Latest / Automated with Brian Heater / Universal Robots’ Roberta Nelson Shea on Balancing Safety and Productivity in Automation

Transcript

Universal Robots’ Roberta Nelson Shea on Balancing Safety and Productivity in Automation · Automated with Brian Heater

0:00
SPEAKER_01: It's very easy for companies to say, well, I don't want to spend any more money.
0:05
SPEAKER_01: I already spent millions and millions and millions of dollars on all this equipment.
0:09
SPEAKER_01: Why should I do something different now from what I did 10 years ago, 20 years ago, 30 years ago?
0:17
SPEAKER_01: I remember when cars didn't have seatbelts.
0:20
SPEAKER_01: Yet, who would ever think of not getting in a car and putting a seatbelt on now?
0:24
SPEAKER_01: So the world changes and the expectations change.
0:28
SPEAKER_01: We have to do better and we have to continuously strive to do better for productivity and for safety.
0:43
SPEAKER_00: Hey everybody, welcome to Automated.
0:45
SPEAKER_00: My name is Brian Heater.
0:47
SPEAKER_00: I am the managing editor at A3.
0:50
SPEAKER_00: First of all, thanks everybody so much for your support of the show.
0:55
SPEAKER_00: As a lot of you know, this is something that we've been working on for like six plus months at this point, and it's extremely gratifying to have it out in the world and to have everyone tell us how uh excited they are for the shows that we've released so far.
1:11
SPEAKER_00: I believe that you're going to be equally excited for today's episode.
1:15
SPEAKER_00: We are speaking with an, I don't want to say unsung, but I will say possibly undersung hero in the world of automation and robotics, Roberta Nelson-Shea of Universal Robotics.
1:29
SPEAKER_00: She has been a huge proponent of safety and safety standards in the world of industrial robotics for decades and decades now.
1:41
SPEAKER_00: Uh, she is also going to be appearing at IRSC, the International Robotics Safety Conference.
1:48
SPEAKER_00: That's November 3rd through 5th in Houston.
1:51
SPEAKER_00: I will also be there as well.
1:54
SPEAKER_00: So you could come hang out and meet us there.
1:58
SPEAKER_00: Um this is a great talk.
2:00
SPEAKER_00: I learned a lot.
2:01
SPEAKER_00: Hopefully, you'll learn a lot as well.
2:03
SPEAKER_00: And if you do, even if you don't, uh, please do all the social things.
2:08
SPEAKER_00: Uh, smash the like button and subscribe and uh do whatever your particular social media requires of you to do.
2:17
SPEAKER_00: Uh, thanks so much for listening, and I will see you on the other side of the interview.
2:22
SPEAKER_00: Berta, thank you for joining us.
2:23
SPEAKER_00: I know you said you're not at your place, but uh you're based in Denmark most of the time, is that right?
2:29
SPEAKER_01: My job is based in Denmark, but I live in the US.
2:33
SPEAKER_01: And so I could be working out of either of two home offices, one in Massachusetts or one in North Carolina.
2:40
SPEAKER_00: Knowing what I know about you, it looks like you're probably maybe on the road more than your home.
2:44
SPEAKER_00: Is that a fair assumption?
2:47
SPEAKER_01: Uh that's a pretty fair assessment.
2:49
SPEAKER_01: Yeah.
2:50
SPEAKER_01: I was last home in May.
2:52
SPEAKER_00: Of 2025, I hope.
2:53
SPEAKER_01: Yeah, 2025.
2:55
SPEAKER_00: This is always an interesting conversation that I that I've had with people over the years.
2:59
SPEAKER_00: People ask me what I do for a living, and I've had to um you know, I've had to come up with sort of the pithy edited version of what it is.
3:06
SPEAKER_00: So when you have that conversation with a stranger, they ask you what your job is, what's your stock response.
3:13
SPEAKER_01: Well, I usually give my title and then hope I don't have to elaborate.
3:18
SPEAKER_01: Yeah, my title is Global Technical Compliance Officer.
3:22
SPEAKER_01: And uh what to be careful about is I'm responsible for the compliance of our products and for managing business risks associated with our products what we put on the market, because compliance is a really broad topic.
3:39
SPEAKER_01: Um, we have it legal compliance, accounting compliance, HR compliance, and so forth.
3:46
SPEAKER_01: Somebody says, Oh, I got a question for compliance, and then it turns out it's really not for me.
3:50
SPEAKER_00: Yeah, it's the other compliance department at the company.
3:53
SPEAKER_01: So the technical hopefully helps with that.
3:56
SPEAKER_00: Yeah, I I tend to think of you, um, you know, you you and I spoke uh a couple of automates ago.
4:03
SPEAKER_00: You actually you first really came on the radar at Automate 2023 when you won the Engelberg, but I I tend to think of you as kind of an ambassador.
4:11
SPEAKER_00: Is that fair?
4:12
SPEAKER_01: I guess it could be one interpretation.
4:15
SPEAKER_01: It might be ambassador.
4:16
SPEAKER_01: Uh another one is uh, or what how I like to think of myself is uh I really try to evangelize or or preach the notion that automation can be a tremendous boon to all businesses for for their own profitability, but it can be safe also.
4:36
SPEAKER_01: So we can have the two happen.
4:38
SPEAKER_01: One is not contrary to another.
4:41
SPEAKER_01: And I think that we've we've really uh shown that to be the case in robotics, where we have um very high adherence and interest in safety.
4:51
SPEAKER_01: Um, but yet here we are in the forefront of technology and changing technology.
4:57
SPEAKER_00: Was that the sense in the early days when you really started doing this in earnest that those two ideas of productivity and safety were necessarily in conflict with one another?
5:07
SPEAKER_01: I have never thought that, but I've waged my own battle, you know, Joan of Arkish.
5:12
SPEAKER_01: But um, as I said one other point in time, we know how her life ended.
5:16
SPEAKER_01: So hopefully I'll be more successful.
5:19
SPEAKER_00: And uh the Yeah, it didn't end particularly well for her.
5:23
SPEAKER_00: No, it didn't end well for her.
5:24
SPEAKER_01: So I hope my ending will be much better.
5:27
SPEAKER_01: Hopefully, I'll survive it all.
5:29
SPEAKER_01: Uh I just have always believed it.
5:31
SPEAKER_01: And unfortunately, there's some people that do not.
5:34
SPEAKER_01: And I think that we've proven that it is possible in robotics because we have a high degree of interest in safety.
5:42
SPEAKER_01: And yet at the same time, our whole business is about being more productive.
5:48
SPEAKER_00: I was reading an interview you did not that long ago, and it struck me that you said robots are like essentially already safer than most industrial machining right off the bat.
5:59
SPEAKER_00: Is that right?
6:00
SPEAKER_01: Yeah, I firmly believe that.
6:01
SPEAKER_01: But part of the reason it is that way is that there's been a tremendous amount of interest.
6:07
SPEAKER_01: And in fact, at times you could almost say, why so much interest here and not elsewhere?
6:12
SPEAKER_01: And maybe that's the sensationalization of the Terminator and all sorts of concerns of robots going wild.
6:19
SPEAKER_01: But it is an industry that because it's automated, people can be surprised.
6:24
SPEAKER_01: They might not expect that there'd be motion or something happening, and then there it is.
6:29
SPEAKER_01: And it's with these surprises that more likely people could become injured.
6:34
SPEAKER_01: We've had such interest in safety that it has helped drive standardization process.
6:41
SPEAKER_01: It's very easy for companies to say, well, I don't want to spend any more money.
6:46
SPEAKER_01: I already spent millions and millions and millions of dollars on all this equipment.
6:50
SPEAKER_01: Why should I do something different now from what I did 10 years ago, 20 years ago, 30 years ago, 40 years ago, you know, 50 years ago.
7:00
SPEAKER_01: I remember when cars didn't have seatbelts.
7:03
SPEAKER_01: Yet, who would ever think of not getting in a car and putting a seatbelt on now?
7:08
SPEAKER_01: So the world changes and the expectations change.
7:11
SPEAKER_01: We have to do better, and we have to continuously strive to do better for productivity and for safety.
7:19
SPEAKER_00: Yeah, you you you bring up automotive, and I think a lot about this anecdote that you hear where uh historically and probably even to this day for a lot of automotive manufacturers, there's kind of a cost-benefit analysis of how much it costs to actually implement safety versus how much are we going to pay in lawsuits and medical bills, have to pay to the family.
7:40
SPEAKER_01: Unfortunately, automotive is noted for that, and some of that's because of some uh court cases.
7:46
SPEAKER_01: But the automobile industry is heavily regulated.
7:50
SPEAKER_01: And I can tell you that at UR, we don't do a cost-benefit ratio or analysis that safety costs this much.
7:59
SPEAKER_01: From our perspective, safety is a requirement.
8:03
SPEAKER_01: The products we put on the market have got to be the best that we can make them to perform and also to have safety built into them.
8:13
SPEAKER_01: We can't make anybody use it right.
8:15
SPEAKER_01: How many times do you see people drive down the street and not wear a seatbelt?
8:19
SPEAKER_01: How many times do you see toddlers crawling around in a front seat and not being in a car seat?
8:25
SPEAKER_01: So we can't make people do what they should be doing, but we can put products on the market that enable this.
8:33
SPEAKER_01: And that's how we go about it.
8:36
SPEAKER_00: So so break that down for me a little bit.
8:39
SPEAKER_00: Obviously, you've been really involved over the years on the standard side of things, in terms of actually, you know, making people, in terms of enforcing these rules, in terms of making sure that people are using systems the right way.
8:52
SPEAKER_00: What's the line between a safety standard and regulation?
8:56
SPEAKER_00: And like, is there a point in which safety can truly be enforced?
9:02
SPEAKER_01: Oh, that's a super good question.
9:04
SPEAKER_01: Uh the only way safety is enforced is by governments.
9:08
SPEAKER_01: And so if you think people worry about an OSHA inspector, and I will tell you that there's a great many companies and businesses that don't worry about an OSHA inspector, other than the fact that somebody is coming in and it's not a business as usual because whatever was planned now is a little bit different.
9:27
SPEAKER_01: But if you truly do try to have a safe workplace and you use equipment and machinery that is as updated as you can, let's be realistic.
9:38
SPEAKER_01: You could have a very old machine and it would not meet today's standards.
9:43
SPEAKER_01: But you have ways by which you can address the risks associated with it.
9:48
SPEAKER_01: In the robotics industry, we're, I'm gonna say almost spoiled.
9:52
SPEAKER_01: I wish we could get more spoiled, but we're a little bit spoiled for the fact that there is a real claim to be made by buying newer.
10:02
SPEAKER_01: Because what do you get?
10:03
SPEAKER_01: Higher speeds, you get uh greater resolution, you get more than what you had before.
10:12
SPEAKER_01: So as people automate, they think of more and more challenging things to automate and more and more things to do with their robotics.
10:22
SPEAKER_01: And with doing more, then it means that you end up buying new robots, whether to replace or they're just you're adding to the population that you have in your facility.
10:34
SPEAKER_01: So for a lot of, you know, typical industrial machinery, it's bought and it is used until the company goes out of business.
10:43
SPEAKER_01: Hopefully that doesn't happen, but um, it's just used forever.
10:47
SPEAKER_01: Yes, there are some robots that are used as long as possible, but the reality is they tend to swap out and go to new because there's real benefits.
10:58
SPEAKER_01: So we're in a we're in a great industry from the standpoint that the expectation is that a robot will only be used for X number of years.
11:08
SPEAKER_00: And robots get software updates and firmware updates, which older machines don't get as well.
11:13
SPEAKER_00: So at least theoretically, you can continue to uh state the safety.
11:18
SPEAKER_01: But let's be realistic.
11:20
SPEAKER_01: You still have got a lot of mechanical parts.
11:23
SPEAKER_01: And with the mechanical parts, you can get to the point of does it really make sense to prolong this, or is this time to retire the machine?
11:33
SPEAKER_01: Um, with robots, yes, there is metal.
11:36
SPEAKER_01: So yeah, you've got metal in it, and you say, Oh, that should last forever.
11:41
SPEAKER_01: Um, but really the magic is in the controllers and the software.
11:46
SPEAKER_01: So at some point in time, you can't update the software.
11:49
SPEAKER_01: Well, that that would be like saying, I would like to update my Texas Instrument 50, and there is no update to it.
11:59
SPEAKER_01: Uh, that's the same thing with phones.
12:01
SPEAKER_01: When when is it that we get rid of our cell phones?
12:04
SPEAKER_01: Oh, because we've run out of memory.
12:06
SPEAKER_01: Or it turns out the battery suddenly doesn't work well anymore, or you want to use some new feature that only works on a newer version software, which can only work on faster processors.
12:18
SPEAKER_01: So in the technology world, you end up updating hardware and software because that's the only way you get all these great new advantages.
12:30
SPEAKER_00: I I think something that people tend to uh take for granted, myself included, is this idea of safety being being built in through, you know, through controllers, uh, through software.
12:41
SPEAKER_00: But again, I I get the the sense, having read some of the things that you've written, that that wasn't necessarily always the case.
12:48
SPEAKER_01: Uh no, it wasn't.
12:50
SPEAKER_01: Well, at one point in time, we really didn't have the affordable technology to do what we do today.
12:56
SPEAKER_01: Uh, one example of that is that uh, and it's not so long ago, but on the order of 10 years ago, some chips and sensors used for vision would be$20,000 unit one to buy one.
13:14
SPEAKER_01: And nowadays, because of the advancements and also the demands, even within cars and driving, thinking about self-driving cars, the sensors and vision and processors for all of that.
13:27
SPEAKER_01: Um, those same sense those same sensors and chips that might have cost you$20,000 are now in the order of less than$20.
13:39
SPEAKER_01: Can you say that about anything else?
13:43
SPEAKER_00: Uh yeah, I the only the only comparable thing I would say that about, and this is again, you know, my backgrounds, but I think it's had a very similar uh uh result are uh components that go into the smartphones, right?
13:55
SPEAKER_00: Once you really start scaling them, they get really small, they're really cheap, and then they just sort of start to transform industries around them.
14:01
SPEAKER_01: That's happened in automation.
14:03
SPEAKER_01: We're we're getting all of the benefits that have occurred due to this very massive scale of consumption of faster processors and faster sensors and better vision and better cameras for everything that everybody's enjoying in the cell phone today that actually does impact us in the industrial sector, where we'll never get to the point in which it's millions of units sold every year, millions, and then competition comes in and then it's billions of units sold every year.
14:34
SPEAKER_01: So uh we get the advantages that come along with that consumer goods business.
14:40
SPEAKER_00: In addition to some of the components coming through, in addition to the advances that have been made in autonomy around self-driving cars, do you do you feel like there's anything else the industrial robotic industry can take from autonomous vehicles, from that kind of, I guess, that that that journey, the difficulty that uh they've had over the years?
15:05
SPEAKER_01: There's a lot that not only can be taken, but is being taken on a daily basis.
15:10
SPEAKER_01: Uh, think about what's happening with mobile robots.
15:13
SPEAKER_01: They have the issue of trying to recognize obstacles, all sorts of different sizes and shapes.
15:20
SPEAKER_01: And I remember reading once that for self-driving vehicles, they can recognize a bus, but if a bus is on its side blocking the road, they don't know what it is.
15:33
SPEAKER_01: I don't know if that's still true because for every problem that arises in short order, there's solutions.
15:39
SPEAKER_01: Um, but mobile robots have benefited from everything that is benefiting autonomous vehicles.
15:46
SPEAKER_01: And it will continue to be that way.
15:49
SPEAKER_01: Uh, the notion of the industrial robot, which is the arm waving in the air, uh, we're getting to the point in which we're using AI to try to do better about various tasks.
16:01
SPEAKER_01: And that's uh kinematics, it's object recognition, it's trying to do smarter faster, smarter, faster.
16:08
SPEAKER_00: Obviously, there's always a push of technology, and one of the pushes um has been this idea of moving away from the cage in an industrial setting.
16:18
SPEAKER_00: But in a lot of ways, the cage is a pretty it's an elegant solution and and not 100% of the time, but it mostly works.
16:26
SPEAKER_01: I actually am really glad you mentioned this because I I don't understand, and I have given some talks on this where I've tried to make fun of it and I've called it foof, fear offenses.
16:38
SPEAKER_01: There is this whole thought that new modern machinery is not going to have any guarding.
16:45
SPEAKER_01: We're not going to keep people separate from it.
16:48
SPEAKER_01: But the reality is that uh the easiest thing to do is to keep people and machinery apart.
16:54
SPEAKER_01: So that's simple and easy to do.
16:57
SPEAKER_01: What's problematic, and people are trying to do more and more, I've got to say, complex safeguarding, what's problematic is the fact that some of this is really difficult.
17:10
SPEAKER_01: We could say, oh, let's get rid of the guards.
17:12
SPEAKER_01: Oh, okay, so let's use, and I'm going to be careful to say this because I don't believe there's such a thing as a collaborative robot.
17:18
SPEAKER_01: There can be collaborative applications and collaborative work cells.
17:22
SPEAKER_01: And if we take a robot that's designed to be what we call power and force limited, and yes, that is the world my company is involved in, um, that world with absolutely positively no safeguarding, meaning you don't have these cages, fences, which everybody talks about.
17:41
SPEAKER_01: That's foof.
17:42
SPEAKER_01: Um, but anyway, uh, if we don't have them, then, and then we have to make sure that the application is such that if there's a contact from any part of that robot application to a person, absolutely worst case, we have a slight injury.
17:59
SPEAKER_01: Slight, not serious, uh, not something that will damage a person permanently.
18:07
SPEAKER_01: And so we say, okay, that makes complete sense.
18:11
SPEAKER_01: But let's be realistic.
18:13
SPEAKER_01: As the payload goes up, the robot, it's mass, it gets larger because it's got to move more heavier things.
18:20
SPEAKER_01: So, what do people want to do?
18:23
SPEAKER_01: They want to be able to have a robot operate infinitely fast, be infinitely uh responsive, and to stop immediately and to be able to do this predictably if anybody gets near it.
18:42
SPEAKER_01: Well, we've gotten better and better with safety sensors.
18:46
SPEAKER_01: So nowadays you can have that quote unquote cage-free, which I only think that applies to chickens, but you can have that cage-free robotic work cell, but you're still using safeguarding.
18:59
SPEAKER_01: You're using safety functions built into the robot, you're using safety sensors that enable the robot to move absolutely as fast as it you want your production or cycle count to be.
19:11
SPEAKER_01: And when a person nears it, it slows down.
19:14
SPEAKER_01: The dream in the long term is to have, if you want to say, that psychic safeguard.
19:20
SPEAKER_01: It knows where people are, it knows where the robot is, it knows the intention, meaning the path of a person, and the robot will magically dance around and keep away from the person.
19:33
SPEAKER_01: Are we at that point that we have got the dancing of the robot to keep away from a person?
19:40
SPEAKER_01: And that robot slowing down when a person is near and then speeding back up.
19:45
SPEAKER_01: Uh, are we there?
19:46
SPEAKER_01: We're there from a more crude standpoint.
19:49
SPEAKER_01: So we can do this, but it is not that elegant dance uh that we envision long term.
19:58
SPEAKER_01: So long term we could get there.
20:00
SPEAKER_01: Short term, we're improving ourselves much more so than we once did.
20:05
SPEAKER_01: At one point in time, it was the whole notion of guarding and cages everywhere.
20:10
SPEAKER_01: And the only safe state for a person to be near a robot was with it off and no power.
20:16
SPEAKER_01: And the reality is we can still have a safe state or a safe situation for people and workers.
20:25
SPEAKER_01: Um, but stopping is not necessarily the exact answer.
20:29
SPEAKER_01: It could be you slow down to a sufficiently low speed that any impact won't hurt a person.
20:36
SPEAKER_01: And no one is surprised by motion because you're things are going slowly.
20:41
SPEAKER_01: Um, that's a wonderful world.
20:43
SPEAKER_01: We're not quite there yet.
20:44
SPEAKER_01: We don't have the safe dancing robot that is always right there, but not touching.
20:50
SPEAKER_01: You know, almost like what kids talk about, you know, it's not touching me.
20:55
SPEAKER_01: It's so close.
20:56
SPEAKER_01: That's the dream.
20:57
SPEAKER_01: That's the dream.
20:59
SPEAKER_00: This opens up a whole nother interesting topic of conversation, which is that, and another lesson to be learned, I think, from automotive is that stopping isn't always necessarily the best solution.
21:09
SPEAKER_00: Sometimes stopping can actually cause more problems.
21:12
SPEAKER_01: It can.
21:13
SPEAKER_01: Uh, an example to that would be if you think about uh a nuclear reactor.
21:19
SPEAKER_01: You don't just hit the off switch because you'll blow up, right?
21:24
SPEAKER_01: Uh we're not at that sort of a critical state with robotics in that stopping the robot creates a lot of hazards.
21:34
SPEAKER_01: Stopping a robot too fast, meaning very abruptly, so that the motion is not smooth, that could cause some problems.
21:42
SPEAKER_01: It could very well be, though, we can keep people safe.
21:46
SPEAKER_01: It's just that we don't completely stop.
21:49
SPEAKER_01: We could keep away from a certain area.
21:53
SPEAKER_01: Uh, we could also have it, as I said, go much slower.
21:57
SPEAKER_01: Or if you want to think about it, sometimes.
22:00
SPEAKER_01: You think about people when they are going to bump each other in the hallway, they just kind of move a little bit to the side so that they don't bump shoulders.
22:10
SPEAKER_01: And I would get that more into the scene of how do you make it so that we can have safety such that the robot just keeps ducking and avoiding, but it still keeps doing its work.
22:22
SPEAKER_01: Now, while that's a dream, we also have to be realistic.
22:25
SPEAKER_01: And that is the best way machinery runs is to keep it running.
22:31
SPEAKER_01: So the more interruptions, so if you stop it, then you have to restart it and you have to make sure everything starts up well.
22:38
SPEAKER_01: But if you keep slowing down, that's not a good thing either.
22:42
SPEAKER_01: It's just that you do want that as a response if somebody starts approaching the machine.
22:48
SPEAKER_01: Like the last thing you want to do is have a robot cell right next to a traffic aisle and have no guard on the traffic aisle side or no obstruction on the traffic aisle side.
23:00
SPEAKER_01: Because what is always the case, the restroom is just there.
23:05
SPEAKER_01: The brake area is just there.
23:08
SPEAKER_01: And so you're going to take a shortcut through the cell, and then you're going to stop everything.
23:14
SPEAKER_01: And you could think, well, that didn't bother the person, the person was safe, but we've just interrupted or slowed down production and made for a longer cycle time, which is profitability.
23:27
SPEAKER_01: So I think that the notion of trying to have the absolutely sublime, sublime in dreamlike state of a robot dancing away from people needs to be balanced against the fact of let's just make sure that that robot is working as much as it can for production and that any interventions, or if you want to say people coming close, that we try to minimize those because that actually causes a loss of productivity.
24:00
SPEAKER_01: So we have a balance that always has to be worked on.
24:04
SPEAKER_01: I have yet to ever hear anybody say, I have plenty of floor space.
24:09
SPEAKER_01: I have plenty of floor space.
24:11
SPEAKER_01: No problem.
24:12
SPEAKER_01: I'll install that robot a mile away within the building.
24:17
SPEAKER_01: Uh, and nobody will ever go near it.
24:19
SPEAKER_01: No one ever says, I have so much space.
24:23
SPEAKER_01: There is always the lament that the floor space, and in fact, that's a quote from someone that I I've known for many years at General Motors, and his response was factory floor space is more valuable to us than any amount of expensive equipment we use to fill that space.
24:41
SPEAKER_01: So floor space is a premium.
24:43
SPEAKER_01: Making really good use of it means a lot.
24:46
SPEAKER_00: I suppose one example I could potentially point to, and again, it's a relatively new one, of a robot stopping or powering down that could potentially harm someone are humanoids, right?
24:59
SPEAKER_00: Like these entering the marketplace are really opening up the safety conversation in new ways.
25:04
SPEAKER_01: It certainly isn't.
25:05
SPEAKER_01: In fact, I just went to the inaugural safety standard meeting for that particular sort of technology, which is either going to be called dynamically stable, or which means statically unstable, or actively controlled for stability, because the humanoid, and it's not just humanoid, because humanoid, if you ask somebody what's a humanoid, or they say, Oh, it's got legs.
25:31
SPEAKER_01: They say, Okay, so no arms.
25:33
SPEAKER_01: Well, it could have arms.
25:34
SPEAKER_01: Okay, so no head, nothing that is evocative of a human body.
25:40
SPEAKER_01: Oh no, I expect it to look like a human.
25:42
SPEAKER_01: So could it be on a wheeled base?
25:45
SPEAKER_01: So take a typical, you know, mobile robot, which to me is basically a rectangle, a scooter with wheels, but it's stable if it's not moving, right?
25:56
SPEAKER_01: If you take that and you put what looks humanoid on top of it, is it still a humanoid?
26:04
SPEAKER_01: And the safety issues with dynamic stability are quite frankly that if you stop abruptly, the robot falls.
26:14
SPEAKER_00: Yeah.
26:15
SPEAKER_01: It falls and they're not very good about recovery.
26:20
SPEAKER_01: They're not like little kids who just kind of back up and so light that they don't hurt anyone, but these are heavy machines, and if they fall over and fall onto a person, a person could be seriously injured.
26:34
SPEAKER_01: So that is a real challenge in that workspace, meaning uh humanoid type or anything that is that requires power to maintain stability.
26:48
SPEAKER_01: Um, this will still be able to be accomplished because, but what it means is that robot that is unstable at a loss of power, that robot's got to have a way by which and sufficient power.
27:03
SPEAKER_01: So suddenly battery management is super important.
27:07
SPEAKER_01: You want to make sure that no matter what, all the time the robot has enough juice so that if it detects that a person will come within its fall area, because it's not just exactly where it is, it's where it would topple over.
27:24
SPEAKER_01: Um, that that robot can go to a hunker-down stable position.
27:30
SPEAKER_01: Um, it could be maybe there's outriggers that come out, think of it like a kickstand, right?
27:36
SPEAKER_01: Um, you have to make sure that doesn't hurt somebody, though.
27:39
SPEAKER_01: Uh, but it might have to basically go all the way down to the ground and spread itself wider with arms and legs so that it is then has a stable position, and then it can retain power so that it could come back up from the loss of stability.
27:57
SPEAKER_01: Or if this really is something go south and the battery's gonna die, um, that it gets to that safe pose or position before the full loss of juice, because otherwise that type of robot could fall on someone, not just fall on someone, but could fall on very valuable product.
28:18
SPEAKER_01: Can you imagine a humanoid handling semiconductor wafers?
28:23
SPEAKER_00: Yeah, or or at the Fabriger Egg Factory.
28:26
SPEAKER_01: Yeah, or oh, yeah, that's a really good one.
28:29
SPEAKER_01: I like that Fabriger egg factory.
28:32
SPEAKER_01: Um, and it it within that area, and damage to product could be considered catastrophic to business.
28:40
SPEAKER_01: And uh so there, I think there long term there will be a place for the foreign factor.
28:47
SPEAKER_01: I don't know if truly the foreign factor is two legs.
28:53
SPEAKER_01: I mean, it could be three legs.
28:55
SPEAKER_01: Why the obsession of making it human-like?
28:58
SPEAKER_01: Um, because quite frankly, remember what I said about productivity and cycle time?
29:03
SPEAKER_01: No one ever says also, oh, speed doesn't matter at all.
29:08
SPEAKER_01: It doesn't even have to do its work.
29:10
SPEAKER_01: We'll be happy.
29:12
SPEAKER_01: We love looking at it, you know.
29:14
SPEAKER_01: That that's that's not realistic.
29:16
SPEAKER_01: So we we have the issue that we want automation to operate as fast as it can, uh, yet people be safe around it.
29:26
SPEAKER_01: And right now, in that dynamically stable, meaning statically unstable world of automation, you have the continuous challenge of not just safety, but the fact of if it can become unstable easily, why are you purposely building a form factor that makes it difficult?
29:48
SPEAKER_01: You know, you think a three-legged stool is always balanced.
29:52
SPEAKER_01: I mean, you could fall off it, but it's always balanced.
29:55
SPEAKER_01: And why does it have to be two legs?
29:58
SPEAKER_01: Why did it does it even need legs?
30:00
SPEAKER_01: Because legs, you know, think about going upstairs.
30:04
SPEAKER_01: How often do you trip if it turns out the stairs are a little bit different?
30:09
SPEAKER_00: More often than I would care to admit on a podcast.
30:12
SPEAKER_01: Well, I'll admit I fall.
30:15
SPEAKER_01: So uh that's how I broke a foot once upon a time is I just stumbled at a curb.
30:20
SPEAKER_01: And it's so I think there's a lot of interest in humanoid because let's be realistic, we love the idea of a robot that seems human-like, or many people do.
30:30
SPEAKER_01: I actually find it a little bit disturbing.
30:33
SPEAKER_01: Um but I I did go to one research facility which was fascinating, and they had a robot that would interact, talk with you, and it had a very humanoid looking face with eyes and would blink, and there was expressions that would be part of the experience of talking.
30:56
SPEAKER_01: And I found it a bit creepy.
31:00
SPEAKER_01: You know, there's no other way of doing it because no matter what, you look at it and you see this reminds me of a person, but it's not a real person, you know, and think about people you meet that you say, oh, that person is fake or phony.
31:15
SPEAKER_01: It's not a real person, there's no no real substance inside.
31:19
SPEAKER_01: And and when the fat turns you off, if it's a human, think about if it's not even a human, and you know that the expressions seem different.
31:31
SPEAKER_01: Or there is a pause while looking up and then a response.
31:39
SPEAKER_00: That's my understanding of how the Uncanny Valley operates, is sort of the the closer that you get to realistic, the the creepier it becomes.
31:48
unknown: Yeah.
31:48
SPEAKER_01: Oh, maybe for somebody involved in the design, they find it absolutely fascinating because it's their baby.
31:56
SPEAKER_01: Uh, but I'm kind I'm in the camp, it's a little bit creepy.
32:00
SPEAKER_01: I would like the automation for it to remind me that it is automation and not to try to confound my senses.
32:08
SPEAKER_01: The form factor is interesting in that you can take any existing facility and say, I can use this quote-unquote humanoid type robot.
32:18
SPEAKER_01: Uh, there are tremendous challenges with stability.
32:21
SPEAKER_01: There's tremendous challenges going up and down stairs, uh, and there being some difference in the stair or the floor surface right at that transition time in stairs.
32:34
SPEAKER_01: And the last thing you want to do is have a robot fall down a flight of stairs and land on the backs of people or the fronts of people.
32:44
SPEAKER_01: So there's there's tremendous technological challenges.
32:48
SPEAKER_01: Uh, some of them I think that are interesting purely from a technical standpoint, because who doesn't like a good technical challenge?
32:55
SPEAKER_01: And some of it is fascinating.
32:57
SPEAKER_01: We were just hosted by Boston Dynamics at the meeting for the actively controlled stability uh safety standard.
33:05
SPEAKER_01: And seeing all of their videos and movies, as well as in person seeing a number of their robots, really fascinating.
33:15
SPEAKER_01: It really is.
33:16
SPEAKER_01: If you're an engineer, it's a fabulous geek fest.
33:19
SPEAKER_01: You really just enjoy it just because of the challenge of it and uh and what they've accomplished, because it's really quite fascinating.
33:28
SPEAKER_01: But let's think about trying to get work done.
33:30
SPEAKER_00: From the perspective of somebody who's been in the industry for some time, watching the humanoid story specifically play out, uh, this rush to get them out, and and perhaps safety being an afterthought, whether you're almost seeing echoes of some of kind of the early days when you were working on this.
33:49
SPEAKER_01: This is the interesting thing in robotics.
33:51
SPEAKER_01: It's actually been the opposite in robotics.
33:53
SPEAKER_01: Uh so for example, uh, I was at the very first robot safety standard meeting where we all that's the only thing we called a robot, was the arm.
34:02
SPEAKER_01: And now we've got a myriad, a myriad of types of robots.
34:06
SPEAKER_01: And that was in January of 1982.
34:10
SPEAKER_01: So I can say my claim to fame is that I have outlived or outworked everybody else who's been involved in robot standardization.
34:21
SPEAKER_01: Doesn't mean I was better.
34:22
SPEAKER_01: I just outlasted them.
34:24
SPEAKER_01: And uh at that meeting, the big push was we, meaning automotive, we had all the major players in the room, uh, as well as Alcoa and another company, I can't remember.
34:38
SPEAKER_01: But anyway, all in the room say, we want to make use of robots, we don't want to be damaging humans doing this heavy work.
34:48
SPEAKER_01: We want to make use of robots, and we want to have automation that is more deployable, more reconfigurable, reprogrammable.
34:58
SPEAKER_01: We really want it.
34:59
SPEAKER_01: But the only way our factory workers are going to accept this is if there's a safety standard and it's met.
35:06
SPEAKER_01: So from the get-go, there was a push.
35:10
SPEAKER_01: Let's have a safety standard so that this technology will be accepted.
35:16
SPEAKER_01: So the concern was acceptance in the workplace.
35:21
SPEAKER_01: And I actually do see that now.
35:24
SPEAKER_01: There uh actually, one person who is in this field right now is Melanie Wise, and she had mentioned we want a safety standard because that is what is holding back some acceptance of these robots.
35:40
SPEAKER_01: So, in order for them to sell and make more robots, they need to have a safety standard that is thoughtful, well done, and so that they can meet it.
35:52
SPEAKER_01: Now, what's typically the challenge of safety standards is sometimes you don't like what you asked for.
36:01
SPEAKER_01: As an example, a number of by the time we get towards the end, uh the first robot safety standard was published in 1986, started in 82.
36:13
SPEAKER_01: And the next one was 92, and then the one after that was 99.
36:21
SPEAKER_01: At the 99, there were all these protests of, oh my god, look at what I've got to do for safety.
36:30
SPEAKER_01: So there were really there was some resistance um to the fact that the safety standard kept increasing requirements to provide for safety.
36:41
SPEAKER_01: And so there was resistance by 1999 saying, well, it seems like it's been pretty good, but that went through.
36:49
SPEAKER_01: And I can tell you that when we came time for the 20, we published the update to that in 2011, but in ISO.
36:58
SPEAKER_01: So our domestic standard went international, became an ISO standard.
37:03
SPEAKER_01: And at the publication of the 2011 standard, all I heard was, no, I can't give up the 99 standard.
37:10
SPEAKER_01: I love it, I use it, I know where everything is.
37:14
SPEAKER_01: So it'd be your standard that you become that you didn't like becomes your warm, cozy blanket that you love.
37:22
SPEAKER_01: And by the time that happens, it's time for an update.
37:25
SPEAKER_01: Why?
37:26
SPEAKER_01: Technology changes.
37:27
SPEAKER_01: And while the requirements become a bit, I'm not gonna say stricter, in some ways stricter, but there's more requirements if you want to do more than simply shut down things.
37:42
SPEAKER_01: And within robotics, we've really embraced the notion that we have a lot of ways we can reduce risk within our control systems, sensor systems, safety systems.
37:55
SPEAKER_01: Let's do it.
37:56
SPEAKER_01: So the robotics world, I really think leads the issue of safety and machinery.
38:04
SPEAKER_01: We do things before other industries.
38:07
SPEAKER_00: So it sounds like from the beginning that that labor was a big part of implementing safety standards.
38:14
SPEAKER_00: Is is that right?
38:15
SPEAKER_01: Yes, labor was a big part of resisting it for fear of worker safety.
38:20
SPEAKER_00: Of resisting automation.
38:21
SPEAKER_00: Yes.
38:22
SPEAKER_00: Yes.
38:22
SPEAKER_01: Um, because a fear of worker safety.
38:26
SPEAKER_01: If you do a job manually, you know what the risks are that you're putting yourself through.
38:31
SPEAKER_01: Uh, sometimes they could be high risk, sometimes, you know, lifting up something heavy all day long, sooner or later, you will have a back problem.
38:41
SPEAKER_01: But initially it was a concern for worker safety around robots and robot cells.
38:47
SPEAKER_01: Then later, workers embraced it.
38:49
SPEAKER_01: And why is that?
38:51
SPEAKER_01: The robots were being used to do the work that was really not good for workers between either breaking bodies or injuring bodies.
39:01
SPEAKER_01: Um, I don't know if you have you ever done welding.
39:05
SPEAKER_00: I haven't done welding.
39:06
SPEAKER_01: Okay.
39:06
SPEAKER_01: Well, I tried once, I was a horrible failure at it.
39:09
SPEAKER_01: But anyway, welding, holding the weld gun and the apparatus and so forth, and then, you know, it's heavy.
39:18
SPEAKER_01: Can you imagine manipulating this weld gun while you're contorting your body in and out of car bodies?
39:26
SPEAKER_01: Because it's all different angles and shapes and things.
39:29
SPEAKER_01: And probably wearing a mask at the time.
39:31
SPEAKER_01: And you're wearing a mask.
39:32
SPEAKER_01: And I could tell you the first that I tried welding for was a few hours as part of training, safety training, which was really good.
39:40
SPEAKER_01: But as soon as you did this to make your helmet drop, you're blind.
39:45
SPEAKER_01: And the only thing you see is the weld flash, but you can't really see the metal.
39:51
SPEAKER_01: So these people who do manual welding, they're really skilled artists.
39:57
SPEAKER_01: They're waiting to see a sheen of a puddle.
40:00
SPEAKER_01: It's a highly skilled in an art form.
40:04
SPEAKER_01: But the thing is, we're not bringing in new welders into the marketplace.
40:09
SPEAKER_01: Uh, I have yet to hear a child say, I want to grow up and be a welder.
40:16
SPEAKER_01: Now, people who do welding take great pride in it, and they should, because it is highly skilled, and as I said, also a bit of an art form.
40:24
SPEAKER_01: We don't have enough welders to do all the work that's needed.
40:28
SPEAKER_01: So, yes, we should be implementing more and more robotics.
40:33
SPEAKER_01: We don't have enough people to do what can be very tough work.
40:38
SPEAKER_01: I don't know if you've ever done wire crimping, uh, but that's actually okay.
40:42
SPEAKER_01: Well, you know, you're cutting conductors, but you might think, yeah, no problem.
40:47
SPEAKER_01: I can cut and strip this and I can put on this onto it.
40:52
SPEAKER_01: You really, that's a tough job on your hands and your forearms.
40:57
SPEAKER_01: And so I don't know of anybody that says, uh, I really want to do all those cable harnesses and hand terminate every single conductor.
41:08
SPEAKER_01: It's that's tough work.
41:11
SPEAKER_01: So we want to have robots that will help people to do work that is more enjoyable.
41:18
SPEAKER_01: And that takes advantage of that which makes people unique, which is their brains, their ability to think through problems, their abilities to try to manage all this automation.
41:31
SPEAKER_01: Um, I've had the pleasure of talking with a number of uh operators in factories that at one point in time they did the tasks that our robot was brought in to do.
41:42
SPEAKER_01: They're now robot programmers.
41:45
SPEAKER_01: And they said, I really had no interest in telling anybody I worked the line.
41:51
SPEAKER_01: I'm really proud right now to say I program robots and I manage these many robots and making sure that they're up and they're doing their work well.
42:04
SPEAKER_01: And I know what good work is because I did that before them.
42:08
SPEAKER_01: And I think that's a wonderful thing.
42:11
SPEAKER_01: We should be creating opportunities.
42:13
SPEAKER_00: It also speaks to another potential issue here.
42:17
SPEAKER_00: You know, we tend to think of roboticists as people who have been working in labs and maybe haven't gotten their hands dirty.
42:23
SPEAKER_00: But you you alluded to this safety training, and it sounds like for you there was a lot of value in learning how to weld.
42:31
SPEAKER_00: So, you know, perhaps part of implementing safety is making sure that the people who are actually building these systems know how those jobs on the floor operate.
42:43
SPEAKER_01: That's absolutely right.
42:45
SPEAKER_01: If uh someone who does integration work or is trying to brainstorm, really think through what would be cool, they should have a notion as to what are what are the steps and tasks that people do.
42:59
SPEAKER_01: Um the only thing, though, that everybody will say who's in automation, to have a robot doing something exactly like how a human did it is not truly the automated way to do it.
43:14
SPEAKER_01: Because humans do work which are based upon their own limitations of reach and strength.
43:22
SPEAKER_01: Whereas a robot could do something that a human body can't do reasonably, and it will then be quicker and faster.
43:32
SPEAKER_01: But a person might not be thinking that way.
43:36
SPEAKER_00: But but having done the job is a good way to uh understand those limitations.
43:42
SPEAKER_01: Uh, I think that's very true.
43:43
SPEAKER_01: And also welding is uh like a process operation.
43:47
SPEAKER_01: If you don't understand what a good weld is, how do you program a robot to do welding and do it well?
43:54
SPEAKER_01: So we need the people who have done that work to know what a good work product is.
44:00
SPEAKER_01: And then they need to use their smart brains and for them to think through how to get the robot to do it far better and quicker than how a person did it, because now they have the automation to do so.
44:16
SPEAKER_01: So, one example many years ago, robots did one sibling.
44:20
SPEAKER_01: They used to talk about turkey farms in an automobile line because it looked like the robots were turkeys bobbing down, right?
44:27
SPEAKER_01: And uh and they still had conveyor lines and so forth.
44:32
SPEAKER_01: Within that was in the 80s, beginning of the 90s.
44:38
SPEAKER_01: By the end of the 90s, you saw robots where they'd have an end effector which might have two or three positions.
44:46
SPEAKER_01: One position it might be welded, and the end effector would rotate, and then it would pick up and move the car body or apart further down the line for something else to be done.
45:00
SPEAKER_01: And what did that get rid of?
45:02
SPEAKER_01: That got rid of the overhead gantries, some things on the floor, and suddenly you have more floor space again.
45:10
SPEAKER_01: And then suddenly people have greater visibility to what is happening inside that fenced, that fenced area, because you've got big heavy things moving around.
45:23
SPEAKER_01: And uh it almost takes 10 years for somebody to have an aha moment of why are we still doing it that way?
45:32
SPEAKER_01: We had to do it that way because a person wasn't gonna lift up the car body as and run it through a factory.
45:41
SPEAKER_01: Um and so automotive really did pave the way for robotics initially.
45:47
SPEAKER_01: Um, but now the bigger or newer growth areas are outside of automotive or tangential to automotive.
45:56
SPEAKER_01: Uh, we're talking about logistics, warehousing, semiconductor uh assembly.
46:02
SPEAKER_01: I remember there was a gentleman from way back in the first safety standard.
46:08
SPEAKER_01: He was from Digital Equipment Corporation, which doesn't exist anymore.
46:12
SPEAKER_01: It's called DEC, and out of Massachusetts.
46:15
SPEAKER_01: And he said, I have a dream that a robot will be able to do assembly for PCBs as well as some other complex things to this resolution.
46:29
SPEAKER_01: And when it can do that, we will buy thousands annually.
46:36
SPEAKER_01: Now, of course, they're not in business anymore, but the that dream, that that need still exists.
46:43
SPEAKER_00: That is another interesting element of again getting back to the humanoid conversation.
46:48
SPEAKER_00: We were talking about dynamic stability, but if everybody's pushing towards, I know that this is a somewhat controversial term, but general purpose robots, like that's gonna, again, completely blow up the conversation, right?
47:03
SPEAKER_00: I mean, in a sense, it's easier to regulate or it's easier to create a safety standard for one robot that's designed to do one thing particularly well.
47:13
SPEAKER_01: For the humanoid, it almost has to be developed for a specific thing.
47:17
SPEAKER_01: It's gonna lift tote bins, it's going to have basically a scanner-reader, and it's going to scan-read barcodes everywhere, for example.
47:29
SPEAKER_01: Uh so uh that quote-unquote humanoid aspect ends up being more finished, more dedicated purpose or purpose build than the typical arm type robot.
47:45
SPEAKER_01: So the typical robot arm, that safety standard for that product goes to the tool plate.
47:53
SPEAKER_01: What you've got is the means for manipulation and movement, but you don't know what is at the end.
48:01
SPEAKER_01: At the end, could be a weld gun, could be a camera, could be a gripper, could be section cups.
48:07
SPEAKER_01: It's almost an endless array of things that it could be used for.
48:12
SPEAKER_01: So an industrial robot, as in that ARM world, is a general purpose machine readily deployable in many different ways and reconfigurable.
48:27
SPEAKER_01: But once you put it to use, that's when you know what you need to do for safety.
48:32
SPEAKER_01: And I'd say that's a little bit of a challenge.
48:35
SPEAKER_01: Uh it really is.
48:37
SPEAKER_00: So you again alluded to this the claim to fame that you're the um the the the I guess the one person who's still kind of in the industry from that uh 1982 meeting.
48:50
SPEAKER_00: Um, you know, there's the uh there's longevity conversation.
48:55
SPEAKER_00: That's a whole like the health conversation is a whole different one.
48:58
SPEAKER_00: But as far as actually staying in the industry, what do you attribute that longevity to?
49:03
SPEAKER_01: I I found it fascinating.
49:05
SPEAKER_01: So I uh robotics is fascinating.
49:07
SPEAKER_01: I I I have a bachelor of science in mechanical engineering and a master's of safety engineering and safety management.
49:16
SPEAKER_01: So I find the whole motion aspect fascinating and control of the motion to be fascinating.
49:26
SPEAKER_01: But my real passion is safety.
49:29
SPEAKER_01: I I really truly believe we can do better with our machinery, the design of our machinery, and I'm including robotics within that, uh, if we do it from the get-go.
49:44
SPEAKER_01: If we say, okay, we want to make sure we don't have sharp corners, we want to make sure we have smooth surfaces, we want to make sure that we can safely control speeds of the robot because that has uh a great influence on the potential of injury and the severity of injury.
50:06
SPEAKER_01: And so we keep challenging ourselves to find more and more ways by which to make more safety be built into robots and to be more easily integrated and deployed.
50:21
SPEAKER_01: Um, at one point in time, you could get huge control cabinets with banks, if you can believe it, of relays, which everybody you know would say, what's that now?
50:32
SPEAKER_01: Um inside them and with you know, wiring everywhere and all sorts of things.
50:39
SPEAKER_01: And nowadays that same cabinet could look minuscule and it has configurability to it, the safety built into it, where you don't have to do as much from an external standpoint, and you have a longer life.
50:55
SPEAKER_01: So the way we provided safety for way back days with stopping, and we talked about that each it's nothing else you can do with stop.
51:03
SPEAKER_01: If it stops, it's better than if it's moving, right?
51:06
SPEAKER_01: And with stopping, the way that that was accomplished was with contactors.
51:11
SPEAKER_01: If every time you make the robot stop, you actually use those contactors, they don't have a long life.
51:22
SPEAKER_01: They have to be replaced.
51:24
SPEAKER_01: And those big contactors for those big heavy robots were really expensive.
51:30
SPEAKER_01: As soon as we didn't have to use contactors, meaning we could stop but keep power retained, that just goes to the bottom line.
51:42
SPEAKER_01: You could say almost that some robots got replaced because you didn't have to keep replacing contactors.
51:50
SPEAKER_01: And who would have ever thought of that at one point in time?
51:55
SPEAKER_00: So you don't expect to be stopping anytime soon.
51:58
SPEAKER_01: Me?
51:59
SPEAKER_01: Uh no, I'm I when I say I was the one continuous person, the outlasted part of it is I'm older.
52:07
SPEAKER_01: And uh at some point in time I will retire.
52:10
SPEAKER_00: Sure.
52:10
SPEAKER_01: But I am still really enjoying myself uh within the robotics field and seeing all of the advancements and the improvements.
52:20
SPEAKER_01: And I also still feel like I have got to be carrying that torch and saying we can do better.
52:29
SPEAKER_01: Just because what we did was good enough 40 years ago, it doesn't make it good today.
52:36
SPEAKER_01: And I can say the same thing from the 2011 standard that uh came out that everybody cried about the 1999.
52:44
SPEAKER_01: The difference between then and now is huge.
52:48
SPEAKER_01: And we can continue to do better.
52:50
SPEAKER_01: We just have to push ourselves that way.
52:53
SPEAKER_01: We've been fortunate that we've had a lot of assistance, not just with technology, but also other standards.
53:02
SPEAKER_01: The whole concept of functional safety did not exist when we started in robotics.
53:09
SPEAKER_01: It does now.
53:11
SPEAKER_01: We do so much with control systems, sensors, and actuators now that are done safely.
53:19
SPEAKER_01: It's inconceivable to the person that I was 15 years ago.
53:25
SPEAKER_00: Well, uh, it's been an absolute pleasure, Roberta.
53:28
SPEAKER_00: Thank you so much for joining us.
53:30
SPEAKER_01: Well, thank you.
53:31
SPEAKER_00: There you go.
53:32
SPEAKER_00: That was the great Roberta Nelson Shea of Universal Robotics.
53:36
SPEAKER_00: I look forward to continuing this conversation in some form at IRSC November 3rd through 5th.
53:44
SPEAKER_00: Thanks so much to her, thanks so much to you for watching and making it all the way to the outro of this episode.
53:52
SPEAKER_00: If you did, I assume that means you like it, or at least you're hate watching, but uh thanks for engaging either way.
53:59
SPEAKER_00: Please like and subscribe, and we will see you next week for another episode of Automated.