Model of driver behavior

To move fi om the hierarchical structure of the driver task to working models of driver behavior, we now need to consider the variables that affect these decisions, the limitations placed on us as decision makers, and the needs and biases that we bring into the driving situation. That is the role of driving models to explain and predict driver behavior in the context of the driver s environment, personal goals, and information-processing limitations. The two classes of models that are described below approach the issue from different perspectives, but they supplement each other more than conflict with each other and both are useful for understanding driver behavior. 

Many of us like to think that v e behave in a rational manner. This is not always the case, and economists often use the rational man model only as a straw man, to demonstrate and understand biases in the actual behavior of people, especially in their purchasing decisions. Our decisions are biased in many ways, and only recently have some of the psychological biases been understood (Tversky and Kahneman, 1992). Still, there is reason to our behavior at least on many occasions, and at least within limits of the information available to us. The challenge to the rational model of driver behavior is to allow for all our limitations and biases. Conceptual approaches to explaining and predicting driver behavior in the context of a process of rational decisions have been offered by Sivak (2002), Fuller (2005), and Parker and her associates (1992). 

Motivational models of driver behavior are labeled as such because they emphasize the driver motivations - rather than the driver capacity - as a key determinant of the driving style and safety. Fuller s model incorporates the motivational aspect through the driver s constitutional features but certainly does not make that the heart of the model. Motivational models assume that most of the time we do not allocate all of our attentional capacities to the safe negotiation of our car. Safety is just one motive, and - judging by the marketing strategies of the... 

There are at least three theoretical approaches or models of driver behavior that can and have been used to account for the relationship between speed and crashes. Each model views the driver and the fraffic environment from a different perspective, and each leads to somewhat different conclusions — all of which have received some empirical support. Two of the models — ftie information processing and the risk homeostasis - have been described in some detail in... 

A detailed breakdown of the specific human factors that precipitated the crashes in MAIDS is provided in Table 16-3. The causes are grouped into the four human information processing mechanisms used in information processing models of driver behavior (see Chapter 3). 

Fiq. 4.7 A simplified task analysis model of driver behavior (from Wilde, 1982)... 

In the present model, the driver is only used as a precondition to certain transitions (T1 and T2). But CPN could be used for the modeling of human behavior. This is an interesting feature as it places the specification model in its context of use and allows the integration of human influence in safety and security analysis. 

The driving state is normally continually monitored (by the driver and/or a system) in order to make corrections on any or all of these levels if required. Detailed applications, variations and refinements of this model can be found in the literature [4, 10-12]. Classically, active safety systems, e.g.. Dynamic Stability Control (DSC), have been designed to provide support at the stabilization level. At this level, the target quantities are generally well defined in terms of vehicle physics. Preventive pedestrian protection, which is in the focus of this thesis, addresses primarily the maneuvering level and thus involves additional eomplexities in control—particularly those involving the interpretation of driver behavior and the interaction of system actions with the driver. 

Michon, J. A. 1985. A critical view of driver behavior models What do we know, what should we do. In L. Evans and R. C. Schwing (Eds.). Human Behavior and Trt fic Safety. New York Plenum Press, pp. 485-520. 

Modjtahedzadeh Hess, R. 1993. A model of driver steering control behavior for use in assessing vehicle handling qualities, ASME J. Dyn. Syst. Meas. Control, vol. 15, no. 3,456-464. 

The best-known motivational model - and the one that has been most frequently challenged — is the risk homeostasis theory of driving behavior. The first formulation of this model was probably Taylor s (1964) risk-speed compensation model, which postulated that drivers adjust their speeds in accordance with the perceived risk. More recently the model has been expanded by Wilde (1998,2002) to include and account for a host of driver behaviors. Because of the controversy it has generated and the research that it has spurred, it will be described here in some detail. According to Wilde, we strive not to minimize risk (or maximize safety), but to reduce (or increase) it to a non-zero level with wdiich we feel comfortable. Because different driving situations have different levels of inherent dangers, we constantly strive to adjust our behavior to maintain a relatively constant risk level. The continuous adjustment process, similar to diat of a room thermostat, is displayed in Figure 3-12. 

Michon, J. A. (1985). A critical view of driver behavior models what do we know, what... 

Summala, H. (1985). Modeling driver behavior a pessimistic prediction In Human Behavior and Traffic Safety (L. Evans and R. C. Schwing, eds.). Plenum Press, New York. Summala, H. (1988). Risk control is not risk adjustment the zero-risk theory of driver behavior and its implications. Ergonomics, 31,491-506. 

For a presentation with a slightly different emphasis than found in this approach, see Glenn Blomquist A Utility Maximization Model of Driver Traffic Safety Behavior. Acci[Pg.52]

When studying complex systems, engineers often employ computer simulation models. Such a model may incorporate empirically based mathematical models as components of the total model. For example, Thomasson and Wright (5) used computer simulation to study traffic at a two-way stop controlled intersection. They first made empirical studies of driver behavior at the intersections and found that the phenomenon could be separated into several well-defined events ... 

The two models given in this example cooperate and are closely linked. Thus the transition T8 from OS to RV mode shall not be possible unless the second model is executed which means that the driver needs to show his intention to reverse and then acknowledge the command. Of course, one of the advantages of such modeling is the automatic construction of paths in the model. And from paths can be derived the required preconditions on the driver behavior. 

Predicting which driver will be involved in a future crash has always been one of the more difficult tasks of the safety pro. The American Transportation Research Institute (ATRl) may have provided some help in this matter. In 2005 ATRl released the results of a study entitled Predicting Truck Crash Involvement Developing a Driver Behavior-Based Model and Recommended Countermeasures. ATRl provides insight into which driving behaviors tend to lead to an increased chance of future crash involvement. The study also determined what companies with low crash rates are doing to prevent behaviors that lead to crashes. 

The importance of subject experiments (or behavioral studies) is founded on the fact that for active safety, driver behavior is more important than the driving characteristics of the vehicle [17]. However, human behavior is subject to a large variability [43], which can be modeled, e.g., on the basis of experiments [87]. The findings from... 

The driver model is represented by Block 2, which in fact is a controller that enforces the conditions prescribed by the NEDC. The controller uses the speed information of NEDC as input and triggers acceleration and braking. The controller unit of the whole system is represented in Block 3. It coordinates the different technological systems that are involved. The behavioral model of the battery module is represented by Block 4. 

The ABM approach is employed in the modeling of car traffic because it is able to capture individual drivers preferences (Galus et al., 2012). In particular for electric car mobility, ABM takes advantage of high resolution road network maps and allows for an accurate mapping between the transportation network and the electricity infrastructure. However, in order to evaluate interdependencies between road and electricity network, detailed data on the electric equipment, drivers behavior, and power demand are needed. 

---