The end-result of out project was one of consistent performance, which we are very happy about. The Marauder’s Map website, voice-activated commands and door-opening mechanism all worked very well, provided the commands were pronounced clearly and correctly. Starting work on the map website was very important, as was figuring out the correct specs for a servomotor that could successfully generate enough torque to turn the door. There was a preliminary idea to use Raspberry Pi microprocessors in the design. The team attempted to access a Raspberry Pi with MakerHub, an add-on for LabVIEW developed by diligent. Complications were encountered and the team was unable to use the software because of compatibility issues with the outdated LabVIEW add-on, and the current version of Raspbian. Using ThingWorx to store and update the boolean variable for signaling a password-match was initially implemented, but issues with the cloud caused this to be dropped in favor of shared variable usage. The LabVIEW code that accessed the Google speech recognition API could not be used on a myRIO (instead it was run on a desktop computer) because the myRIO could not convert audio input into a wave file. The initial design for the door-opening mechanism had the servomotor positioned very close to the handle. This design saw many revisions and was made successful due to a combination of sound advice from multiple students and the TA, as well as improvisation and resourcefulness.
While the end-result was functional, it would have been nice if the wired microphone could be switched out for a Bluetooth-wireless microphone. This would prevent the need for a wire to trail from a computer inside the room to a mounted microphone outside the room. The team attempted to work with Bluetooth at first, but the transceivers required an amplifier. The team did not foresee this issue when planning out the system. Instead, multiple extension cables were required to achieve the necessary length, and we were limited to the computer that was closest to the door. To conclude, we have included a SWOT analysis below:
- Voice-activated map and door-opening mechanism had very consistent performance.
- Mounting the front panel and motor system is not too difficult with velcro and clamps.
- The button can be pressed once or held down while operating the door-opening mechanism, which helps with outcome consistency for first-time users.
- There is a time delay of several seconds between the microphone receiving the voice command and the door opening. It is not yet known how to reduce this delay.
- The user may not always know how to properly pronounce the password.
- Other password combinations, or different rooms being opened altogether, are within the realm of possibility.
- Since the microphone is wired, this limits the number of computers available to run the LabVIEW software. They need to be as close to the door as possible.
- The clamping technique for mounting the servomotor may not work for other doors, since there is a large required gap for the clamps to fit under.
- The string connecting the servomotor and the door handle may loosen over time, which can prevent sufficient torque from being generated to open the door.