Analysis of movement data is a new trend of research in GIScience. Recent and emerging positioning technologies and wearable sensors such as smart watches have led to increases in the availability of various kinds of data for moving objects. Thus, new exploratory tools and knowledge discovery techniques are required to extract meaningful information, discover interesting patterns, and explore the dynamic behaviour of moving objects in order to transform raw trajectory data into useful information to be used in reality interpretation and decision making. Especially, analysis of movement observations, which contain information about the movement of each individual entity and the underlying mechanisms, are of great interest. These observations are key to the study and understanding of movement behaviour.
It is essential for movement behaviour studies to take into consideration the so-called movement parameters (MPs), e.g. speed, acceleration, or direction, which are key to characterize the movement of objects. For example, in some sports, it is important to be aware of the heart rate or speed patterns of athletes through running or walking at different times of the day in order to detect the athletes’ performance. In such cases, analyzing a persons’ movement observations in terms of space (x, y, z) and time (t), through considering movement attributes of each person and contextual information (e.g. the environmental information such as temperature) will allow a better understanding of behavioural movement patterns and effect of different parameters on patterns as well.
This article proposes analysing walking observations to extract behavioural pattern of attributes (such as speed and heart rate) of a person to examine the influence of different conditions on behavioural movement patterns. For this, a segmentation approach is proposed to project the “movement parameter profiles” into a pattern for analyzing changes in different movement attributes of each individual during walking in order to identify the effect of each personal context (e.g. gender of the person) and environmental context (e.g. day time, tiredness, weather condition, etc.) on the movement patterns. A measure to assign distance value is then used to examine the effect of contexts on the movement. As expected, the results show that time series of each attribute for each person have a unique pattern describing the persons’ movement behaviour, which changes in different contexts. For detecting the effect of different personal and environmental conditons on the behavioral movement patterns of people during walking, the comparing procedure was implemented on movement parameter patterns with the same conditions and different conditions respectively and a running index was introduced to model the similarity and persons performance during walking. This model was used to predict the similarity between the movement parameter patterns with different conditions. Finally, by comparing the predicted value and the similarity which calculated by proposed approach, we could reach the effect of different situation on behavioral movement pattern of people during walking.