Behaviour Monitoring and Interpretation: Facets of BMI Systems

Behaviour Monitoring and Interpretation: Facets of BMI Systems

Björn Gottfried (University of Bremen, Germany)
DOI: 10.4018/978-1-61692-857-5.ch021
OnDemand PDF Download:


The body of this chapter is structured in the following way. An initial example is presented in Section 1. It makes clear what BMI is about from the point of view of an application example. Simultaneously, a framework is outlined that derives from this example. It will be shown how other application examples fit into this very same framework in the following sections. Each of those following examples introduces typical facets of BMI systems: spatial scales are relevant in that each object to be monitored is found at a specific spatial scale (Section 2); direct and indirect observations are to be distinguished since objects are either directly observed or indirectly by means of changes that occur in the environment (Section 3); monitoring processes either occur in reality or in virtual spaces or in mixed reality scenarios (Section 4); behaviours are either purposeful and active, or they reflect typical everyday behaviours (Section 5); it can be distinguished whether a monitoring system works by means of deploying local or allocentric techniques (Section 6); eventually, similar as each application is found at a specific spatial scale, temporal scales are to be distinguished considering both durations and the speed with which observed behaviours are carried out (Section 7). A discussion section closes this chapter with a couple of issues (Section 8): different purposes of BMI applications are identified; their relationships to related areas, namely Ambient Intelligence (AmI) and Smart Environments (SmE) is discussed; the importance of AI methods is pointed out; ethical issues are considered. Finally, an outlook on future work is presented (Section 9).
Chapter Preview

The Walking Behaviours Of Pedestrians

One of the most fundamental behaviours of both humans and animals consists in getting from one place to another place. In order to monitor such locomotion behaviours a whole bunch of technologies are available today (Millonig et al. 2009). Furthermore, several authors propose different methods in order to evaluate sets of positional data, i.e. locomotion behaviours. The purpose of such investigations varies from studying which paths people take (Andrienko et al. 2007) to how their particular walking style looks like (Millonig et al. 2008). We shall initially present a study which is about walking behaviours of pedestrians and about the walking styles that can be distinguished when observing them going shopping.

A Study on Walking Behaviours

In their project, (Millonig et al. 2008) are aiming at the classification of pedestrian walking behaviours. Their study does not rely on positional data alone, but considers further factors that might influence walking styles, e.g. emotional or life style related factors. Their motivation is that mainly in urban areas people are interested in the use of mobile tools for wayfinding combined with location based services in order to coordinate activities. Then, the following problem is identified. What is called an optimal route or useful information might significantly differ among people. Therefore, a tool which can be judged to work successful, should tailor service information to the individual context of the user. In their study the authors follow the assumption that a comprehensive investigation of walking behaviours will reveal specific patterns which are typical for different individuals. Such patterns can later be employed as essential parameters for determining the individual context of a pedestrian navigation and location based service tool. Since there are endless walking patterns of pedestrians conceivable, they confine their study in an appropriate way by addressing just those pedestrians who are going shopping. It is then the aim to look for whether typical shopping patterns can be distinguished, and hence, whether people can be categorised into different types of shoppers.

In order to analyse the observed patterns they employ several qualitative-interpretative and quantitative-statistical methods. For simplicity, we summarise their methodology as follows. First, people are observed without knowing anything about it; for this purpose a hand-tracking tablet computer is used by the observer who tracks the pedestrians under observation. Second, the people who have been observed get interviewed in order to let them tell a little bit about themselves, their intentions and social background. Finally, they get further tracked, this time equipped with a Bluetooth Smartphone or a GPS logger for indoor and outdoor tracking, respectively. The acquired data is basically analysed by using speed histograms and by clustering the obtained trajectories. As a result there are three types of shoppers identified, namely swift, convenient, and passionate shoppers. These types of shoppers can be easily employed in a pedestrian navigation and location based service tool.

Complete Chapter List

Search this Book: