Welcome to my blog! (and thoughts about Human-Machine Automation)

Hello everyone! Welcome to my professional/portfolio blog. I am a first-year Masters student at Tufts University, in the Mechanical Engineering part of the Human-Robotic Interaction Program. (Lost yet? I probably need footnotes to explain it properly, but try the links for now).

HRI is a relatively new program, focused on (wait for it): Humans, and Robots, Interacting.

(It's kind of nice to have what I'm doing be exactly what it says on the tin, for once.)

When I'm not at school, I am probably with this crew:

I've been a full-time parent for 11 years, more or less, and this is my first time as a full-time student in 16 years, so it's an exciting big change for everyone.

I'm taking 3.5 classes this term:

• CS-133 Human-Robotic Interaction

• ME-134 Robotics

• ENP-162 Human-Machine System Design

• and Mechanical Engineering Seminar

I'm sure you'll hear all about all of them here, and I'll be updating the project pages with projects, but this blog and my entire portfolio website exists thanks to ENP-162, so the blog should have quite a bit of Human-Machine Systems content. Speaking of which, some general thoughts:

Human-Machine Automation / Human Factors in Design

Generally, I feel like the human factors part of engineering system design is terribly neglected, especially in the design phase. This is not particularly a hot take - you can find dozens of people on the internet bemoaning the lack of human factors integration in early design and requirements derivation phases of the engineering cycle and discussing how to fix it.

One of those papers notes in their abstract that "Human Factor analysis is mostly performed a posteriori, when a prototype of the product can be manufactured". This is funnier if you misread the Latin, but just as appropriate: in my professional experience, not only are human factors considered after the design process is complete and in review (there was literally a "human factors guy" at our design reviews when I was working on the DDG-1000, whom we'd never seen before that day) but they have also been somewhat the butt of the engineering world, with security and manufacturability engineers.

HFE are not generally available during the design process, and therefore their role in reviews is to try to make as few changes as possible to have the design meet minimum standards. Actual users are completely unavailable during the design process, or even the review --their great-great-grandbosses might be in the room representing the Very Important Customer, but the actual technicians, sailors, and soldiers who operate the machines, replace the cables, sit in the room that's 90ºF on one side and 55º on the other, etc.

So why is the human part of human-machine automation the "posterior" of the design world? Basically, because people are difficult. They are hard to measure, hard to quantify, hard to sort into categories -- even their basic physical characteristics are difficult to accommodate!

As an example, on the left above is a chart of hand span (the furthest that one's hands can reach, pinkie to thumb, from a site on sizing piano keyboards for smaller hands. On the right, a photo of a fillet weld from a discussion on weld fatigue tolerance (that looks really interesting but is not relevant here)

According to an engineering forum, fillet welds have a minimum size and are allowed to be up to 10% of that size smaller at any point along the length. In comparison, the range of human adult hand sizes (even discounting the entire under-18 population) is 5" - which is 83% of the size of the smallest hand!

I'd rather design for a 10% variance than an 83% variance -- wouldn't you?

Ok, Human Factors is hard so we ignore it in the design phase. Does it matter?

It does! (I bet you knew I was going to say that, given what you know about me so far, but bear with me here.).

The human in human-machine interface or human-machine design is put first because they're the reason the whole design exists. The purpose of automation is to make tasks easier, typically for humans*.

If the whole purpose is to make things easier for the humans*, the design should work well for them. If it's going to work well for them, their needs should be designed in from the beginning. Mic drop, the end.

But you said it was hard! How are we doing to do that?!‽‽‽

Stay tuned! More on that in a future blog post.

*If we are making automation to make tasks easier for oxen, or dolphins, or dogs, or aliens, we should center them and their needs and preferences in the design instead, but I'm going to talk about human-centered design as shorthand here, because "intended-end-user-centric-design" is unwieldy.