Hydrosys

Remotely controlled cameras

During on-site analysis, multiple cameras can be accessed, delivering both high quality images and video streams. All cameras are registered in their environment and thus can be used to overlay "augmented reality" visualisati ons. Cameras are mounted at the handheld computer setups, at sensorscope sensor stations, below the blimp, and at a pan-tilt unit. Some of the cameras can be controlled. Of particular interest of using multiple cameras are the spatial awareness factors involved in switching between and comprehending all the views external to the user. The consortium is performing a series of studies to tackle this issue.

Results

To support multi-perspective viewpoints, low-power wireless camera systems have been developed that can be operated in a self-sustained way in outdoor environments. These systems will be deployed on our large site in La Fouly, Switzerland. Freely controllable viewpoints are available by using a remotely controllable pan-tilt unit via the handheld. Furthermore, multiple studies have been performed to investigate the mental effects (situation awareness and mental workload) on exploration of footage from multi-camera systems for truly understanding the site. These studies form the basis for developing tailored techniques for the handheld systems as part of the user interface work.

Publications

Veas, E., Mulloni, A., Kruijff, E., Regenbrecht, H., Schmalstieg, D. Techniques for View Transition in Multi-Camera Outdoor Environments. In Proceedings of Graphics Interface 2010 (GI2010), 2010.[PDF]

Blimp (unmanned aerial vehicle)

The objective is to firstly develop systems to capture overhead footage of the sites in both optical and thermal infra-red bands, by reconstructing detailed textures by stitching many captured images. Secondly, a system will be developed to allow used to access the overhead footage, by developing a visualization system and integrating this system with other components. The overhead footage will be captured from cameras mounted to a helium-filled, remote controled blimp. Control of the blimp will be performed using a mixture of onboard sensors (GPS, inertial and a vision system using the cameras as input) and external sensors observing the blimp from the ground. The data collected from the blimp will be processed using a mixture of online (live) and offline processing before being passed to the other components.

Results

The blimp system is operational and has been tested succesfully in Davos. A customized telemetric platform has been developed to attach the equipment for recording the footage to the blimp, taking into regard the lift (weight) limitations at high altitudes. A control structure for flying the blimp is avalaible.

Global sensor network

GSN is a Java middleware environment that runs on one or more computers composing the backbone of the acquisition network. A set of wrappers allow live data from sensor network deployments to be imported into the system. The data streams are processed according to XML specification files. The middleware is built upon a concept of sensors (real sensors or virtual sensors - new data sources created by processing or repeating live data in software) that are connected together in order to build the required processing path. For example, one can imagine an anemometer that would send its data into GSN through a wrapper (many are already available and writing new ones is fast), this data stream could then be sent to an averaging virtual sensor, the output of this virtual sensor could then be split and sent to a database for recording or to a visualization layer for displaying the average measured wind in real time. GSN can be fed with live or archived data for demonstration or exploration purposes.

Results

GSN is integrated with an R server and therefore applying arbitrary R scripts on sensor data is now possible. The integration of GSN with Alpine3D for complex environmental hazard risk assessment based on sensor data is supported. A new fully parametrized CSVWrapper component has been developed to load archived sensor data from CSV files. Smartclients for the handheld platform and cellphones support access to sensor data content. A low-cost WiFi bridge system has been developed to support high bandwidth network access on-site.

Publications

N. Bonvin, T. G. Papaioannou, and K. Aberer. An economic approach for scalable and highly-available distributed applications. In Proc. of the 3rd IEEE International Conference on Cloud Computing, 2010. [PDF]

N. Bonvin, T. G. Papaioannou, and K. Aberer. A self-organized, fault-tolerant and scalable replication scheme for cloud storage. In Proceedings of ACM Symposium on Cloud Computing, Indianapolis, IN, USA, 2010. [PDF]

Bonvin, N., Papaioannou, T., Aberer, K. Cost-efficient and Differentiated Data Availability Guarantees in Data Clouds. In International Conference on Data Engineering (ICDE2010), 2010. [INFO]

N. Bonvin, T. G. Papaioannou, and K. Aberer. Dynamic Cost-Efficient Replication in Data Clouds. In First ACM Workshop on Automated Control for Datacenters and Clouds (ACDC09), 2009. [INFO]