LucioC
commented
11 years ago Unit tests done. More implementation needs to be done.
- Define all the types of modalities events that can occur
- Define the specific combinations for commands
- Implement the commands that will trigger this events to the engine
- Test the integration
Current implementation is not using fusion of modalities. Refactoring is needed.
Sugested approach: Instead of use a direct function call to server, devices should (be able to) send the trigged event identification ex. Swipe right gesture, button lock. A new module would receive these events and then based on its configuration trigger the resulting action.
For REST, a new resource point should be used. Since kinect runs in a local thread it could be implemented throw thread events/notification or using a REST call. In both cases the endpoint/observer should be passed for thread creation.