Spatio-temporal databases Advanced Database Systems Spatio-temporal databases Nguyễn khắc trung Văn đình vỹ phương Trương viết toán Đào thị thu trang 2015

Outline Part 1: Definition Part 2: Spatio - Temporal Data Models Part 3: Comparing and Conclusion Part 4: Questions & Answers

Part 1: Definition on spatial-temporal data modeling A spatiotemporal database is a database that manages both space and time information. Common examples include: Tracking of moving objects. A database of wireless communication networks

Definition on spatial-temporal data modeling Spatial-temporal data model are the core of Spatial-Temporal Information System (STIS). They define object data types, relationships, operations and rules to maintain database integrity. Research both on spatial database and temporal database is basis for spatial-temporal database models.

Temporal semantics Granularity Temporal operations Time density Representation of time Transaction / Valid time Time order Lifespan

Spatial semantics Structure of space Orientation/Direction Measurement Topology

The spatial-temporal semantics Data types Primitive notions Type of changes Evolution in time and space Space-time topology Object identities Dimensionality

Query capabilities Queries about locations, spatial properties and spatial relationships. Queries about time, temporal properties and temporal relationships. Queries about spatio-temporal behaviors and relationships.

Part 2: Spatio-Temporal Model Event-Oriented, Object-Relationship, Object-Oriented, Moving object, …

The land informations system Example the change in real.

Part 2: Spatio-Temporal Model Event-Oriented Event Oriented Spatio-Temporal Data Model (ESTDM) Previous model not identify individual changes or events to data set raster-based event- oriented * Event Oriented Spatio- Temporal Data Model (ESTDM) -Group time-stamped layers -Store changes in relation (not snapshot as Snapshot model) A header file contains information its thematic domain, pointer to a base map, and pointers to the first and last event lists. The base map is initial snapshot of a single theme of interest in a geographic area. Every event is time- stamped and associated with a list of event components to indicate where changes have occurred

Part 2: Spatio-Temporal Model Event-Oriented Amendment Vector *another event-oriented approach, called the amendment vector approach A base state (or a final state) is overlaid with amendment maps, representing the events in the database The advantage of this representation is: information about what happened to the objects is stored in the database. The time domain considers discrete, relative, linear modeling

Part 2: Spatio-Temporal Model Object-Relationship Model (O-R Model) Previous model: enhanced current GISs realistic world However: None covered the description of change processes existing models concern design (rather than: representation of natural environmental changes, processes and events)

Part 2: Spatio-Temporal Model Object-Relationship Model (O-R Model) Different methodologies present: Modul-R Mecosig and Pollen Object-Oriented Modeling Application Data with Spatio- temporal features (MADS). implemented following object relationship The implementation of object-relationship models describe “processes, which act on the geometric attributes of an entity” and illustrate the importance of capturing the processes, which cause change in connection with space and time.

Part 2: Spatio-Temporal Model Object-Relationship Model (O-R Model) MADS incorporates space & time model basics to object- relationship model At the same time, the object properties, which are required to be captured in the spatio-temporaldatabases, are stated. The processes influencing a single object’s geometry are described as the characteristic of the geometry attribute The spatio-temporal processes are visualized through icons in the schema. - spatio-temporal database processes allow users to handle the complex data models required by higher-level abstracted spatio-temporal applications.  Therefore, the types of processes were classified and represented as relationships between involved spatio-temporal objects.

Part 2: Spatio-Temporal Model Object-Relationship Model (O-R Model)

Object-Oriented Data Model (O-O Model) Event oriented: focus on event, capture spatial part in discrete time Object relation: define each object- and relationship between them. Object oriented: focus on object, adapt with OO programming Query: trees x in region Q at time T3

Object-Oriented Data Model (O-O Model) 4 advantages of O-O: A Single Object can represent the whole history of an entity Queries are simple Efficient temporal data handling Uniform treatment of S-T Data handling 1: one obj can content all information of one obj in real 2: relationship between obj 3: time can be store as part of obj 4: time and S is part of obj

Object-Oriented Data Model (O-O Model) Worboys ST-object ST-OBJECT BiTemporal ……. BiTemporal: valid time and transaction time From Start to End Spatio

Object-Oriented Data Model (O-O Model) Wachowicz and Healy (version management) ROOT ST-O G.To.Th complex versioned objects with geometric, topological and thematic properties A new instance of an object with a different identifier is created for every version ST-O G’.To’.Th’ ST-O G.To.Th’

Object-Oriented Data Model (O-O Model) Bonfatti and Monari (Law) ST-OBJECT LAW……. Strt characterize object structure: structure and relationships. laws describe the behaviour of the components geographical structures and phenomena laws describe the behaviour of the components not possible for the LIS(Land Information System) Applications listed as spatio-temporal dominant (simulation modeling, electronic charting, environmental management, etc.) Rel

Object-Oriented Data Model (O-O Model) Rojas-Vega and Kemp - SIDL Structure and Interface Definition Language (SIDL) geographical structures and phenomena laws describe the behaviour of the components not possible for the LIS(Land Information System) Applications listed as spatio-temporal dominant (simulation modeling, electronic charting, environmental management, etc.)

Object-Oriented Data Model (O-O Model) Object Model Technique - OMT Its high-level representation abstraction Representing data varying in 3D space and time presented for storage of data sets concerning environmental monitoring and simulations

Object-Oriented Data Model (O-O Model) Object Model Technique - OMT STO – Data set LAW……. Data set(t) The values in a data set can be spread either regularly or irregularly within the 4D space by x, y, z and t

Object-Oriented Data Model (O-O Model) Object Model Technique - OMT The values in a data set can be spread either regularly or irregularly within the 4D space by x, y, z and t

Object-Oriented Data Model (O-O Model) Object Model Technique - OMT The values in a data set can be spread either regularly or irregularly within the 4D space by x, y, z and t

Moving Object Data Models when we try an integration of space and time, we are dealing with geometries changing over time In general, geometries cannot only change in discrete steps, but continuously, and then we are talking about moving objects.

Moving Object Data Models Time dimension Space vs Time Position Extention when we try an integration of space and time, we are dealing with geometries changing over time In general, geometries cannot only change in discrete steps, but continuously, and then we are talking about moving objects. growing or shrinking regions

Moving Object Data Models Time dimension Linear Discrete/continuous Absolute time model The past presented is data-type oriented with emphasis on generality, closure and consistency Captures the change and movement

Moving Object Data Models Discrete representation “slices” “simple” function on pair (I, u) The past presented is data-type oriented how “moving” types are represented. where I is a time interval and u is some representation of a simple function defined within that time interval.

Moving Object Data Models Discrete representation “slices” “simple” function on pair (I, u) The past presented is data-type oriented how “moving” types are represented. where I is a time interval and u is some representation of a simple function defined within that time interval.

Moving Object Data Models SpatioTemporal predicates dynamic attributes  motion vector (with threshold)  speed,acceleration dynamic attributes, i.e. attributes that change continuously as a function of time, without being explicitly updated. DBMS to predict the future location of a moving object by providing a motion vector motion vector which consists of its location, speed and direction for a recent period of time

Part 3: Comparing and Conclusion Temporal semantic Spartial semantic Spartio-temporal semantic Query capabilities

Temporal All support multiple granularities (except Moving object) Assume time order is linear Earlier: assume discrete modeling of time. Currently: both levels discrete/continuous Some new models: handling events’ duration and history of ST object (lifespan) Moving object: single granularity

Spatial Base on spatial data models Assume vector structure of space (raster is just adopted by 4 models) Only O-O and Moving Object Orientation and direction of ST object during change/movement across space and time dimensions support all type of spatial data handling Vector representation is much simpler and generic, easily be transformed to Raster

In spatio-temporal Data Model Data types Primitive Notions Types of change Dimensionality Event Oriented Data, event & raster structures Event, amendment maps Only Discrete All O-R Point, line, region Entity, relationship Not movement O-O Simple descriptions, point, line, area Object, attribute, relationship All types Moving Objects Moving point, region Sliced representation 2D Comparison related to model domain Common data type: point, line, region, temporal point, temporal interval Approach data type: parametric rectangles and moving points Most capture different dimensionality representation of ST-O

Query capability All models support for attribute and simple queries Composite query -> describe complex and evolving real work object Moving object can resolve any type of ST Query Queries in all 3 categories: temp, spatial, spatial-temporal DB Time dimension is considered to be continuous Some models support variety ST Q, but can not cover all possible physical process

Conclusions The reviewed models vary on following features: Formalisation Implementation Tool Application Spatial model Temporal model

Q&A on the Paper.

Q&A Câu 1: What are talk about of Event-Oriented Model? A. It is easily identify changes or events in data set. B. It uses transaction log to backwards the changes C. It is easily to change the vector-based system D. It uses amendment vector approach and overlay maps

Q&A Câu 2: What is NOT an advantage of using object-oriented approach in STDB? A. Use multiple objects to represent the whole history of an entity. B. Uniform treatment of spatial and temporal data handling C. Queries are simple, because they deal with each single object of an entity D. Efficient temporal data handling

Q&A Câu 3: Moving Object Data Models talk about: A. From an Object to another Object in the time interval B. Attributes of Object can continuously represent by time functions. C. From an Object to another Object by its time functions. D. Temporal part of Object can continuously be recorded.

Q&A Câu 4: Currently, which data model can resolve any type of spatio-temporal query? A. Event Oriented B. Object-Relationship C. Object-Oriented D. Moving object