Unfortunately, the data acquired was too ambiguous. The word ‘boo’ was initially put into the library as a reference to compare to the input signals. However, when we input another signal of the word ‘boo’, the system output the word “key”. This means that the system thought the word was most similar to the word ‘key’ compared to the other words in the library. Evidently, the system was not correct since it did not correctly identify the word as ‘boo’, for this reason, we decided to add a cable shield in attempt to collect more accurate data. However, while adding the cable shield, the Mayoware power pin was burnt and the board no longer word. Unfortunately, this prohibited our group for getting any more data even less getting accurate data.
Throughout this semester we had a lot of troubleshooting with the board, the Mayoware battery shield fell off at some point in the process, the power pin fell off and was incorrectly fixed, which caused a stall in our project. Unfortunately, we were unable to collect more data. However, this project was a great experience in which we learned a lot from. No one in our group knew how to solder or had used particle photon so we learned how to do so in the project. If someone were to continue our project, we would first recommend them to use better EMG signals recognizer, the data we received was a little jerky at times and could have benefited with using another EMG signal recognizer. Furthermore, the Mayoware had a very small data sheet that we thought was not informative enough. Second, we would recommend to perhaps purchase the updated ThingSpeak or another program that would collect data faster with a smaller latency. Although we did not reach our goal of 70 percent accuracy, we did accomplish what we had planned on our grant schedule.