The visualizer plays a central role within our robotics team development. During a Robocup match, it gives an overview of the current match state and diagnostics per robot within one view. It helps in making decisions whether to pull out a robot for servicing. A second functionality of the visualizer is to show people how the robots perceive their surroundings, whilst giving demonstrations to for instance high schools. Lastly, the visualizer greatly helps our internal development cycle, by allowing data stream pausing, playback & low-level inspection.
The visualizer is written in QT/VTK. We develop in Ubuntu 16.04, C++11 and use git(lab) for version control.
Improve the visualizer on whatever the team requires. Some ideas currently identified:
Draw operator attention to a robot which perceives problems, for instance by red blinking
Expand robot details (such as AI action choice) using mouse-over events
Drag-and-drop interface to simulator and command dispatcher
Integrate low-level computer vision diagnostics (which is currently a separate openCV toolset)
Augmented reality: insert a video feed as background, for comparing with ‘ground truth’
You are dedicated to come at least once a week at the Locht to develop for the robots