Hydrosys

Visualization

HYDROSYS' users can access multivariate hydrological and environmental data that will need to be visualized on the handheld devices. In order to support the on-site analysis, visualization methods generally used by hydrologists and environmentals scientists need to be adapted to run in real-time on these small devices. Due to the size of the handheld displays, and the computing power of the devices, the methods need to be extremely optimized. Visualization includes the handling of large map data sets, 3D models of various data sets, flow visualizations, and the inclusion of decision making support including alert mechanisms. The visualization methods will be applied both as overlay on map-like visualizations, and registered to the real-environment using overlay techniques.

In order to compare different kinds of data, focus + context visualization techniques will be developed. These techniques allow the user to relate micro (on-site) and macro (larger range, such as aerial imaging) data in linked views, and to compare different kinds of visualizations from the multivariate sensors. The focus + context techniques will handle large amounts of partially complex data on smaller displays.

Results

At the handheld computers, users are able to browse through sensor data and simulations that can be displayed as point data labels registered in the environment (augmented reality labels) and various optimized simulation overlay, embedded in video content. The cellphones allow browsing of similar data connected to a 3D model of the environment in a list or plot-based fashion. Multiple studies have been performed on modulation of representations, to draw the attention of the user on particular aspects in the environment, so called saliency techniques, of which several implementations exist. Furthermore, in cohesion with user interface studies, a comprehensive investigation of perceptual issues for representation and interpretation of augmented content was performed to optimize display of content on small displays.

Publications

Mendez, E., Feiner, S., Schmalstieg, D. Focus and Context in Mixed Reality by Modulating First Order Salient Features. In Proceedings of the ACM International Symposium on Smart Graphics, 2010. [PDF]

Mendez, E., Schmalstieg, D., Feiner, S. Experiences on Attention Direction through Manipulation of Salient Features. In Proceedings of the IEEE Virtual Reality Workshop on Perceptual Illusions in Virtual Environments (IEEE VR 2010), 2010 [PDF]

Hybrid tracking

HYDROSYS end-users often work in difficult terrains. Tracking users in such terrains requires hybrid localization techniques combining which are complementing each other. This task focuses on the integration of different sensors consisting of GPS, inertial sensors, ultra wide band signalling and computer-vision based techniques to obtain robust results. A global estimation scheme will use both relative measurements between devices and vehicles and absolute measurements through GPS to optimize the individual position estimates. In order to integrate different sensors, their statistics must be known to a high degree of accuracy. Currently, the existing sensors (ultra wide-band, GPS, inertial and magnetic) are being characterized in order to develop a system to fuse the measurements. New sensors including a computer vision system to localize position using the visible horizon are under development.

Results

Within the vehicle setup, a major milestone has been successfully reached in terms of building and setup the system. The setup test has been conducted in Davos on March 2010 with University of Cambridge. The service transmitting the 3D real time localisation data has been implemented and successfully tested. On the other hand, in order to predict the system performance and behavior and also to setup and test the hybrid tracking together with the University of Cambride, Ubisense has been working on the simulation using a real world reference including the system setup. This enables investigating the potential limitations of the system by assessing most of all possible situations and therefore to adapt the hybrid tracking algorithm.

A part of the hybrid tracking platform consists of a ultra-wide band tracking system. This solution consists in equipping a car with a set of sensors connected to a Power Over Ethernet (PoE) switch onboard the truck and track people (equipped with small localization tags) moving in the area around the vehicle. This solution has the advantage of being movable and easy as the sensors are mounted on the vehicle with a fixed configuration.