelligent cruise control in preventing collisions. The models are then subjected to Monte Carlo simulation to calculate the probability of collision. Based on crash probabilities, an expected value is calculated for the number of cars involved in any collision. It is found that given the model assumptions, while equipping a car with autonomous intelligent cruise control can significantly reduce the probability of the collision with the car ahead, it may adversely affect the situation for the following cars.

Dynamic Behavioral Data Collection Using an Instrumented Vehicle

A significant problem that has become increasingly apparent in the development of models of driver behavior over the last few years is the absence of reliable data with which simulated processes, such as car following, may be compared. Obtaining such data, and the associated increase in model validity that this would allow, is clearly becoming of greater importance since a reliable baseline is required against which improvements in traffic flow and safety produced by many advanced transport telematics systems can be judged. One source of such data is an instrumented vehicle; a vehicle equipped with relative distance- and speed-measuring sensors that may be deployed in the traffic stream to collect data that are realistic, accurate, and dynamic. The opportunities for data collection afforded by instrumented vehicles are examined, in particular, the construction and testing of a new facility fitted with an optical speedometer, a radar rangefinder (capable of measuring the distance to, and relative speed of, the next vehicle in the traffic stream), and forward- and rear-looking video camera. Examples are given of the use of the vehicle in several current research projects, the operational strategies for which will be presented and discussed along with output. These include experiments on close-following, lane-changing, and the perception of relative speed. In conclusion, future areas of research and development are examined.

Driver headway: How close is too close on a motorway?

Driver headway has recently become an important question with much attention being given to unsafe headways or ‘tailgating’. This paper reviews a series of recent studies undertaken at the University of Southampton, which have sought to measure and model distance keeping, demonstrating how following distance depends on a wide range of factors, some of which are only recently being explored. These include variations in following distance for any particular driver and the relationship with time to collision, variations in following distances of drivers of differing nationalities and the ability of the driver to ‘read the road ahead’, which may be affected by interaction with different vehicle types. It is demonstrated that providing clear unequivocal statements regarding car following and safety levels, even after such studies, is still far from straightforward.

Drivers' use of deceleration and acceleration information in car-following process

Understanding driver behavior is important for the development of many applications such as microscopic traffic simulation models and advanced driver assistance systems. The car-following process is an important phase of driving behavior and takes place when a driver follows a lead vehicle and tries to maintain distance and relative speed within an acceptable range. A key to improving knowledge of driver behavior during this process is determining the information perceived by drivers that could influence their decisions. It has been believed for some time that the main kinematic parameters that affect driver judgment in car following are the relative speed, the distance separation, and the absolute speed. The research described investigated whether drivers are also able to use information on the lead vehicles deceleration or acceleration during the car-following process through experimental validation of current car-following hypotheses. For this research, an instrumented vehicle was used to collect a large database of car-following time sequences, the analysis of which showed strong evidence that drivers are able to perceive information such as the deceleration or acceleration of the vehicle being followed, although no empirical relationship was determined. An example demonstrating the importance of such perception shows that modeling a driver trying to avoid a collision with a lead vehicle would lose 20% of its fit accuracy if the lead-vehicle acceleration state were not considered.

Vehicle to vehicle communication outage and its impact on convoy driving

Advanced vehicle control and safety systems are rapidly evolving around the globe, allowing the testing of convoying systems. A key element in such systems is the production of a reliable vehicle-to-vehicle communication link, capable of robust operation in a range of operating environments. This paper reports on a modelling study of one such system conducted as part of the Road Traffic Advisor project within the UK, based on a communication link in the millimetric waveband (approximately 60 GHz). The likely outage is examined for a basic inter-vehicle data link. This is used in conjunction with a simulation of convoy behaviour based on a control model to examine the effect of such outages on convoy operation. The paper concludes that the impact of the communication protocol and the communication outages on the control dynamics of an automated convoy would not degrade its performance.

WHAT IS THE ANSWER? AND COME TO THAT, WHAT ARE THE QUESTIONS?

This paper replies to the issues raised by commentators on Brackstone and McDonald’s historical review of car following. It focuses on the implications of satisficing, the determinism assumed in engineering inspired models of car following, and the role that individual differences in perceptual and cognitive performance, as well as motivation, must play in future accounts of traffic behaviours such as car following.