Why automation doesn't have to be all-or-nothing

MW: You know it’s a long answer, right? I’ve always been in the robotics and flexible automation space, and the approach we take in that space is you don’t look at (manufacturing) as an end-to-end process; you look at it as a bunch of components of processes that in the whole make up customers’ end-to-end process. The idea of scalable automation is to take a piece of (a process) and automate that with some tool that you might be able to redeploy or reuse.

This was the original idea behind industrial robots. You could program them for a job, and if that job wasn’t available in a couple years, you could redeploy them to another job. (If I) put 10 robots on the line, I get 10 times the work done. But I only add them one a time. So I’ve taken out the capital risk (by doing) it in a scalable, modular, piece-at-a-time fashion.

PS: What do maintenance and reliability teams need to understand as far as implementing scalable automation?

MW: Typically if you’re going to do it in a scalable fashion, you’re doing it in a modular fashion. If you’re running material handling with a conveyor system, you need one conveyor to run everything. If the conveyor motor breaks down, your line is down. If you’re running 10 vision-guided vehicles to move stuff and one breaks down, you still have 90% of your capacity. Another advantage (to scalable) is your familiarity with the equipment is better because you’re working on the same machine in a variety of places.

The same holds true for people using industrial robots on an assembly line. Using the same model of robot (for a different application) even if it’s a little too big lowers their risk of failure because people are more facile with that machine.

PS: Speaking of robotics, what skills do you think M&O teams need to have when it comes to reprogramming?

[sidebar id="4"]

MW: When I started in the robot business, the joke was, yeah, the robots were going to do the work of your labor on the floor, but you needed to hire a couple of PhDs to program them. Now most of the robots on the market can be taught using walk-through or lead-through techniques. The robots and the robotic equipment are getting smarter, so the skills required for reprogramming them are less. The expertise of the workflow is what’s more important. With our vision-guided vehicles, people who train the vehicles to run around the facility are the guys who drive manual trucks now, because they know the routes better than the engineers in the plant did.

We have one site where 50 trucks are running. And there’s one person per shift responsible for taking care of them.

PS: What about safety concerns? Do you see a point when vehicles could be reprogramming themselves on the fly?

MW: I sit on the standard committee for guided vehicles. In the current standard, if the vehicle goes off the guided path, it has to declare an emergency stop. So now the question is, if they’re rerouting themselves automatically, when are they really lost? The committee’s just starting to address that.

Seegrid set up a thing called “Rules of the Road” to establish who’s got priority. If a human driver wants to turn into an intersection but a guided vehicle is already there, they have to give way to the guided vehicle. This improves safety for facilities that could otherwise look like the Wild West.

PS: What’s driving interest today in scalable automation?

MW: I think what’s happened in the marketplace is the rate of change is increasing. In terms of product change, you know, people like (new) phones to come out every year. That makes it harder for people who build “monuments” or fixed automation systems. With modularity, you have shorter implementation times.