Home » Education » The Evolution Of A Roboticist – Mark Tilden

The Evolution Of A Roboticist – Mark Tilden

If there is celebrity-status in robotics,
surely Mark Tilden merits entry to that club
and for good reason.

 

He is a standout not only for what he has accomplished in the world of robotics but for who he is. Not content with turning the robot world on its head by establishing a brand-new school of thought known as BEAM robots (derived from Biology, Electronics, Aesthetics, and Mechanics, see the July-August Robot , page 78), Tilden has since gone on to succeed in academic circles, build robots for the US militarys Los Alamos Labs and for NASAs space program. And, get thishe can even lay claim to stupendous commercial success with his robotswith 22 million sold and counting!Just as BEAM broke new ground, changing our understanding of how robots can be constructed, Tilden wants to change the way roboticists are perceived by the public and by themselves. Unlike popular stereotypes of roboticists, Tilden is not satisfied to sit in a basement lab and be a squinter (hes done enough of that, thank you). Today he works as a consultant for several companies around the globescientific, education, consumer electronics and otherwise, from a private lab in a high-rise in Hong Kong overlooking the Kowloon Walled City Park, dreaming up his next creations. By any measure he is sitting in an enviable place in the world of robotics.

INSPIRING OTHERS

Tilden beams with excitement when he discusses robotics and has a generous spirit for sharing his ideas and enthusiasm with the robotics community. He wants the community to follow his lead or at least not perpetuate the image that they need to be socially awkward geniuses and suffer for their craft, hiding in dimly lit back rooms. If Victor Frankenstein was alive today, hed have a PR agent, jokes Tilden.

After all, Tilden believes roboticists should feel the right to strut a little. Their chosen field doesnt just encompass the knowledge of one science. It encompasses all of them. Further, it doesnt require just a little artistic flare, but a lot of itfrom aesthetics to movement to personality. The robots of the future will need to embody all the arts and sciences if they are to capture the imagination of the public at large and be accepted and embraced. Only then will they survive and ultimately evolve.

ROBOSAPIEN IMITATING LIFE

Tildens personality and attitude certainly translates into his work. The first Robosapien he built for Wowwee was really Tildens break-out product in terms of commercial success. Despite being encased in the shell of a toy, it was what the market was craving. It was a humanoid robot that you could program and even teach a robotics course with. And yet you could still have it walk around a room burping and farting. That was me, Tilden jokes. All that was missing was the hat and beard!

The signature fedora atop his head, beard and big laugh counterbalance what lies underneath a very serious, accomplished scientist who has succeeded on many levels where countless others have failed. The breadth of his success and experience gives him a unique perspective on a budding industry that he sees moving from its Romantic Period into its Golden Age.

We are where the auto and plane industries were 100 years ago, Tilden observes. He notes that just as there were hundreds of car companies at the beginning of the last century and many more ideas quickly emerging that would improve the technology, robotics, today, is at that same inflection point. What has already been accomplished today will usher in the true promise of the robots to come. Never before has that fulfillment been so tantalizingly close and yet so frustratingly beyond our grasp. Tilden insists that much of what is possible can be achieved today, but it will be up to market forces to determine how fast robots move into the mainstream.


Oh the places youll go! Tilden displays his first BEAM robot, the Rover 1.0 with his latest BEAM-inspired creation for Wowwee, the voice-controlled Joebot.
Tildens thinking on the three laws of robotics is a great metaphor that captures how he approaches his work: with candor, brutal honesty, and a desire to strip it all down to its bare essentials to see whats worth keeping and whats worth leaving behind.

EARLY ROBOT YEARS

When Tilden first dove deeply into the study of robotics in the late 1980s he had already been building mechanisms for a while. Like many roboticists, he was faced with three major problems: One was cost. The incredible cost of the hobbyeven back thenwas prohibitive. Second was the complexity of the hobby, and third was the fact that nearly every prediction about robots had dystopian visions of what the hobby would lead to.

So the problem with being a roboticist was that you were lumped in with Loch Ness Monsters sexers in terms of overall popularity. But when you did actually make something that was pretty cool and worked, it turned out all right. It was pretty interesting to see how people reacted to it, Tilden said.

It wasnt until Rodney Brooks spoke at Waterloo University where Tilden was studying in the late80s

RE-STATING THE ROBOT LAWS

When Tilden became more established in robotics he would do what other roboticists and science fiction writers had not lightly done before or would even think of doinghe tinkered with Asimovs Three Laws of Robotics, which robotics philosophers have regarded almost as Holy Commandments given from on high.

ASIMOVS THREE LAWS

1. A robot may not injure a human being, or, through inaction allow a human being to come to harm.
2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
Tilden wrote his own Three Laws that seemed to him to be more practical.


Mark with one of many minefinder robot prototypes in Yuma, AZ, 1997.

TILDENS THREE LAWS:

1. A robot must protect its existence at all costs.
2. A robot must obtain and maintain access to a power source.
3. A robot must continually search for better power sources.
Or otherwise put into Tilden-Speak (which like his BEAM robots, representssimpler and more direct solutions), a robot must: 1. Protect thyself.
2. Feed thyself.
3. Move thyself to better real estate.

Tildens thinking on the three laws of robotics is a great metaphor that captures how he approaches his work: with candor, brutal honesty, and a desire to strip it all down to its bare essentials to see whats worth keeping and whats worth leaving behind.

EARLY ROBOT YEARS

When Tilden first dove deeply into the study of robotics in the late 1980s he had already been building mechanisms for a while. Like many roboticists, he was faced with three major problems: One was cost. The incredible cost of the hobbyeven back thenwas prohibitive. Second was the complexity of the hobby, and third was the fact that nearly every prediction about robots had dystopian visions of what the hobby would lead to.

So the problem with being a roboticist was that you were lumped in with Loch Ness Monsters sexers in terms of overall popularity. But when you did actually make something that was pretty cool and worked, it turned out all right. It was pretty interesting to see how people reacted to it, Tilden said.

It wasnt until Rodney Brooks spoke at Waterloo University where Tilden was studying in the late80s that Tilden had the chance to talk with Brooks, and when he did, his vision for BEAM robotics took off. Brooks had written a paper and later a book titled Fast, Cheap and Out of Control that advocated the use of simple, inexpensive robot platforms with relatively little computational resources that could work as a team. If, for example, one or two of a group working on the moon broke down, it would not impact the teams ability to carry out the mission. Tilden had already been thinking in these terms and wanted to take this concept to the next level.I basically asked,Whats the simplest, most performance-effective way of building robots? From there the concept of BEAM robotics took root in his mind and Tilden started to challenge the idea that robots needed to have complex artificial intelligence.

Walkman and Walkman, 1992.
During this time Tilden formulated most of his basic philosophies at the core of BEAM robotics and asked his most important questions. Mother Nature built human kind out of nothing. So why are robots so inherently complex? Why is it so hard to do artificially what Mother Nature does so easily and naturally? So this got me thinking about a completely different radical philosophy, he remembers. In 1989 while building the first solar roller, my mind was simultaneously churning out the fundamentals of BEAM minimalist philosophy, knowing how conventional roboticists would have difficulties with it. I could understand the problemwhy make a robot simple and elegant when you can make it wonderful and complex?


BEAM by the hundreds!

PROTOTYPING CHALLENGES

For Tilden, the real challenge facing roboticists isnt finding ideas but getting them into play. The real problem with modern robotics is it takes two or three years to get things organized and get everything sorted and simulated on a computer before anything gets done. Back at the Los Alamos Lab people would come to me with an idea and Id be able to come back with something by lunch, Tilden says of his time working for the military. He is still working that way with Wowwee. He builds BEAM prototypes that later on are run through their paces by engineers so that the movements produced by analog beam circuitry can be digitized and emulated by a microchip.

LOS ALAMOS LAB YEARS

It was at Los Alamos National Labs in the 1990s that he took some time to reflect on robotics, the philosophy driving BEAM and perhaps more importantly where he was going within the field of robotics. These were formative years not only because he was practicing his craft of building robots but also he could reflect on his bottom up approach. Over that time he saw in museums a lot of expensive robots collecting dust from top-down developers. This was partially because moving from the concept stage to the real thing isnt always easy and because the mechanics in robots are more difficult to engineer than most believe.

Today biomorphic concepts are a legitimate science, but back then Tilden recalls that no one else was asking the types of questions he was. I asked a simple question when I was formulating BEAM: If robots were evolved from the simplest form to a complex form (using a human being as an exampleprocess of evolution), how long would it take to make a walking, thinking, listening, talking humanoid? And the answer is: 20 years. This tantalizing speculation by Tilden may foreshadow a faster pace of development than most forecasters have envisioned. That is about the length of time from his meeting Rodney Brooks at Waterloo University in 1989 till today.

WATERSHED MOMENT – INSPIRE THE KIDS

One of his last projects for the military at Los Alamos Labs was developing robots that could crawl into fields and detect mines. Tilden jokes that the challenge was to build a robot smart enough and capable enough to pull itself out of mud and puddles and find unexploded mines without, of course, being smart enough to realize that its job was actually to find unexploded mines!

Tilden then had his next revelation that would change the course of his career and the evolution of BEAM. If you were to drop these cute little robots into a field, guess what? The local kids would pick them up. The natural tendency for humans is to claim things in their domain. It was one of the major reasons to branch out into toys. Putting them out wild and free unfortunately means that they are removed from doing what they are supposed to do [which is to inspire the evolution of robotics]. So why dont you put them in the hands of kids on Christmas morning and wind up giving them a sense of wonder and you can evolve the technology that way, Tilden says, of his decision to try to inspire the next generation of roboticists. It was a watershed moment.


Another alien design, the Roboquad got its inspiration from Mars exploration robot designs Tilden pioneered.

Tildens 2008 Femisapien Robot Companion illustrates beautifully that form does indeed follow function.


HITS & MISSES

Not all of Tildens ideas and robots have been runaway successes. The RS Media, for example, was a more sophisticated version of the Robosapien V2 robot and was built with a Linux brain. This was primarily targeted at the top-down roboticists who needed a better body to go with their robots computational capabilities. Building better cradles for robot brains is something Tilden alwasy sought to do..

The robot worked but was a commercial disaster, in part because it was twice as expensive as many similar Wowwee robots then sold (though still costing less than a quarter of Asimo-style servo robots). You can buy them now on eBay for twice their original cost because they are no longer available. Its like the old saying:I dont care if vengeance is the Lords as long as its quick and I can watch. In 20 years Ive been able to watch a lot!

ROBOBOA

Of all the robots he has created, Tildens most favorite is the Roboboa and that was one that met the most market resistance. Tilden loves the Roboboa because it was a true alien life form. It also was a purposeful robot that could be used as a light, wake-up alarm and room guard as well as be programmed to navigate around a room or follow you around using the wheel at the base of its tail for locomotion. At the time, Tilden believed it would be a natural in college dorms (if you look in the background of the television sitcom The Big Bang Theory you can see various Wowwee-brand robots decorating the shelves).

The Roboboa was released at the same time as the Roboquad (a design Tilden originally created for JPL for Mars exploration, a prototype precursor to the Spirit and Opportunity rovers) because Tilden wanted to see if the marketplace would take to alien-like robots. The Roboquad was adequately bug-like and attracted interest. However, the marketplace was not so kind to the Roboboa (it was released in Asia but not in the US). It did not resonate with distributors. It confused the market. They didnt know what part of the store it belonged in: The lighting section or the toy aisle, Tilden said. In modern markets you cant have a combination product. The only ones who could do this were the Swiss with the army knife.The natural selection which is the market is also challenging the recently released Joebot. In a way, the voice activated Joebot is not up against robots from competitors, but rather the original Robosapien, which still sells tens of thousands units a year. Plus it was released during a recession when a $100 robot competes with many $20 – $30 toys on the market.

A FOURTH LAW OF ROBOTICS? (WELL, SORT OF)

If Tilden had a fourth law of robotics, it might be Robots must be desirable and useful to humans or to put it more in the words of Tilden: Have thyself befriended. Market forces serve as Darwinian natural selection for commercial robots, and market acceptance is an essential ingredient for success. It is something that Tilden is paying closer attention to as he develops his next robots, where he says he has over a dozen prototypes ready for release, but he wants to make sure the market isready for them and the products are right..


Tilden with RoboBoa, a seeing, sinuous alien robot that could be a desk lamp, clock alarm, personal guard, and intelligent roaming R2D2-like companion. It never enjoyed a US release.
I was a hobbyist, then a scientist, and now Im an industrialist, Tilden muses, hinting that the lessons learned from BEAM and robotics is nothing compared to the lessons the market has thrown at him. Those experiences have provided him with an amazing perspective on the industry. What is interesting about that is realizing that now we have this incredible technology. And where do we take it? The fact is its completely different from where you would expect it. I dont speculate, I just have 20 years of real-world experience seeing what my robots can actually do, at cost.“People get nervous seeing robots moving at two meters a second despite the fact that people do it all the time”Tilden says he is not a futurist. He says there are enough guys in suits running around at conferences willing to talk about what role robotics will play in years to come. Instead he speaks about what he has and what works. He says there are many things robots can accomplish in the here and now, but without buy-in from consumers, it will mean that many a good robot and robot company will have short lives.
 

BEAM PHILOSOPHY IN PLAY

If there was any doubt, there is a lot of BEAM in those Wowwee robots. And while those robots may be classified as toys for children there are some serious robot and business concepts at play in their designs. What does it take to create a robot that survives the environment? Thats easy. But to survive the market? Thats hard. To survive a ten-year old boy, you need something nuclear! To accommodate the latter, Tilden leverages his experience at NASA with something called Gold Plating.

You dont over design something by 600px, you over-design it by 200%. The Robosapien looks like a football player so he can survive the horrors of the marketplace, Tilden comments.

The basics of biomorphic design (the scientific branch of BEAM robotics) determine the shape, speed, functionality, and energy efficiency of these robots. While those not familiar with robots may claim the Wowwee line-up devour batteries (see those Amazon reviews), the truth is, they are extremely energy efficient for what they do and how long they last on a single battery round. The BEAM design in Wowwee products like the first Robosapien and the V2 mean that with its mechanical structure each time it takes a step it helps to recharge the battery. The battery weight is placed strategically in the soles of the feet to also serve as a ballast to keep the robot upright. As a point of comparison, many more expensive humanoid robots of equal size or smaller have a battery life that is a fraction of that time.


Built as a carriage for “top-down” roboticists with a Linux brain, the RS Media was packed with programming features. It struggled commercially.

1980S INDUSTRIAL DUMPSTER DIVING

The beauty of BEAM tech is that it also utilizes obsolete outdated components. It is almost an industrial form of Dumpster Diving. In addition to all the components in the Wowwee robot line being toy-grade they are also basically 1980s tech at best. I often think if we were to go back in time to build a robot company in the 1980s, what would it look like? Well, it would look a lot like what weve done, Tilden jokes.

It is amazing when you consider some of the capabilities and functionalities of some of the Wowwee robots. For example, Robosapien V2 can track and distinguish among the colors red, blue and green by means of a radial retina. It doesnt think in raster-scan technology, it uses biologically inspired circular-scanning like our eyes do, Tilden explains.

 

BUILT FOR THE HACKER

For those who may have suspected Tildens robots were, in part, platforms ready to be hacked, you are correct. Every single robot I build is meant for the enthusiast out there, Tilden says. The wires are color coded, the robots can be taken apart using a single screwdriver and you can get replacement parts for the robots at the local Radio Shack.

For those more keyboard orientated, Tildens Wowwee robots have even attracted the attention of programmers and hackers like Robert Oschler, who developed the free Robodance software through which you can program Tildens robots. Occasionally you find someone with a real passion for autonomous robotics and the skills to move the community forward; Robert Oschler is one such. For cost reasons I could never give my robots the sophisticated front-end computer interface I knew hackers would like, so its great when something like Robodance is available to take the field to the next level. His website and videos are amazing as well, and you have to give him and those like him kudos for pushing the philosophy, says Tilden.

THERE AND BACK AGAIN: A RETURN TO PROFESSIONAL ROBOTICS

Tildens journey as a roboticist in a way may come full circle in the next few years. After all, he got his start making robots for adults needing robotic solutions for the military and space exploration, and that led him to turn towards manufacturing toys to inspire a new generation of roboticists. However, safety requirements and other market forces may cause Tilden to venture from creating robotic toys for children that their fathers can enjoy to taking BEAM-inspired robots to the next level: Producing robots for parents that the kids are not allowed to touch. The fact is it takes around 100 safety rules to get a robot into space but over 3,000 to make sure a toy is safe on Christmas morning, Tilden comments.


Tilden poses with his posse.

The Robosapien was one of Tildens most successful robots.
Tilden wont go into many details but hints that he is tackling the social and mechanical issues surrounding constructing human-sized robots. Among the more obvious challenges are durability (making sure they survive a fall without cracking their outer shell), energy efficiency and minimizing fears people have of life-sized robots. People get nervous seeing robots moving at two meters a second despite the fact that people do it all the time, Tilden says.

THE SINGULARITY WILL NOT BE NEAR

Perhaps Tilden is too engrossed in the bottom up approach of robotics, but he just doesnt buy into a future where robots take over the world. You cant market a robot with evil intentions for the same reason you cant sell an evil cellphone. As soon as a machine loses critical benefits there is one rule: Machines that piss poople off get murdered. If you dont think so, just look at your drawer filled with dead, discarded electronic devices, resting comfortably. he candidly says, adding he knows this view flies in the face of what many students are learning from professors today. What BEAM teaches by practicality and actuality is exactly opposite to what they teach in universities today where they cant wait to get to the Hal 9000 fast enough, Tilden observes. In Tildens eyes, the Hal 9000 vision will not come to pass mainly because of economics. Any AI project that for incremental ability requires exponential increase in cost and complexity is a dead end.

Those who think otherwise, Tilden says, ensure the future of his ideas and that of BEAM robotics and its practitioners.

LinksBEAM bot papers, circuits, and articles by Wilf Rigter, http://wilf.solarbotics.net/
BEAM bot pics and links, http://solarbotics.net/
Mark Tilden History http://en.wikipedia.org/wiki/Mark_Tilden http://www.beam-wiki.org/wiki/Tilden
Robosapien Education packages for schools and museums, http://www.q4technologies.com/
RobotsRule (Robot reviews, videos, and Robodance control software), www.robotsrule.com
Solarbotics (BEAM kits, plans, parts, modular gearboxes), www.solarbotics.com, (403) 232-626
WowWee, www.wowwee.com, http://wowwee.com/en/products/toys/rob ots/robotics (800) 310-3033
Wowwee hackers Web Community, http://www.robocommunity.com
 

Words by Thomas Marsh