A Study on Moving Objects. Representation and Reasoning

Content

Abstract

1 Introduction

2 Problem description and motivation

3 Reference framework for movement representation and reasoning

4 Problem description and motivation
4.1 Data models
4.2 Location management
4.3 Indexing and data mining
4.4 Logical representation and reasoning

5 Classification of previous techniques

6 Ontology driven approaches
6.1 Spatial foundations and proposals
6.2 Temporal foundations and proposals
6.3 Spatio – temporal research

7 Conclusions and future work

8 References

Abstract

There are some researches that explore and help to solve the representation of moving objects, mainly focused on improving DBMS and Query Languages for support spatio-temporal information storing and processing. Some other researches present techniques to analyze trajectories using data mining, indexing and math operations. More recently, with the emergence of the Semantic Web, the tracking of trajectories of moving objects became a fully applicational field and an open area for ontology applications. This paper explores moving objects representation and reasoning, aiming to find out foundations for a semantic representation of moving objects approach, for solving some existing problems when considering spatio-temporal aspects and semantically enrich query results over any stored information of moving objects.

Keywords Ontologies, Data Storage Representations, Database semantics, Interactive data exploration and discovery, Knowledge acquisition.

1 Introduction

The constant advances of the computing sciences, the GPS and sensors technologies and the growing necessity of more accurate applications for tracking planes, cars or persons have caused re-searchers to pay special attention to the analysis of the motion semantics, especially to the representation of movement of spatial entities and the associated reasoning. All these spatial objects that change their position or shape over time are called moving objects.

Most efforts have been focused on improving DBMS and Query Languages for support spatio-temporal information storing and processing. Some other researches present techniques to predict trajectories using data mining or indexing. Moreover, with the emergence of the Semantic Web concepts and the known relevance and usefulness of ontologies, some research had been coursed on this vein. There are ad-hoc ontologies for representing moving object semantics but tightly related to a very specific area. Thus, we have identified a lack of formal ontology representations for moving objects and foundational statements for its development, or maybe a researching-path.

There is some research done, focused mainly in the knowledge representation about the motion semantics and the subsequent reasoning. Examples of papers that enclose this work are: [1] [2] [3] [4], considered important previous surveys. Other important papers are: [5] [6] [7] [8] [9] [10], which we classify as foundational or vision papers. Many of these papers deal with the temporal and spatial semantic separately, even when describing the present problematic. Moreover, few moving objects-centered surveys can be found on the literature.

One of the main goals of this paper is to survey the work done so far on this vein, aiming to find foundations for a suitable representation of moving objects semantics. In our case, in contrast to previous surveys, we go quite back before ontological proposals and analyze all from the spatio-temporal perspective, firstly, analyzing the spatio-temporality separately and then as a whole.

Thus, our main goals on this paper can be summarized in:

- Analysis of the motion semantic representation and reasoning problem from a whole spatio-temporal perspective based on an established reference framework.
- Grouping and classification of some important existing techniques given their perspective and intend.
- Identification of important lacks over the present ontology driven approaches.
- Statement of a researching path for moving objects semantics representation using ontology approach.

This paper is structured as follows: After this introduction Section 2 depicts the entire problem from a whole spatio-temporal perspective and gives some motivations issues. Section 3 establishes a reference framework for moving objects representation and reasoning. On Section 4 existing techniques for solving the problem are studied and detailed. Section 5 presents a classification of these techniques. Section 6 analyzes the Ontology-driven approaches and, finally, Section 7 presents conclusions and future work.

2 Problem description and motivation

The analysis of motion semantics have become in an important issue for the correct representation and subsequent reasoning about moving objects in the world (trajectory analysis, movement patterns, movement prediction).

We can differentiate these objects depending on the type of movement they will have. Thus, when tracking mobile phones, planes, cars, or persons we can talk about moving points. When we represent hurricanes, lakes (e.g. that changes extension) or the spread of an epidemic we talk about moving regions. Finally, commonly referred as rarer cases, when we represent something like snakes, a moving float of ships or convoys, we need to talk about moving lines. Most of these application fields are extremely important and some of them need accurate information in order to avoid undesirable disasters, for example aircraft tracking.

Some research had been done and promising ideas are stated today. But there are some interesting challenges:

- how should we represent moving objects
- how to keep track of the past movements and predict future behavior

Figure 1 illustrates the common spatial representation of a moving objects trajectory.

illustration not visible in this excerpt

Fig.1. Representation of moving objects trajectory

In this domain (e. g. Schneider in [1]) moving objects are usually classified taking into account “the constraint of the movement of these objects”. Thus, they can be classified in:

1) Moving objects in unconstrained environments.
2) Moving objects in constrained environments.
3) Moving objects in networks.

In the former classification, a re-classification can be done taking into account “what to do with the information stored: analyze past or predict future”. Thus, this classification can be done according to two different points of view:

- History of movement of the moving objects.
- Predictive movement for the moving objects.

These two ways of re-classification usually re-main into the second and third group, thus, we believe that they can be applied to moving objects in constrained environments and networks.

The tracking of these objects have been used in many applications such as float control and car navigation systems, location based services, GIS, video surveillance and others. Each one of these areas uses a different representation of the motion semantics. Anyway, there is a notable lack of approaches for full integration between applications (e.g. GIS applications) that handles moving objects data and also for reasoning over big amounts of spatio-temporal data.

Ontologies seem to be that missing link for this purpose, but actual approaches are not robust enough when we compare them with previous efforts (database-driven, qualitative reasoning, description logics). Thus, some of the created “body of knowledge” on knowledge representation about evolving objects could be imported to the ontological level. Next sections analyze the previous efforts on this topic.

3 Reference framework for movement representation and reasoning

Besides analyzing the existing proposals, we establish a reference framework for representing and reasoning over moving objects semantics. Given this framework, we will have more criteria for the study and comparison of existing approaches. This framework basically encloses two components:

- Moving objects representation: Could be data models for databases, database schemas modeling approaches, description logics representation, RDF or OWL ontologies, or extension of spatial representations with temporal aspects; it could be any effort that allow spatio – temporal data storing and subsequent reasoning.
- Spatio – temporal reasoning: Based on database query languages, data mining techniques, rule – based reasoning, qualitative reasoning, or ontology query languages that include a spatio – temporal scope. This includes:

I. Location and trajectory analysis.
II. Movement patterns recognition.
III. Future movement prediction.

Having these two components we can guarantee the application of reasoning over spatio – temporal data of moving objects and use it for GIS applications, for instance.

The following section analyzes existing techniques basing the analysis on this reference frame-work.

[...]