Augmented Reality combines the real world environment and computer graphics to display a 3D image. The large volumes of information stored as data in storage devices from as small as CDs to massive databases .This data can be transmitted in reality through data manipulation by use of XML database querying languages such as Xpath, XQuery, XSLT.
In this research I am going to Explore and critically evaluate the developing relationship between Augmented Reality Systems and XML Database.
Augmented reality system and XML database need each other for existence. This is to say the XML database can be the back-end and AR system can be the front-end to the XML Database as it brings to life what is stored in the XML database by displaying that stored information in the form of 3D images.
By definition "Augmented reality (AR) is a term for a live direct or indirect view of a physical real-world environment whose elements are merged with or augmented by virtual computer-generated imagery; creating a mixed reality" . There are three main things to be noted in augmented reality; AR combines real and virtual environments, is interactive in real time and is registered in 3D.
"An XML database is a data persistence software system that allows data to be stored in XML format. This data can then be queried, exported and serialized into the desired format".  Thus, this research as mentioned above is to evaluate the relationship between the two technologies. AR systems are quite complex due to their nature and will need the versatility, adaptability and flexibility of XML databases. This is not to say XML databases are that easy to manipulate as compared to AR Systems. On the contrary, AR systems can be easier to use, they can even be used by a novice but the XML Database require some sort of training before one can put it in use.
Techniques Affecting Data storage/manipulation
There are a two Database types: The XML-enabled Databases and the Native XML Databases. These can be in the form of Relational databases, Distributed databases, Analytical databases, flat file, object- Oriented databases, Oracle databases etc. For the purpose of this research, I am going to focus on the Relational database as most of its products are now XML enabled and are the most commonly used today.
XML enabled databases provide for the capability to transfer data between XML documents and data structures of the database, this transfer capability typically relies on the specification of the XML schema.
All XML databases now[update] support at least one form of querying syntax, just about all of them support Xpath for performing queries against documents while others support XSLT as a method of querying results retrieved from the database and others support XQuery to perform querying.
"Query includes XPath as a node-selection method, but extends XPath to provide transformational capabilities. This approach helps in combining Relational and XML data with much ease" .
Both data filtering and transcoding have to be applied to the data from a geospatial data source in order to obtain a data set suitable for an AR model. Data transcoding describes the conversion of the data representation used by a GeoDBMS into an XML-based encoding for the description of the AR model.
Data Integration/Management Issues
Load balancing and object migration is one of the centre issues here as the requirements of the AR application might be compromised because the mobile wearable computers might not be able to track and augment fast enough to satisfy the AR requirements. So a decision has to be made here as to how much computation will be performed on the mobile wearable device and how much should be performed on the stationary system. With the right balance being made, bandwidth and lag will be saved. 
There is developing relationship between augmented reality systems and XML Databases in the form of feedback. This is brought about by the interaction between any type of mobile unit in operation and the stationery system. For instance the satellite navigation where a user can use a stylus to impute the required information/direction and the message is immediately processed and directions requested for will be displayed on the screen with a voice actually alerting the driver where to turn and how far he has got to travel.
Further more to the above point of feedback, there is an issue of seamless integration of the virtual world and the real world. Because the real world is not static, the virtual objects have to be seamlessly integrated into the environment and in such a way, the AR paradigm will be exploited as the computers will not only augment the real world but also track and analyse how it changes.
According to Mathias Haringer and Regenbrecht, H.T in their article: A pragmatic approach to Augmented Reality Authorising, describe the Augmented Reality (AR) authoring system "PowerSpace" which allows fast and comfortable generation of AR worlds. "The system presented uses the functionality of a 2D presentation program (Microsoft PowerPoint) as the basis for the composition of 3D content. An MS PowerPoint export is used to generate an XML-based extensible description of a presentation. This description is enriched by 3D content with the help of an editor. The content of this presentation is finally converted into 3D scenes and is used in an AR-Viewer" 
Coverage of Current/Future Research and Development
Advances in computer technology in recent years have facilitated the study
of augmented reality systems and evolved many new applications on a variety
of application domains. Typical applications in which AR has been applied
are: manufacturing, maintenance and repair, medical, military, collaborative interior design, entertainment, cultural heritage and fashion design. 
As with any new technology, there is bound to be challenges but how the challenge is addressed is what matters, with the AR systems, they must be structured in a way that they are flexible to address a wide range of AR applications. Thus a design to make an extension to the Geography MarkupLanguage (GML3) specifically targeted at AR is on the way. GML is an XML-based encoding for exchange of geographical Objects.
The Global positioning system (GPS) still has a limitation: that is, it's only accurate to within 9 metres (30 feet) and doesn't work well indoors.
Another downside to AR systems is that, there is over reliance on AR applications and they end up missing what's right infront of them.
There is also the issue of privacy with the some of the possibilities offered by AR applications with image recognition software, for instance an iphone user can just point his phone at someone with a similar device and be able to get information about them from their face book, twitter, Amazon.
Despite some of the concerns, AR systems have so many possibilities. For instance one can get to know above a city where they haven't even been by just pointing the AR-enabled device (phone) towards that direction, locate places of interest.For those in the construction industry, they can save on materials by using virtual markers to designate where a beam should or which structural support to inspect.
In conclusion, it would be important to note that there is an increasing relationship between XML databases as it has become a popular format for online distribution of datasets that AR systems need in order to display all these fascinating thoughts that are stored in their 2D forms. However the concept of AR system is still not that popular because of its cost of implementation.