This week was largely spent wrapping up the integration phase and preparing for testing. We implemented a three-way serial protocol in which the Due requests sensor data from the Micro, and after receiving said data it requests the ID of the most likely candidate object type that is being touched. Once we established that the hardware for this connection was working we began soldering components to protoboards for the final glove assembly. We also made more progress on the refinement of the integrated velocity data from the accelerometer, but sadly there are still some issues involving some weird data being received in certain edge cases. Given that the rotational data works fine and is intuitive to use, we’ve decided to table the translational data for now. We began outlining formal procedures for the testing phase of both our individual subsystems and the integrated system. We also made significant progress on the Max synth. Finally, we continued to work on refining the object/contact detection subsystem in a variety of ways: by experimenting with a custom-trained model and by adjusting the glove tracker’s region of interest. Though the implementation of these refinement techniques wasn’t fully successful, we still have time in the coming week to address the root issues and we have some ideas for how to do so.
As of now, we’re pretty much on schedule. This week will largely be focused on testing and refinement so that we have a working system for the final presentation. There are some minor bugs with the serial data parsing and we are continuing to refine our detection methods, but ultimately we expect that we will have an adequate product ready to go for the presentation.
This week we intend to dedicate most of our time to testing and refinement. Since we have already begun outlining our testing plans we should be ready to dive right into testing soon. We also plan to finish the final assembly of the glove. It is very crucial that we have the assembly finished as soon as possible in order to start testing.