I achieved my weekend goals. Not only did I change the Kinect service to a polled model, but I have completed and tested the basic logic for my position engine. That being said, I have a long way to go regarding the fine tuning of the thresholds of the algorithm as well as the smoothing parameters of the Kinect — but that can come later. I suspended work on the navigation portion of the project to concentrate on my Kinect services, with the idea that if I couldn’t get them to work it wouldn’t really matter if the other components worked. Now that I have finished and proven the template code for the Kinect I can return to navigation. After that is completed the final challenge will be integrating all of the components as a state machine. I have some ideas regarding that structure, but I’ll talk about those later.
I have run into some issues with VPL and my particular work habits. I really like creating small logic components and then later integrating them into the project, a practice that is difficult with VPL. I can understand why that is the case since many times the message flow between services depends on context, but it is still an issue. I have compiled smaller activities into services with mixed success. The tutorials say that it is usually just a matter of going into the C# code and fixing some references which I may have to investigate that farther in order to simplify my overall VPL logic.