|
|
BEHAVIOURAL RESPONSES TO DYNAMIC ROUTE GUIDANCE (DRG) SYSTEMS / P G Jackson
In: Paper to be presented at the PICT International Doctoral Conference, 28th-30th March
1994.
University of London Centre for Transport Studies
P G Jackson
Centre for Transport Studies
Department of Civil Engineering
Imperial College
London
February 1994
ABSTRACT
The possibility of providing traffic and navigation information to drivers via Road Transport Informatics will soon become a reality. Although a large number of computer simulations have been run with the aim of increasing understanding of this complex equipment and its effects upon drivers and the road network, very little field research has been carried out. In particular, very little is known about individual drivers' responses to Dynamic Route Guidance and Information (DRG) systems. In addition, a number of transport researchers have, in recent years, cited cognitive maps as playing a major part in DRG, both in teens of their importance for determining travel behaviour and in terms of DRG's ability to extend the spatial knowledge that they contain.
This paper has been written with the aim of clarifying some of the misconceptions which surround cognitive maps and thus describes the mechanisms which are involved in the process of acquiring spatial knowledge. Following a review of the acquisition process and the internal structures which have been proposed to hold such knowledge, the paper draws on theory and research to predict some of the potential consequences of DRG use. The paper examines the way in which drivers acquire information during tasks such as route planning and wayfinding and compares this with how using DRG might affect the execution of such tasks. The importance of the information's salience to the individual and of the individual's goals and purposes are highlighted in considering how a driver evaluates the information presented.
1 INTRODUCTION
The effect of increased traffic congestion in modem urban areas has been the cause for concern throughout the industrialised world for more than a quarter of a century. Whilst the problem has been well documented and researched, the various approaches adopted to change the situation have met with little success. Improvements to highways, traffic control systems and to the vehicles themselves have all gone some way to raising the quality of the driving environment but have had little effect upon the traffic congestion which blights modem urban areas across the world. The problem of traffic congestion is seen to be compounded by the fact that poor route knowledge leads to drivers taking inefficient routes, thus wasting time and other resources in unnecessary mileage. This waste has been estimated as being up to 12% of the current total mileage and time spent on the road (Wootton, Ness and Burton, 1981).
Recent technological advances, however, offer transport researchers and planners a possible means of significantly improving the efficiency of the transport network, resulting in reductions in congestion levels, improvements in the levels of safety and generally raising the quality of the driving experience. The technology which is envisaged to have such an effect upon these problems is known by the generic term 'Road Transport Informatics' (RTI). RTI, and more specifically, dynamic route guidance and information (DRG) systems (See Section 2 below), have been predicted to offer many benefits both to the individual driver and to the road system as a whole. In particular, it is predicted that route guidance and information systems will be capable of expanding drivers' route knowledge, which in turn will have a dramatic effect upon the network, providing a means of reducing wasted mileage and time.
Safety, ethical and experimental control considerations have prevented extensive field use of DRG systems, although it should be noted that a number of field teals have taken place (e.g. LISB, in Berlin), while others such as ADVANCE in Chicago are planned (ADVANCE, 1990). Consequently, the benefits which have been predicted are predominantly based on research which has used DRG systems either in simulations, mock-ups or in laboratory experiments:
"While these may be of value for initial exploration, their ability to induce real-life responses remains in question - very little validation work has been carried out on them." (Watling, 1993, p.24).
Similarly, surveys which have explored drivers' stated preferences with regard to information provision assume that such responses will compare with drivers' real world preferences. Whilst such approaches are valuable and indeed, essential during the development of technologically advanced equipment such as route guidance and navigation systems, it would be unwise to base predictions of their benefits solely upon one, inconclusive area of research.
1.1 Structure of the paper
However, a more fundamental problem exists. This paper considers the central prediction that DRG will expand drivers' route knowledge. Whilst the cognitive structures which have been hypothesised to hold such spatial knowledge ('cognitive maps') are often referred to in transport research, no attempt has been made to clarify the term, or describe the acquisition and formation of cognitive maps, their structure or how new information affects them. Consequently, assumptions about the structure of cognitive maps and predictions about the effects of additional information are made with little evidence to support these predictions. This paper aims to clarify the concept considered so central to the success or failure of route guidance and information systems. To enable predictions to be made about drivers' responses to additional transport information, and its effects upon the driver and the network, we should understand a number of aspects of the process correctly known as environmental cognition.
First, we need to understand how individuals acquire information about their environment: Section 3 reviews this process and the mechanisms proposed to handle the information. Section 4 considers the factors that affect the uptake of information. This is of crucial importance to route guidance navigation systems as the style and content of the information presented may determine drivers' willingness to respond to such information. we need to understand the structure of cognitive maps so that information can be presented in such a way that it is compatible with these existing structures, section 5 reviews this aspect of research on cognitive maps. Finally, we need to know what effects new information about an environment will have on this existing representation, how is the cognitive map updated as a result of receiving new information? Section 6 considers this question and, based upon the above research findings, compares the performance of wayfinding and navigation tasks with and without the benefit of DRG systems.
Through answering these questions it is hoped that we will gain a better understanding of the process by which individuals, and particularly drivers, acquire, mentally store and represent, retrieve and use information about an environment such as a transport network. Through gaining such an understanding it is hoped that DRG systems can be designed most effectively, ensuring that the system has the best chance of being accepted by drivers as a viable aid to driving.
2 WHAT IS RTI?
Before considering these questions, what driver information systems are we concerned with? RTI covers a number of technologically advanced techniques for providing drivers with motoring information via a combination of information and telecommunication technologies. RTI systems cover three generic functional areas:
The technology required for ATMS is now available and indeed there are examples of such systems already in operation around the world. In contrast AVCS are in the early stages of development and to a large extent their continued development is dependent upon the successful deployment and wide spread acceptance of ADIS - It could be argued that the possible implementation of some of the higher level services such as collision avoidance will depend heavily upon drivers' perceptions of the system over the next ten to twenty years; if drivers are unwilling to trust RTI for relatively simple, low-level tasks such as route guidance they will be still more unwilling to use these systems for complex procedures which entail placing full trust in the equipment.
A range of methods for providing information exist, including television and radio channels dedicated to broadcasting traffic information; electronic pre-trip route planning systems; on-board navigation systems, roadside variable message signs (VMS), parking guidance systems and the subject of this paper: dynamic route guidance (DRG) systems.
2.1 Facilities and benefits offered by DRG systems
DRG systems utilise advanced electronics and telecommunications technology to provide drivers with a range of possible types of information (in the case of DRG, via a dashboard-mounted display) about the network. Stergiou and Stathopoulos (1989) list 18 potential facilities that such systems might eventually provide, although in the short to medium term this list can be abbreviated to providing information about:
The provision of such information is predicted to have a dramatic effect upon a number of aspects of traffic systems and their users. Bonsall (1992) has summarised the effects of 'In-Vehicle Route Guidance and Information systems' as follows:
Hence, RTI systems aim to affect key elements of the traveller's journey such as route, cost, speed and accessibility (Bonsall, Pickup and Stathopoulos, 1991). For the individual driver this has the advantage that awareness of the travel conditions and congestion problems for a particular day will be increased, while for the efficiency of the network, the provision of information has the potential of eliminating poor route choices and thus diminishing excess travel times (Ben-Akiva, de Palma and Kaysi, 1991). In addition, the potential to provide information regarding the current state of modes of public transport (such as current average journey time to a specific destination) could influence drivers to select an alternative means of commuter transport for the day with the potential to reduce traffic on the roads.
However, such predicted benefits rest on assumptions that drivers will be willing to use information about the network in preference to their own knowledge and experience and that the information provided will be processed similarly to information acquired in the real world. The following section looks at the acquisition of spatial knowledge in real world environments to test these assumptions.
3 THE ACQUISITION OF SPATIAL KNOWLEDGE
3.1 Environmental cognition
For more than thirty years psychologists, geographers and planners have explored the processes by which individuals come to know about and represent their spatial environment. This area of research, environmental cognition, developed out of early attempts to identify specific features within the urban environment which make it memorable or contribute to our overall image of that environment (e.g. Lynch, 1960; Appleyard, 1969; 1970). Recent research has placed mole emphasis upon the reciprocal nature of person-environment relationships. Environmental cognition is seen to refer not only to specific features within the environment, but as much to events and happenings within the environment, the meaning attached to these elements, the emotions aroused by them and to the significance that they have to us as individuals and groups (Moore and Golledge, 1976).
3.2 Cognitive maps and their significance to transport research
With respect to transport research, the study of environmental cognition is important because the processes that it explores are central to activities such as wayfinding, navigation and route choice selection. In particular, what are commonly referred to as cognitive maps (Downs and Stea, 1973; Kaplan, 1973; Tolman, 1948) play a major role in influencing an individual's behaviour (Downs and Stea, 1973). The perception, prediction and evaluation of possible courses of action can be represented by cognitive maps (Kaplan, 1973); they are seen as coping mechanisms which enable the individual, faced with a vast amount of information, to establish both the location of valued articles and a means of getting to them from the individual's present location.
There have been a number of references in recent years to the importance of cognitive maps to transport research (e.g. Bovy and Stem, 1990; Wenger, Spyridakis, Haselkorn, Barfield and Conquest, 1990; Schofer, Khattak and Koppelman, 1993). Within the field of transport, however, there has been kale attempt to explain the reasons for this importance or to describe the processes which are involved.
3.2.1 Cognitive maps
The term cognitive map was originally coined to describe the process by which rats learnt a maze (Tolman, 1948) - as a metaphor for the process by which they learnt the correct route, in that they appeared to develop an internal map-like representation of the environment However, Tolman's use of 'map' has long been seen as potentially misleading in its implication of a structure with some resemblance to a cartographic map, suggesting an ordered, organised representation of spatial information. On the contrary, research suggests that, unlike real maps, cognitive maps are likely to be distorted and fragmented (Appleyard, 1970; Stevens and Coupe, 1978). Cognitive maps enable us to construct a mental representation of an environment which cannot be seen from one vantage point alone (Arthur and Passini, 1992) and so they enable us to build a meaningful whole from the various parts which have been previously perceived. This definition is important because it emphasises the subjective meaning that the individual attaches to the representation. The cognitive map is formed not just through perception of environmental features, but those features which are perceived are shaped by the individuals' purposes, goals, past experiences and future objectives with regard to the environment in question. Consequently, cognitive maps are likely to contain features which make the map meaningful to that individual.
3.3 Cognitive mechanisms facilitating spatial understanding
Wayfinding and navigation through familiar and unfamiliar environments can be seen as examples of spatial problem solving (Arthur and Passini, 1992). Clearly, the successful resolution of such problems is highly dependent upon the information available to the individual. However, when faced with a new environment we are bombarded with stimuli through all the senses - how do we manage to make sense of this wealth of available information and know what information to retain and what to ignore?
It has been hypothesised (Kaplan, 1973, 1976) that the cognitive mechanisms we employ during problem solving exercises such as wayfinding evolved because of their importance to the survival of the organism. Thus, our present day ability to build cognitive maps arose through this evolutionary process. Our primitive ancestors were ill equipped to fight against faster animals with better sensory capacities of sight, hearing, sense of smell and the protection afforded by claws, fangs and thick skins. To compensate, these first 'humans' had to develop systems to enable them to compete on near equal terms. Kaplan proposes that the systems which developed were oriented towards processing and making sense of the enormous amount of spatial information available. Thus, early Man evolved strategies, speed of thought and the ability to differentiate, select and process relevant cues from the much wider barrage of available but irrelevant information. Kaplan proposes that the necessity for strategy and speed of thought and action, coupled with the relative scarcity of salient information led to the evolution of the following capacities:
The organism needed the ability to anticipate future events to enable intelligent decisions to be made quickly. Furthermore, when faced with a novel situation it was necessary to be able to abstract from previous situations to enable the transfer of prior experience to the present. This in turn required the ability to generalise from situation to situation. However, whilst such abilities would be important, they alone would not be sufficient to enable competition on equal terms. The success of the species was dependent upon the ability to come up with appropriate solutions to problems never before encountered: responsible innovation. Kaplan proposed a number of mechanisms which might deal with these requirements.
3.3.1 Anticipation
The ability to anticipate events is of central importance to functioning in the environment. However, having encountered a particular event, a huge number of possible sequences could arise from its occurrence. Such a multitude of possibilities requires an economical coding system. Kaplan proposes that such economy would be best handled by a system which allows considerable overlap. This would ensure economy both of storage and also of experience: a cognitive map, built up from many different sources would enable access to the representation from a large number of experiences.
3.3.2 Abstraction: the need for a hierarchical structure
Abstraction is aided by a number of mechanisms which aim to reduce the mass of information with which the individual is faced. First, in recognising an object and creating a representation, only the object's most salient and frequently occurring features will be included. Second, the representation must enable generalisation to a large number of circumstances and not solely those which gave rise to it. Abstraction is further aided by the capacity to shift scale; we are capable of making the cognitive leap from a representation of a room in a house, to a representation of a whole town, all with little effort. Although we take this process for granted it is most important in reducing detail and enabling abstraction.
Kaplan hypothesised a hierarchical storage system (the 'concept of layers') to facilitate these processes. Recent empirical evidence (e.g. Hirtle and Jonides, McNamara, 1986; McNamara et al., 1989; Stevens and Coupe, 1978) offers strong support for the existence of such a hierarchicaI structure. It has been suggested that our spatial representations are composed of nested levels of detail, which can be represented in graph-theoretic trees, such that global and local properties of an environment can be represented at different levels of the tree (McNamara, Hardy and Hirtle, 1989). Hierarchy is clearly an essential characteristic of cognitive maps, allowing for economy of storage, flexibility of use and the ability to infer relationships between separately encoded representations. Indeed, McNamara et al., (1989) have shown that subjects tend to impose an hierarchical structure on spatial layouts to be remembered even when none is apparent, suggesting that hierarchical encoding facilitates storage and memory of larger amounts of information.
3.3.3 Representations: analogical or propositional?
One of the most important aspects of cognitive maps for DRG designers is the form taken by stored representations. Research in the area of visual imagery has long been concerned with the form in which information is internally represented. Broadly speaking, mental representations can be divided into those which are analogical and those which are propositional:
"Analogical representations tend to be images which may be either visual, auditory, olfactory, tactile or kinetic. Propositional representations are language-like representations which capture the ideational content of the mind, irrespective of the original modality in which that information was encountered." (Eysenck and Keane, 1990, p.206)
Some have argued that the form of mental representations can be best accounted for within a purely propositional framework (Pylyshyn, 1973; 1979). However, it is now generally agreed that information may be represented in both analogue and propositional formats (Kosslyn, 1980) and indeed such a system is desirable when one considers the amount and variety of information the human mind has to represent (Eysenck and Keane, 1990).
In addition to the distinction between representational formats, a further distinction has been made between types of knowledge. Facts about the world are represented as 'declarative' knowledge, whereas information about how to act is represented as 'procedural' knowledge. Thus, with regard to wayfinding and navigation, salient objects are represented in the form of declarative knowledge and enable the individual to know of, and recognise these features upon encountering them. Routes, on the other hand, being composed of a series of decision points, are represented as procedural knowledge (Gärling and Golledge, 1989).
Procedural knowledge tends to be more abstract in nature, concerned with identifying appropriate behaviour for a given situation. It is reasonable to assume that such knowledge will be more difficult to represent in an image format than objective facts about the world. However, whilst it is tempting to assume that declarative knowledge will be stored in the form of images, and procedural knowledge in the form of propositional statements, it is unlikely that such a clear cut distinction exists. Some items will be more amenable to being represented in visual imagery, while abstract concepts will perhaps rely upon propositions to encode the information Moreover, an individual's total knowledge about an item is likely to be made up of a wide range of images and propositional statements, accumulating over the course of one's experience with that item.
3.3.4 Filling in the gaps: responsible innovation and the Search mechanism
The mechanisms discussed thus far and the processes which they are considered to enable, deal with speed of action and the reduction of information. In reality, however, perhaps the most challenging cognitive exercise is one which requires intelligent decision making in the face of novel circumstances. Clearly, this is a process faced daily by travellers wayfinding through new environments. Where there is no prior experience to fall back on, the individual must make a decision which not only fulfils the requirements of the situation, but which does not endanger the individual.
The proposed 'responsible innovation' mechanism uses prior experience by seeking a solution to a novel problem utilising representations acquired through past, similar experiences. To this extent the process is not random, but it is nevertheless innovative in that it involves the creation of new connections between the present and past experience.
3.3.5 The use of heuristics
A problem usually involves an initial knowledge state and a goal knowledge state. Newell and Simon (1972) proposed a number of strategies, which they called heuristic methods or heuristics, which individuals use in problem solving to reduce the number of intermediate states required to go from initial to goal state. As wayfinding and navigation are examples of spatial problem solving it is reasonable to assume that individuals use a number of heuristic methods to enable them to reduce the quantity of information they are faced with when navigating or wayfinding. One of the most important heuristics proposed is means-end analysis (Newell and Simon, 1972) whereby the individual compares their present state with a 'goal' state and tries to solve the problem by creating sub-goals to reduce this difference.
Environmental cognition research (e.g. Byrne, 1979; Tversky, 1981; Stevens and Coupe, 1978) suggests that similar heuristic strategies are used to encode large scale space. Byrne (1979) showed that subjects estimated route distances as longer when they contained many turns than equally long routes with few turns. Byrne also found that angles remembered along a route tend to be recalled with a bias towards a right angle. Tversky (1981) identified two storage heuristics: the alignment heuristic (figures are represented as more aligned than in reality) and the rotation heuristic (figures are aligned more towards the closest cardinal point than in reality). Strategies such as these enable individuals to make generalisations across situations, retain more spatial information and abstract innovative solutions for novel problems. Additionally, they explain some of the distortions present in people's cognitive representations of an environment.
4 THE CONTENT OF COGNITIVE MAPS
4.1 What features of the environment are 'memorable'?
While it is important to know how cognitive maps are formed, DRG designers will want to know what features of an environment are retained and how these are organised in an individual's cognitive map, in order that route guidance and information can be based around the most memorable features. Cognitive maps develop not only in terms of the quantity of information they contain, but more importantly in terms of the quality of the detail contained. Indeed, developmental theorists emphasise that development is a process of increased differentiation and organisation. The development of spatial knowledge results from a discrepancy between what is observed in an environment and what has been stored in the individual's cognitive map of that environment (Hart and Moore, 1973; Moar and Carleton, 1982; Siegel and White, 1975; Thorndyke and Hayes-Roth, 1982).
The content of an individual's initial representations of a new environment have been the subject of much debate. Early theories (e.g. Siegel and White, 1975) suggested that route schemata were essential to the of cognitive maps. Our initial representations are composed of distinctive landmarks within the environment, but with very little information regarding the spatial relations between landmarks. With increased experience, more knowledge is acquired of the sequential order of landmarks along a route and eventually the routes learned may become integrated into a configurational representation of the environment, what is referred to as 'survey knowledge' (Siegel and White, 1975).
Siegel and White emphasised the importance of routes in the development of spatial knowledge, they were seen as the building blocks of cognitive maps because each route caused a new representation to be formed and through increased experience these separate representations became incorporated into one overall representation of an area. However, recent research (Byrne, 1979; Moar and Carleton, 1982) has suggested that the emphasis upon routes was perhaps overstated. Rather than having to acquire a separate schemata for each route and then combining them into a more global picture, Moar and Carleton (1982) found that if the routes intersect in an obvious manner a cognitive map may be formed which is based on the network of routes rather than separate schemata for each.
4.1.1 Distinctiveness
Clearly, some environmental features will be more memorable than others, but the extent to which these can be identified as universal constants is questionable. Early research (Lynch, 1960; Appleyard, 1969; 1970) concentrated upon the identification of distinctive features of the environment, but more recent work ((Kaplan, 1976; Gärling, Book and Lindberg, 1984; Golledge, 1978; Mainardi Peron, Baroni, Job and Salmaso, 1990) suggests that distinctiveness is not limited solely to the physical features of a landmark. Certainly a building which is visually distinctive will be perceived more readily, but this is affected by the landmark's inferred distinctiveness: a building may stand out from other features in an environment because we know something about its structure which sets it apart from its neighbours, despite visual similarities. This would imply that inferred distinctiveness is highly subjective, based upon the individual's past experiences, whereas a landmark which is truly visually distinctive would be so for all who view it. Furthermore, a building may be perceived as functionally distinctive, having some particular salience for the individual, or serving a particular function. This element of subjective salience is likely to produce a cognitive map which, in addition to features of the environment, will also include non-spatial information (Gärling, Book and Lindberg, 1984; Golledge, 1978; Hirtle and Jonides, 1985).
4.2 Factors that affect the acquisition of spatial knowledge
4.2.1 Primary versus secondary learning
A distinction has been made between primary learning (direct contact with the route which is being learnt) and secondary learning such as studying a map or a model of the route (Presson and Hazelrigg, 1984). Mode of knowledge presentation would certainly appear to play a role in the extent to which spatial knowledge develops. While a number of studies (Allen, Siegel and Rosinski, 1978; Cornell and Hay, 1984) have shown that it is difficult to obtain configurational knowledge from secondary learning alone, some recent studies have suggested that subjects can acquire good configurational knowledge as a result of a combination of secondary learning and exposure to the test environment. Hunt (1984) has shown that subjects who have experienced an environment via video simulation in conjunction with viewing a model of that environment, can develop good wayfinding abilities. In her study of nurses' knowledge of a complex hospital setting Moeser (1988) showed that nurses who had been working at the hospital for two years had developed good route knowledge but had very limited configurational knowledge. By contrast, she demonstrated that a group of subjects with no prior experience of the hospital could develop good configurational knowledge after an intensive training program involving learning floor plans of the building. Hunt's and Moeser's results have been offered as examples of situations where configurational knowledge obtained through secondary learning can be used in the real environment (McDonald and Pellegrino. 1993). However, in both cases it should be noted that, in addition to the gained from some forth of secondary learning, both Hunt's and Moeser's secondary learning groups were given additional help which could be construed as varying forms of direct or primary experience of the environment. In particular, Moeser's naive group were given a tour of the hospital, which could have helped to articulate the spatial representation they acquired through studying the floor plans in much the same way that conventional primary learning would contribute to the development of an internal representation. This objection apart, the study does demonstrate that configurational knowledge does not necessarily develop as a result of prolonged experience in an environment.
The results of studies such as those of Hunt and Moeser prompted researchers to question whether representations obtained via secondary learning might in some ways be preferable or superior to those obtained as a result of primary learning. However, other research (Evans and Pezdek, 1980; Presson and Hazelrigg, 1984; Sholl, 1987; Thorndyke and Hayes-Roth, 1982) has shown that while representations formed through secondary learning may be very precise, they are orientation specific, that is, the information thus encoded is available only in the orientation in which it was originally learned. Representations formed through primary learning, although less precise, are far more flexible and can be manipulated to allow a range of orientations. This is an important distinction in that, as Neisser (1976) argues, the main purpose of a cognitive map is as a means of orienting the individual within their environment and is therefore likely to be egocentric in its orientation. A spatial representation which permits only one perspective may be sufficient to enable a subject to score highly on experimental tests, but is of less use when that individual has to abstract new routes and relationships between parts of the environment as is required in wayfinding. One could also question the validity of comparing knowledge gained from intense, specific training programs (Moeser's naive subjects were required to learn floor plans to a level of 100% accuracy) with the natural process of spatial knowledge acquisition that accrues as a result of prolonged daily experience?
4.2.2 The purposive nature of cognitive mapping
Wayfinding and navigation can be broken down into a number of interrelated processes: the development of a plan of action; the formation of a travel plan and finally, the execution of the travel plan (Gärling, Book and Lindberg, 1986). Central to each of these processes is the information available to the individual. Information about the environment can be obtained from various sources: relevant media; word of mouth; direct observation and from the individual's cognitive map of the environment. We have seen that a number of cognitive mechanisms are employed to reduce and select information. However, one of the most important aspects of cognitive maps, and one which has all too often been overlooked by researchers from all disciplines, is the extent to which an individual's cognitive map of an environment reflects his or her purposes within that environment. In related areas, Canter (1983) emphasised the purposive nature of people's evaluations of places, Ittelson (1973) pointed out the extent to which perception directs purposeful actions, and previously Gibson (1966, 1979) argued that people perceive environments within the framework of 'affordances'- the functional significance and meaning an environment holds for the perceiver, i.e. how the environment is seen to be able to assist the individual in the completion of personal aims. Clearly, such evaluations will be highly dependent upon the individual's reasons and purposes for being in the environment: these will influence the meaning and significance the individual attaches to the particular setting and will consequently shape the individual's behaviour in that environment.
4.2.3 Environmental roles
The interactions which take place between people and environments are ongoing activities and are thus difficult to specify for any given point in time (Canter, 1977). Similarly, with respect to cognitive maps, whilst research has attempted to identify specific features which can be predicted to be included in an individual's cognitive map of an environment (reviewed by Moore, 1979), the wide variation of elements identified that this may be more difficult than was originally thought. Certainly, it would be useful for DRG designers to know which environmental features are most 'memorable', so that such features can be included when giving route directions, but the point is that such features are not universally memorable. Rather than trying to identify specific aspects of the environment or the individual which might affect environmental cognition, an alternative approach is to consider the pattern of interactions that an individual has in a particular place. It was noted earlier that much of our behaviour is strongly goal orientated, and thus we base our evaluations of an environment partly upon the extent to which it facilitates these goals. Consequently, the role that an individual has, in and with regard to, a particular setting - their environmental role (Canter, 1977) - will have a major influence on the way that they conceptualise that setting. Unlike observations of behaviour made at a specific time which might be in isolation, a person's environmental role is likely to be built up over a much longer period of time, reinforced by repeated occurrences and as such is likely to be a good predictor of their likely behaviour in that particular setting (Canter, 1986).
This is a key concept which has direct significance for DRG research. Goal seeking is an essential part of automobile travel (Appleyard, Lynch and Myer, 1964). Therefore, it is important to consider an individual's purposes with respect to the environment and the consequent affordances that environment is perceived to provide an individual, as these will naturally affect the individual's conceptions of, and behaviour in that environment. Clearly, a driver's evaluation of an environment can be expected to differ from that of a pedestrian. Similarly, in providing the driver with information about the route, features considered obvious, memorable and salient when standing at a particular vantage point may not be so considered when moving through the environment in a car travelling at speed. Additionally, recognition of purposes and affordances helps us to understand the ways in which different groups of drivers evaluate driving and the particular network on which they habitually travel. For a commuter driver, the primary aim is to reach their workplace by a specific time, and therefore the car, the network and the prevailing conditions thereon will be evaluated in terms of the extent to which these hinder or facilitate this goal. By contrast, for a driver who uses the car as part of their job, such as a taxi driver or sales representative, for a large part of the day the car is their workplace. For such drivers, their evaluation of driving and of the network will be shaped by the extent to which these make performance of the job easier or more difficult (Jackson, 1992).
A good example of the influential nature of environmental role and of the affordances offered by an environment is provided by a re-examination of the results of Moeser's (1988) study. If we consider the environmental role of each group, the reasons for the results become clearer. The environmental role of the established nurses would have been shaped by the extent to which the hospital facilitated their work. The spatial representation they formed of the environment was specifically aimed at enabling them to achieve their primary objectives in relation to that environment, consequently they obtained excellent knowledge of the routes between places which they traversed daily. As such, their representation would have been particularly strong with respect to these routes. This was sufficient for them to perform their duties efficiently. Configurational knowledge was not acquired because such a depth of knowledge was not required, indeed learning such detail could have had a detrimental effect upon their primary task, that of nursing. By contrast, the naive group - psychology students fulfilling part of their course requirements - would have had a weak environmental role. They had a clear motivation for performing well and indeed were subjected to the procedure until such time as they had obtained a perfect score, the lack of a relationship with the test environment turns the exercise into little more than a memory testing exercise for the naive subjects. The clearest point to come out of the study is that rather than indiscriminate learning taking place, individuals adopt a purposive approach to learning about their surroundings; their learning is specifically oriented to the goals which they have to achieve with respect to the environment. For the naive subjects this entailed acquiring a comprehensive configurational knowledge of the environment. For the nurses learning the spatial environment was only important in the extent to which it allowed them to competently perform their nursing duties.
4.2.4 Familiarity
Whilst Moeser (1988) demonstrates the modulating effects of environmental role on the influence of familiarity in spatial cognition, it is clear that familiarity does play a significant role in the process. As we become more familiar with an environment our experience of the area increases not just the amount of information we obtain, but also the differentiation and integration of that information. Appleyard (1970) found that with increasing length of residence (and thus increased familiarity) the use and frequency of spatially dominant maps increased. Similar results have been reported by Moore (1975) and Evans et al. (1981). More recently, a number of studies (Gale, Golledge, Pellegrino and Doherty 1990; Mainardi Peron, Baroni, Job and Salmaso 1985, 1990) have offered further support for the influential nature of familiarity and in doing so have explored the concept in more detail. Gale et al. (1990) argue that familiarity exists along four dimensions: locational knowledge, visual recognition, place name recognition and interaction with the place. A further distinction is made by Mainardi Peron et al. (1985) in differentiating between acquaintance and functional familiarity, the latter referring to repeated contact with a place for a specific purpose - goals being attained through activity occurring in the place.
5 PREDICTED RESPONSES TO DRG SYSTEMS
The question most central to an understanding of the effects of Dynamic Route Guidance and Information systems concerns the effect that additional information has upon an existing cognitive representation of an environment. Thus, before considering some of the effects that DRG will have upon the driver, the following section considers the updating of existing knowledge.
5.1 The updating process and the information supplied
When we experience an environment for the first time, the initial representation formed is highly contextualised, (McNamara, Altarriba, Bendale, Johnson and Clayton, 1989) dependent upon the specific circumstances in which the representation was acquired. As a result, the representation is unitary, parts cannot be retrieved without accessing the whole. Over time, the representation becomes decontextualised - objects within the environment are experienced in a wide range of contexts and become disassociated, perhaps leading to multiple internal representations. Thus, having experienced features within an environment over a prolonged period we are able to retrieve specific detail about one element within that environment without having to access our spatial knowledge of the wider milieu within which the feature occurred (McNamara et al, 1989). This decontextualisation is another example of the increased organisation and differentiation which characterises developing cognitive maps.
With the provision of new information two situations exist: first, where the information is additional and in keeping with existing knowledge, and second, when the information is additional but incongruent with existing knowledge. If the additional information is in keeping with that which already exists then the hierarchical structure will allow this new piece of additional detail to simply be added at the relevant level, rather like adding another 'branch' to the 'tree'.
If the additional detail is incongruent with existing information, however, in that it contradicts some aspect of the existing knowledge, the individual needs to evaluate the new information. This evaluation process will be shaped by a variety of situational factors, and particularly by the individual's familiarity with the environment in question. Familiarity will produce a strong, well formed representation which will be resistant to change, particularly in response to a single piece of contradictory information. The information is therefore likely to be rejected. Unless the new information is of obvious immediate and long term importance (e.g. a major permanent change to the road layout) the existing representation will be updated only when the same piece of contradictory information occurs repeatedly.
It is reasonable to assume that, when drivers are provided with additional contradictory information, their perceptions of the information will determine their willingness to respond to and accept the information. Behavioural research on drivers' responses to traffic information in a number of associated areas (e.g. parking guidance information, roadside displays, radio traffic broadcasts) strongly supports this idea. Such research shows that information should be clear and easy to comprehend when moving at speed (Allen, 1993). If it is difficult to process the information or advice contained in a message drivers are likely to ignore the advice altogether (Bonsall, 1992). Above all the information should be accurate, reliable and credible (Shirazi, Andersen and Stesney, 1988; Polydoropoulou, Ben-Akiva and Kaysi, 1994) and therefore needs to be frequently updated to enable it to respond to changes in the network. In addition, information should be perceived as salient to the individual's situation, it is therefore essential that the system menu accurately reflects driver needs and is flexible enough to be adaptable to a wide range of different user's requirements (Bonsall, 1991b).
To enable an explanation of the possible effects of route guidance systems the present paper considers the various benefits which have been predicted for DRG and compares the performance of these tasks with and without such assistance. In this way it is possible to compare both the positive and negative effects of carrying out complex spatial tasks in both conditions. In the short to medium term DRG systems aim to offer three particular services: route planning; route guidance and wayfinding; and traffic and congestion information. The following section considers each of these functions with and without the benefit of DRG in light of the above review.
5.2 Route planning
5.2.1 Without DRG
In pre-planning a trip without DRG some secondary source will be employed which, in the absence of direct route experience will provide a good overview of the journey, but will not be encoded in sufficient detail to allow navigation of the route. However, if the individual has some generic experience or knowledge of a destination in the same geographical area, the use of the secondary source might help the abstraction process, enabling the driver to recall representations gained from the previous trip and to abstract possible junctions which might be those represented on the map.
From a psychological perspective, the main purpose of such pre-planning is to reduce stress and anxiety caused by having to negotiate an unknown route. Certainly, a number of physiological studies have identified traffic situations of high complexity as one of the more stressful aspects of driving, associated with increased heart rate and high blood pressure (Littler, Honour and Sleight, 1973; Rutley and Mace, 1970). It is reasonable to assume that negotiating an untravelled route is of comparable complexity and hence is comparably stressful.
5.2.2 With DRG
A driver pre-planning a route with the assistance of DRG could take one of a number of courses of action. At one extreme he could rely exclusively upon the information provided by the route guidance system. At the other end he could choose to ignore such information. Somewhere in the middle lies the likely course of action that an individual will take until such time as they have sufficient positive or negative experience of the system to enable a decision to be made about its subjective usefulness. Until then it is likely that the individual will use the system as an aid in addition to their own experience, maps and other secondary sources of information. Assuming that the system has the capacity to display the most suitable route (based on prevailing road conditions at the time of enquiry) then the presentation format will be of importance (Dingus and Hulse, 1993; Wierwille, 1993). In addition, the system's potential for anxiety and stress reduction is clearly related to its success on the first few trials, if the information provided is perceived to be useful then the driver will gain confidence in the system and anxiety and driver stress will be reduced.
5.3 En route guidance
5.3.1 Without DRG
En route guidance and wayfinding without DRG is largely through roadside information displays. The ability of a driver to process such information is dependent upon the design of the display as well as the nature of the information presented. The driver is faced with having to select correct routes at high speed and often with little preparation. Such situations are potentially stressful and dangerous, particularly on motorways or in 'bumper-to-bumper traffic'. The only alternatives for a driver seeking route guidance are: to stop and ask directions; consult another source such as a map; or to continue on with their journey in the hope that future signposts will indicate a means of getting to the correct route. Evidently, en route guidance assistance can be a most stressful driving experience. This would suggest that such an experience will be well imprinted in memory, in addition, research by Evans, Skorpanich, Gärling, Bryant and Bresolin (1984) indicates that stress itself can affect environmental cognition. The degree of anxiety which might be caused by such situations is highly dependent upon the driver's reasons for making the journey; losing one's way en route to a job interview, for example, is likely to be far more stressful than becoming lost while on a touring holiday. In this sense there is an element of purposive evaluation attached to wayfinding exercises, the driver evaluates each new environment to see whether it affords opportunities to enable completion of a predetermined task; in this case wayfinding.
In summary, wayfinding and en route guidance without DRG can be characterised as high-stress, anxiety causing and potentially dangerous in terms of the amount of time spent searching for route information rather than attending to the road and surrounding traffic. At the same time, however, the stressful nature of the situation is likely to make it particularly memorable or meaningful and thus a side effect might be the development of a strong cognitive representation of the area.
It should not be assumed that all effects are necessarily negative, however. It has been argued (Winston, 1982) that satisfaction can be derived firstly from doing an activity (process utility) and secondly from having done the activity (goal utility). With regard to wayfinding it is possible that there is a sense of achievement and personal satisfaction gained from both the planning and successful execution of a wayfinding exercise.
5.3.2 With DRG
One of the main effects to be considered when drivers use DRG for wayfinding is the effect that it will have upon cognitive map formation. Of all the tasks which a driver has to perform, none articulate and develop their cognitive representation of the spatial environment more than navigating one's way through that environment (Carr and Schissler, 1969; Siegel and White, 1975). Even in non-stressful situations the experience of directly orientating and navigating oneself through the environment contributes to the formation of configurational knowledge. Such direct contact with the environment helps to develop simple route knowledge and contributes to the overall representation formed of the surrounding environment. With the assistance of DRG it is questionable whether this same intimacy with the environment would be retained.
The processes described by Kaplan: abstracting from existing knowledge; generalising from this knowledge to the new situation; searching existing knowledge for similar patterns; and coming up with new solutions to the present problem, are all crucial in helping to articulate the cognitive representation of the new environment. It is therefore probable that an individual's spatial representations of the environment would suffer as a result of DRG usage. The situation could perhaps be compared to the debate which surrounds the move away from books to television as primary source of entertainment for children over the last fifty years, instead of relying upon the imagination to conjure up images, the images are presented in their entirety, little effort is required of the person receiving them, be they seated in front of a television screen or a windscreen.
It could be argued that, even if reliance upon DRG does create impoverished spatial representations, such a situation is hardly a disaster. However, the process of building up cognitive representations clearly has positive effects on the individual. It is therefore necessary to consider the peripheral influence of DRG upon the positive effects that the driver gains from task completion. Will the satisfaction arising from having achieved something be reduced because of a reliance upon DRG and not upon internal cognitive processes? In addition, the positive effect on an individual's self image and confidence that comes as a result of developing one's spatial knowledge should be considered. In all these examples it is not that DRG will have a negative effect, but that reliance upon DRG in preference to natural processes may negate some of the positive effects. Whilst of little consequence on their own for the success or failure of DRG such factors may contribute to the driver's overall impression formed of DRG.
The success of wayfinding using DRG, like pre-planning of a route, will depend upon the driver's perception of the information presented as valid, reliable, consistent, accurate and above all salient to that driver's particular circumstances. To a large extent this will be shaped by the driver's purposes for taking a specific route. This suggests that the information supplied to drivers will need to be very flexible to cope with the range of circumstances to which it might be applied. If the driver perceives the information to be of use then he might choose to follow the directions suggested. If these perceptions then prove to be accurate then it is reasonable to assume that stress and anxiety will be reduced as a consequence. It is still likely that in the initial stages the driver will be uncertain about the information, and may wish to rely upon prior experience rather than on a piece of equipment. Indeed, it could also be argued that the anxiety and stress which arose from ignorance of the route might be replaced by anxiety and stress as a result of relying upon an unfamiliar and untried system.
5.4 Traffic and congestion information
5.4.1 Without DRG
One of the most stressful experiences that the driver has to encounter is being delayed in congested traffic, be it on a major motorway or on urban streets, particularly when on the way to or from work, meetings etc. Without DRG such situations are likely to cause considerable stress, frustration, anxiety and anger, although the extent to which drivers suffer these problems is dependent upon characteristics of their journey (Novaco, Stokols and Milanesi 1990; Jackson, 1992), their ability to adopt mechanisms to enable them to cope with the stress caused (Gulian, Debney, Glendon, Davies and Matthews, 1989; Jackson, 1992) as well as a range of individual factors. For alternatives to the congested route, the driver must either rely upon traffic reports from radio, television or other media source (e.g. computer networks and cable TV stations dedicated to traffic information in the USA) or upon previous experience of the network and knowledge of alternate routes which might enable a diversion away from the congestion. Media reports are problematic in that the information, being based on real time reports from helicopters or light aircraft, is often no longer valid by the time it has been received. Experienced drivers tend to rely upon previous experience, gradually building up a cognitive representation of the surrounding area and side roads which connect various parts of the network. This reliance upon "rat-runs" has prompted many local authorities to make such roads 'no-entry' areas or to design-in bumps along the side route to cut down on speeds and reduce the route's attractiveness as an alternative. With regard to the urban traffic situation, therefore, it could be argued that drivers are faced with ever-decreasing alternatives: the few that do exist are likely to be equally busy. Assuming that the driver does know of an alternative, without the benefit of DRG he is faced with the decision to either stay on the present road and hope that the situation improves, or to try the alternative, aware that it could be worse than the existing route.
DRG's predicted benefits are partially based upon the assumption that in congestion situations drivers do not change route because they are ignorant of alternatives. However, research (Mannering, 1989), suggests that in such situations drivers may well be aware of alternative routes, but simply prefer to stay on the congested route. Drivers are simply unwilling to run the risk of switching routes with no guarantee of success, when they know that the time taken on the habitual route will always fall between certain extremes (e.g. 15 minutes on a good day, up to 45 minutes on a bad day). It would appear that DRG's main enemy might well be the habitual nature of drivers' behaviour- a number of studies (E.g. Golledge and Zannaras, 1973; Huff and Hanson, 1986) imply that, having established a small set of viable alternatives the driver gives up any further search activity and settles into this pattern of habitual behaviour. In keeping with the theory of boundedly rational behaviour (Simon, 1955) drivers 'satisfice' rather than 'optimise' their behaviour (Mahmassani, Chang and Herman, 1986).
Recent research (Bonsall et al., 1991; Jackson, 1992) might help to explain why this is so. Jackson (1992) showed that one of the main attractions of the car over other modes of commuter transport was the perceived control which this mode provides - control over departure time, route taken, as well as control over the immediate environment - the car is the driver's own personal space. Results of surveys reported in Bonsall et al. (1991) support this idea: drivers taking familiar routes preferred to be given information rather than guidance, clearly they felt confident in their own abilities and were unwilling to hand over control to a machine. When unfamiliar with the network, however, they were happy to be given guidance - on unfamiliar routes the driver's perceived control is much reduced and consequently additional information is a means of gaining more control over the situation. Thus, one of the problems for DRG is that although it aims to increase personal control over trip decisions, in familiar networks drivers may perceive it to be taking away some of this control - resulting in a reluctance to rely upon the equipment.
If we consider in more detail the dilemma with which the driver is faced it is possible to see why drivers are so unwilling to change routes. In a congestion situation let us say that the driver is faced with a choice between route A, the present, obviously congested route and route B. a known alternative which, not being the driver's habitual route must in normal circumstances be perceived by the driver to be inferior to route A. The driver knows how congested route A is at that moment in time, previous experience and familiarity with route A might also give him some indication of the likely delay time. The present and immediate future state of Route B. on the other hand, is unknown. If the driver chooses to take route B and it is found to be significantly less congested than A then the driver will certainly feel very positive about the decision (so much so that it might prompt him to reassess the status of the two routes). However, if route B is found to be as congested or even worse than A then the driver will certainly suffer more frustration and stress than would have been the case had he done nothing. Extending from the idea that we gain satisfaction from successfully completing a task, if we undertake a task as a proposed solution to a stressful situation, and it proves unsuccessful to the point that it makes our situation even worse, then it is reasonable to assume that the result would be even more stress. Not only has the driver suffered the stress arising from being in a congestion situation, but he has also suffered the additional stress of having made (in his mind) a wrong decision; what could be referred to as 'missed opportunity stress' - remaining on the habitual route would not have been so bad after all. It is possible that drivers go through an abbreviated version of this process, recalling previous negative experiences and the resultant impact upon affective state, to rationalise the decision to do nothing.
5.4.2 With DRG
The challenge for DRG is not simply to provide drivers with knowledge of existing alternative routes, but to provide them with sufficient detail about the current state of those alternatives with regard to the driver's present route. It is erroneous to assume that drivers are unaware of alternatives. From an information processing approach one could say that a driver who has habitually travelled the same route for a number of years will have abstracted from existing knowledge, generalised from prior experience of route structure and therefore will certainly have developed a cognitive representation of the likely spatial arrangement of connecting and adjacent roads. It is perhaps more accurate to say that drivers lack accurate knowledge about the current conditions of these alternatives, and are therefore unwilling to change routes.
The above discussion emphasises the need for DRG information to be accurate and up to date, describing the present situation both on alternative routes and, to enable comparison, the driver's present route. It also stresses the usefulness of information provision regardless of whether its content is negative or positive. Clearly, if alternative routes are less congested then information about their condition might prompt the driver to change routes, resulting in the benefits which have been predicted both for the individual and for the network as a whole. However, if the situation on alternative routes is worse than the driver's present route it is possible that the driver will use the information to rationalise his own situation: he may feel that, although his situation is bad, at least it is not as bad as it would be if he were on the alternative route. In this way the sense of helplessness that a driver feels when stuck in heavy traffic (particularly when on a motorway where there are fewer opportunities for route changing) might be partially alleviated.
6 CONCLUDING COMMENTS
It has been suggested that "the potential for expanding and enriching cognitive maps through ATIS (Advanced Traffic Information Systems) may be substantial" (Schofer et al., 1993 p.108) The present paper would argue that this is an erroneous assessment of the effect of DRG upon driver's spatial cognition, in that it assumes that receiving the information via DRG will have the same effect as would gaining such information via direct contact with the environment. It is correct to assume that DRG will have an effect upon environmental cognition, but rather than expanding drivers' cognitive maps, research cited suggests that providing route knowledge by DRG could actually hinder the development of driver's cognitive representations of their spatial environment. Research has shown that gaining spatial knowledge via DRG is a form of secondary learning and thus differs from the kind of knowledge gained by primary learning (i.e. direct contact).
In order to predict the possible effects that information provision might have upon drivers we need to understand the processes which are in action when a driver acquires and uses spatial knowledge under normal circumstances, i.e. without the benefits of DRG. The review has highlighted the value of trying to understand the cognitive processes which facilitate our acquisition of spatial knowledge, and has reviewed some of the mechanisms which have been suggested as means' of processing and organising such information. With the benefit of this knowledge it has been possible to make some tentative predictions as to the effects that DRG might have upon drivers, with the ultimate aim of assessing the likelihood that drivers will use DRG. From these predictions it is clear that drivers' continued use of DRG will not depend simply upon its capacity to provide information about upcoming congestion. Indeed, it could be argued that to have any chance of gaining widespread acceptance DRG must provide drivers with knowledge of the traffic conditions both on their own route and all possible alternatives heading for the same destination. In short, the characteristic which will have most influence on a driver's decision whether or not to utilise information provided will be its salience to the individual driver's situation and therefore DRG will need to be flexible enough to cope with a wide range of demands which might be placed upon it, determined by the particular driving patterns of its users.
The present paper has argued that driver's willingness to accept and use DRG will in part depend upon their perception of DRG as a useful driving aid. This assessment will be quickly formed, probably as a result of a fairly limited number of trials. A number of factors will contribute to the driver's overall perception of DRG, it being shaped not only by the obvious tangible benefits such as reductions in time and distance travelled, but also by a combination of the positive and negative psychological effects which DRG will have upon the individual. Moreover, drivers' tendency towards habitual route choice would imply that the savings projected for these more obvious benefits will be insufficient reason on their own for drivers to respond to the information provided. As a consequence, greater consideration of the possible psychological benefits of DRG, such as a means of stress relief, are urged. Meanwhile, on the negative side, having noted the positive feelings of increased self confidence and satisfaction that arise through the process of developing one's spatial knowledge, it is argued that DRG use could result in a reduction in such benefits and that this might have an impact on one of the aspects of driving which make it the popular mode of transport that it is today.
Stergiou and Stathopoulos (1989) expect the first main users of DRG systems to be professional drivers such as truck and taxi drivers, and suggest that private drivers will be more hesitant in DRG use. This is based on the assumption that professional drivers will derive greater benefits from DRG by the nature of their job and the consequent time spent driving. However, the present review would suggest that these two groups will be more unwilling than most to rely upon DRG for information, preferring to use their own knowledge based on years of experience. Professional drivers, more than most, evaluate their driving environment with respect to their purposes, aims and goals. The network will be perceived in terms of the opportunities it offers them to achieve goals, with the consequence that the cognitive representations they form will be structured by their purposes. Moreover, a taxi driver's representation is not only highly detailed but
"has superimposed upon it the flux of traffic as this is governed by the time of day and day of the week." Gladwin, (1970) p.224
Armed with such highly developed cognitive representations, taxi drivers are perhaps the least likely group of drivers to use DRG.
Clearly, a program of systematic research is required in the field to test out these many predictions, although as Gotts and Bonsall (1993) point out, ethical and safety considerations make such studies problematic. Further survey work should be undertaken which assesses drivers' information requirements: the present review would suggest that the most effective approach for DRG would be to allow drivers to select information salient to themselves from a menu of available databases. There is a need for experimentation which studies driver behaviour with the aid of DRG over a period of time and in as natural a setting as possible, the ADVANCE program currently in progress and using large numbers of real drivers in field experiments may, if behavioural scientists have sufficient involvement, provide data which would be invaluable in answering some of the questions raised above.
7 ACKNOWLEDGEMENTS
This paper has benefitted from the advice, guidance and comments of my Ph.D supervisor, Dr. Kay Axhausen. I am grateful to him and also to Professor Tommy Gärling of the University of Umea, Sweden for reading and commenting upon earlier drafts of the paper. I would also like to thank Peter Bonsall of the University of Leeds, Institute for Transport Studies and Robert French for their assistance in locating specific conference papers. The opinions contained within this paper, however, are solely those of the author.
8 REFERENCES
ADVANCE (1990) Scope, feasibility and cost of a dynamic route guidance system demonstration, Final Report to Illinois Department of Transportation.
Allen, G.L., A.W. Siegel and R.R. Rosinski (1978) The role of perceptual context in structuring spatial knowledge, Journal of Experimental Psychology: Human Learning and Memory, 4 (6) 617-630.
Allen, P.A. (1993) Driver response to parking guidance and information systems, Traffic Engineering and Control, 34 (6) 302-307.
Appleyard, D. (1969) Why buildings are known: a predictive tool for architects, Environment and Behavior, 1 (2) 131-156.
Appleyard, D. (1970) Styles and methods of structuring a city, Environment and Behavior, 2 (1) 100-118.
Appleyard, D., K. Lynch, and J. Myer (1964) The View from the road, Cambridge, Mass; MIT Press.
Arthur, P. and R. Passini (1992) Wayfinding: People, Signs and Architecture, McGraw-Hill, New York, NY.
Ben-Akiva, M.A., A. de Palma and I. Kaysi (1991) Dynamic network models and driver information systems, Transportation Research, 25A (5) 251-266.
Bonsall, P.W. (1991a) Response to route guidance, Paper presented at the 23rd Annual Conference of the Universities' Transport Study Group, University of Nottingham, January 1991.
Bonsall, P.W. (1991b) The implications of findings on drivers' requirements for and response to, route guidance, Report prepared as part of the DRIVE Project V1011 Integration of Dynamic Route Guidance and Traffic Control Systems ("CAR-GOES") (Deliverable 21)
Bonsall, P.W. (1992) The influence of route guidance advice on route choice in urban networks, Transportation, 19 (1) 1-23.
Bonsall, P.W., L. Pickup and A. Stathopoulos (1991) Measuring behavioural responses to road transport informatics, in Advanced Telematics in Road Transport, 1457-1487, Elsevier, Amsterdam.
Bonsall, P.W. and T. Parry (1991) Using an interactive route-choice simulator to investigate drivers' compliance with route guidance advice, Transportation Research Record 1036, pp. 59-68.
Bovy, P.H.L. and E. Stem (1990) Route Choice: Wayfinding in Transport Networks, Kluwer Academic Publishers, Dordecht.
Byrne, R. (1979) Memory for urban geography. Quarterly Journal of Experimental Psychology, 31(2) 147-154.
Canter, D.V. (1977) The Psychology of Place, Architectural Press, London.
Canter, D.V. (1983) The purposive evaluation of places: a facet approach, Environment and Behavior, 15 (6) 659-698.
Canter, D.V. (1986) Putting situations in their place: Foundations for a bridge between Social and Environmental Psychology, in A Furnham (Ed.) Social Behaviour in Context, 208-239, Allyn and Bacon, Boston.
Carr, S. and D. Schissler (1969) The city as a trip: Perceptual selection and memory in the view from the road, Environment and Behavior, 1 (1) 7-35.
Conquest, L., J. Spyridakis, M. Haselkom and W. Barfield, (1993) The effect of motorist information on commuter behavior: Classification of drivers into commuter groups, Transport Research C, 1 (2) 183-201.
Cornell, E.H. and O.H. Hay (1984) Children's acquisition of a route via different media, Environment and Behaviour, 16 (5) 627-641.
Dingus, T.A. and M.C. Hulse (1993) Some human factors design issues and recommendations for automobile navigation information systems, Transportation Research C 1(2) 119-131.
Downs, R.M., and D. Stea (Eds.) (1973) Image and Environment' Cognitive Mapping and Spatial Behavior, Aldine, Chicago.
Emmerink, R.H., K.W. Axhausen, P.Nijkamp, P. Rietveld (1993) Effects of information in road transport networks with recurrent congestion, Working Paper No. LSEOR.93.9 Operational transport networks with recurrent congestion, Paper No. LSEOR.93.9 Operational research, London School of Economics and Political Science, December 1993.
Evans, G.W. and K. Pezdek (1980) Cognitive mapping: Knowledge of real world distance and location information, Journal of Experimental Psychology: Human Learning and Memory, 6 (1) 13-24.
Evans, G.W., M.A. Skorpanich, T. Gärling, K.J. Bryant and B. Bresolin (1984) The effects of pathway configuration, landmarks and stress on environmental cognition, Journal of Environmental Psychology, 4 (4) 323-335.
Eysenck, M.W. and M.T. Keane (1990) Cognitive Psychology: a student's handbook, Lawrence Erlbaum and Associates, London.
Gale, N., R.G. Golledge, J.W. Pellegrino and S. Doherty (1990) The acquisition and integration of route knowledge in an unfamiliar neighbourhood, Journal of Environmental Psychology, 10 (1) 3-25.
Gärling, T., A. Böök, and E. Lindberg (1984) Cognitive mapping of large-scale environments, action plans, orientation and their interrelationships, Environment and Behavior, 16 (1) 3-34.
Gärling, T., A. Böök, and E. Lindberg (1986) Spatial orientation and wayfinding in the designed environment: a conceptual analysis and some suggestions for postoccupancy evaluation, Journal of Architectural Planning and Research, 3 (1) 55-64.
Gärling, T. and R.G. Golledge (1989) Environmental perception and cognition, in E. H. Zube and G. T. Moore (Eds.), Advances in environment, behavior and design (Volume 2, pp.203-226), Plenum, New York.
Gibson, J.J. (1966) The Senses Considered as Perceptual Systems, Houghton and Mifflin, Boston.
Gibson, J.J. (1979) The Ecological Approach to Visual perception, Houghton and Mifflin, Boston.
Gladwin, T. (1970) East is a big bird, Harvard University Press, Cambridge, MA.
Golledge, R.G. (1978) Learning about urban environments, in T. Carlstein, D. Parkes and N. Thrift (Eds.), Timing space and spacing time (Volume 1, pp.76-89) Arnold, London.
Golledge, R.G. (1987) Environmental Cognition, in D. Stokols and I. Altman (Eds.) Handbook of Environmental Psychology, 1, 131-174, Wiley, New York.
Golledge, R.G. and G. Zannaras (1973) Cognitive approaches to the analysis of human spatial behavior, in W.H.Ittelson (Ed.) Environment and Cognition, 59-94, Seminar, New York.
Gotts N.M. and P.W. Bonsall (1993) A new technique for investigating drivers' response to In-Vehicle Route Guidance Information (IVRGI) Systems, Paper presented at the 25th Annual Conference of the Universities' Transport Study Group, University of Southampton, January 1993
Gulian, E., L.M. Debney, A.I. Glendon, D.R. Davies and G. Matthews (1989) Coping with driver stress, in F.J. McGuigan, W.E. Sime and J. MacDonald Wallace, (Eds.), Stress and Tension Control, 173-186, Plenum Press, New York.
Hart, R.A. and G.T. Moore (1973) The development of spatial cognition: a review, in R.M.Downs and D.Stea (Eds.) Image and Environment: Cognitive Mapping and Spatial Behavior, 246-288, Aldine, Chicago.
Hirtle, S.C. and J. Jonides (1985) Evidence of hierarchies in cognitive maps, Memory and Cognition, 13 (3) 208-217.
Huff, J.O. and S. Hanson (1986) Repetition and variability in urban travel, Geographical Analysis, 18 (2) 97-114.
Hunt, M.E. (1984) Environmental learning without being there, Environment and Behavior, 16 (3) 307-334.
Ittelson, W.H. (1973) Environment perception and contemporary perceptual theory, in W.H. Ittelson (Ed.), Environment and Cognition, Seminar, New York.
Jackson, P.G. (1992) Car commuting, its benefits and disbenefits: a preliminary investigation of the theory of Incongruent Impedance, unpublished M.Sc. Thesis, University of Surrey.
Jackson, P.G. (1994) Behavioural responses to RTI, Working paper, Centre for Transport Studies, Imperial College, London.
Kaplan, S. (1973) Cognitive maps in perception and thought, in R.M. Downs, and D. Stea (Eds.) (1973) Image and Environment: Cognitive Mapping and Spatial Behavior, 63-78, Aldine, Chicago.
Kaplan, S. (1976) Adaptation, structure and knowledge, in G.T. Moore and R.G. Golledge (Eds.) Environmental Knowing, 32-45, Dowden, Hutchinson and Ross, Stroudsberg, PA.
Khattak, A., A. Kanafani and E. Le Colletter (1994) Stated and reported route diversion behaviour: Implications on the benefits of ATIS, Paper presented to the 73rd Annual Meeting of the Transportation Research Board, Washington, D.C., January 1994.
Kosslyn, S.M. (1980) Image and Mind, Harvard University Press, Cambridge, MA.
Littler, W.A., A.J. Honour and P. Sleight (1973) Direct arterial pressure and electrocardiogram during motor car driving, British Medical Journal, 2 (3) 273-277.
Lynch, K (1960) Image of the City, MIT Press, Cambridge, MA.
Mahmassani, H.S., G.L. Chang and R. Herman (1986) Individual decision and collective effects in a simulated traffic system, Transportation Science, 20 (4) 258-271.
Mainardi Peron, E., M.R. Baroni, R. Job and P. Salmaso (1985) Cognitive factors and communicative strategies in recalling unfamiliar places, Journal of Environmental Psychology, 5 (4) 325-334.
Mainardi Peron, E., M.R. Baroni, R. Job and P. Salmaso (1990) Effects of familiarity in recalling interiors and external places, Journal of Environmental Psychology, 10 (3) 255-271.
Mannering, F.L. (1989) Poisson analysis of commuter flexibility in changing routes and departure times, Transportation Research B, 23 (1) 53-60.
McDonald, T.P. and J.W. Pellegrino (1993) Psychological perspectives on spatial cognition, in T. Gärling and R.G. Golledge (Eds.) (1993) Behavior and Environment. Psychological and Geographical Approaches, Advances in Psychology 96,47-82, Elsevier Science Publishers B.V., Amsterdam.
McNamara, T.P. (1986) Mental representations of spatial relations, Cognitive Psychology 18 (1) 87-121.
McNamara, T.P., J. Altarriba, M. Bendele, S.C. Johnson and K.N. Clayton (1989) Constraints on priming in spatial memory: Naturally learned versus experimentally learned environments, Memory and Cognition 17 (4) 444-453.
Moar, I. and L.R. Carleton (1982) Memory for routes, Quarterly Journal of Experimental Psychology 34A (3) 381-294.
Moeser, S.D. (1988) Cognitive mapping in a complex building, Environment and Behavior, 20 (1) 21-49.
Moore, G.T. (1975) The development of environmental knowing: an overview, in D. Canter and T. Lee (Eds.) Psychology and the built environment, Architectural Press, London.
Moore, G.T. (1979) Knowing about environmental knowing; the current state of theory and research on environmental cognition, Environment and Behavior, 11 (1) 33-70.
Moore, G.T. and R.G. Golledge (1976) Environmental Knowing: Concepts and Theories, in G.T. Moore, and R.G. Golledge (Eds.) Environmental Knowing; Theories, Research and Methods, 324, Dowden, Hutchinson and Ross, Stroudsburg, Pa.
Neisser, U. (1976) Cognition and Reality: Principles and implications of cognitive psychology, Freeman, San Francisco
Newell, A. and H.A. Simon (1972) Human problem solving, Prentice Hall, Englewood Cliffs, NJ.
Novaco, R., D. Stokols and L. Milanesi (1990) Objective and subjective determinants of travel impedance as determinants of commuting stress, American Journal of Community Psychology, 18 (2) 231-257.
Polak, J.W. and P. Jones (1991) Assessing the responses of travellers in Birmingham and Athens to in-home pre-trip information, EURONETT Deliverable 27, Transport Studies Unit, University of Oxford.
Polydoropoulou, A., M. Ben-Akiva and I. Kaysi (1994) Revealed preferences models of the influence of traffic information on drivers' route choice behavior, Paper presented to the 73rd Annual Meeting of the Transportation Research Board, Washington, D.C., January 1994.
Presson, C.C. and M. Hazelrigg (1984) Building spatial representations through primary and secondary learning, Journal of Experimental Psychology: Learning, Memory and Cognition, 10 (4) 716-722.
Pylyshyn, Z. (1973) What the mind's eye tells the mind's brain, Psychological Bulletin, 80 (1) 1-24.
Pylyshyn, Z. (1979) Imagery theory: Not mysterious - just wrong, Behavioural and Brain Sciences, 2, 561-563.
Rutley, KS. and D.G.W. Mace, (1970) Heart rate as a measure in road layout design, Research Report No. RRL-LR 347, Road Research Laboratory, British Ministry of Transportation, London.
Schofer, J.L., A. Khattak and F.S. Koppelman (1993) Behavioral issues in the design and evaluation of Advanced Traveller Information Systems, Transport Research C, 1 (2) 107-117.
Sholl, M.J. (1987) Cognitive maps as orienting schema, Journal of Experimental Psychology: Learning, Memory and Cognition 13 (4) 615-628.
Shirazi, E., S. Anderson and J. Stesney (1988) Commuters' attitudes towards traffic information systems and route diversion, Transportation Research Record 1168.
Siegel, A. W. and S.H. White (1975) The development of spatial representation of large-scale environments. In H.W. Reese (Ed.) Advances in Child Development and Behavior, 10 (1) 9-55, Academic Press, New York.
Simon, H.A. (1955) A behavioral model of rational choice, Quarterly Journal of Economics, 69 (1) 99-118.
Stergiou, B. and A. Stathopoulos (1989) Traffic models and Road Transport Informatics (RTI) systems, Traffic Engineering and Control, 30 (12) 580-586.
Stevens, A. and P. Coupe (1978) Distortions in judged spatial relations, Cognitive Psychology, 10 422-437.
Thorndyke, P.W. and B. Hayes-Roth (1982) Differences in spatial knowledge acquired from maps and navigation, Cognitive Psychology, 14 (4) 560-589.
Tolman, E.C. (1948) Cognitive maps in rats and men, Psychological Review, 55 (4) 189-208.
Tversky, B. (1981) Distortions in memory for maps, Cognitive Psychology 13, 407-433.
U.S. Department of Transportation (1989) Discussion Paper on Intelligent Vehicle-Highway Systems.
Watling, D. (1993) Urban Traffic network models and dynamic driver information systems, Transport reviews (Forthcoming).
Wenger, M., J. Spyridakis, M.D. Haselkorn, W. Barfield and L. Conquest (1990) Motorist behavior and the design of motorist information systems, Transportation Research Record, 1281, 159-167.
Wierwille, W.W. (1993) Demands on driver resources associated with introducing advanced technology into the vehicle, Transportation Research C, 1 (2) 133-142.
Winston, G.C. (1982) The timing of economic activities: I;irms, households, and markets in timespecific analysis. Cambridge University Press, Cambridge.
Wootton, J., M. Ness and R.S. Burton (1981) Improved direction signs and the benefits of road users, Traffic Engineering and Control, 22 (5) 264-268.