How does Athena, the first voice-operated assistant specifically designed for manufacturing, handle working in a noise-filled industrial environment? As you can imagine, that is the number one question we get when introducing Athena to OEMs and end users. After all, Alexa, Siri and Google Assistant are usually called on in quiet spaces such as the home or office. Can you imagine trying to use them next to noisy machinery? The results would be less than optimal. Jay Requarth, our Vice President of Technical Development, explains the science that allows Athena to shine, even when the decibel levels are high.
It’s all in the programming
According to Jay, when it comes to dealing with loud noise environments Athena’s software has two important components: the voice engine and the classifier.
The voice engine, which gives Athena the ability to recognize an operator’s speech, is programmed to concentrate on certain key words and phrases related to machining, and ignore random sounds and words that do not match these parameters. “This minimizes false positives or phantom commands,” Jay said. “For example, she has been trained to know the difference between, ‘set the alarm’ and ‘works like a charm’.”
The classifier is another piece of software that listens through the mic and comes up with ways to filter out louder, non-verbal sounds. Within the classifier are profiles of noises to which the mic is programmed to react in certain ways. “If the decibel level is high enough, the mic will turn off or level out the sound,” Jay said. “So, if you’re in an environment where there are multiple noise sources like an air hose and a loud press brake, those profiles can be taught to adjust to the particular sounds.”
Where Athena is today in regards to noise, voice and feature capability is a result of over two years of development, testing and implementation. Our OEM partners give us feedback so we can continuously add to the profiles in the classifier and filter out additional sources of noise. “We release various refinements to OEMs and end users on a regular basis so that all customers benefit from the suggestions,” Jay said. “In especially challenging and unique environments, we are prepared to work with OEMs and end-users to tailor a voice profile to their specific needs.”
Check one two
Athena, as a package to OEMs, does not include a headset and microphone. “We do offer hardware recommendations,” Jay said. “A quality noise cancelling headset with a particular microphone array (two directional mics integrated into boom) is required in a production environment.”
How loud is loud? To give you a frame of reference, the NFL stadium in Seattle has recorded a noise level of 136 decibels (dB). OSHA’s permissible exposure limit for noise is 90 dB for all workers for an 8-hour day. “Athena works very well at this sound level,” Jay said. “But we are constantly meeting with microphone manufacturers and researching software, looking for ways to improve Athena’s performance. I don’t think we will ever stop chasing perfection.”
If you would like more information on how Athena can simplify machine operation and dramatically reduce training costs, click here to Contact Us.