IocTober Day 31 - “Enhance 57 to 19”

We’re creating an IoT solution for every day of “IocTober”.

November 2019 is the setting of Ridley Scott’s seminal 1982 motion picture Blade Runner. We didn’t get undetectable humanoid robots (or, perhaps, that’s what they want you to think), but it is remarkable how much of Scott’s dark future we did get.

Have you notied we can now speak to our appliances without a second thought?

I have a video microscope used for Printed Circuit Board (PCB) inspection, but under high magnification it can be quite difficult to locate the component you want to inspect.

I designed this 3D printed cartesian table to allow precise positioning of the PCB. Initially positioning was via hand cranks, but that got tiring very soon. Addition of some miniature stepper motors makes it remote controllable via a controller pad.

Now pick up a soldering iron and tweezers, and you’re out of hands.

What would deckard do?

Let’s adopt a microphone array intended for turning a Raspberry Pi into an ersatz amazon echo, and provide hands free navigation.

I’m using a Raspberry Pi zero, and the Keyes microphone hat.

Software is Node-RED, plus Amazon Lex. Lex is handy becuase it allows you to define a limited vocabulary and a conversation tree, which delivers better recognition accuracy than an unlimited vocabulary. Such an approach is ideal for appliances where you do not want your kitchen appliance to misparse “make me some toast” as “make me a ghost”.

There are a pleasing number of voice frameworks out there aimed at producing conversational appliances. I think we are in the early days still of learning what to do with this technology.