The Story Behind The Story: Mobile Robots – Jeanne Dietsch
Robots are neat - no question, and we love working with them as much as anyone - maybe more. The only thing we like better is the people behind the robots. Understanding what makes a robot tick is fun. Understanding what makes a robot creator tick is an absolute blast. So, when Jeanne Dietsch dropped by Robots Dreams (electronically) we jumped at the chance to hear the story behind Mobile Robots.
Discounting the toy companies that make toy robots, there are extremely few companies in the robot game that can boast about having thousands of robots in active use at customer sites.
There are even fewer that can trace their roots back ten years or more. Mobile Robots is one of the rare few that can honestly make both of those claims.
First, here’s the ‘official’ backgrounder on Ms. Dietsch:
JEANNE DIETSCH, CEO & CO-FOUNDER. Ms. Dietsch began her career in educational marketing and was chosen at age 26 to be the educational products manager of the Weber Costello company. Her entrepreneurial career started at age 29 in partnership with IDG Chairman Pat McGovern in the TALMIS affiliate of IDG. Dietsch independently founded the first market research company specializing in e-commerce in 1995 and co-founded the robotics group later that year. As co-founder and visionary of MobileRobots, Dietsch sets strategic and tactical directions, builds partnerships and heads sales, marketing and administrative efforts.
We asked her to explain some of the background behind the formation of Mobile Robots, Inc. and she generously obliged. Here’s what she had to say-
“We started back in 1995 as a garage operation, selling robots for a company called "Real World Interface" (RWI), owned by Grinnell More, who got his start ten years earlier, building robots for Rodney Brooks. Grinnell had expanded his customers beyond MIT to CMU and Sebastian Thrun, who was still in Germany at that point, and a handful of others around the world, but was still an engineering firm: no documentation, no brochures, etc. And the least expensive robot was $20,000! So we helped him turn his operation into a business and co-launched a new, little $2,000 robot called the Pioneer from our company, ActivMedia Robotics.
The Pioneer changed university robotics programs. Suddenly there was a programmable robot you didn't have to build yourself and nearly any university could afford at least one. In addition, you didn't have to spend half your time repairing it. The Pioneer moved forward intelligent robotics software dramatically. But Grinnell was much more interested in the defense industry; he sold his company to iRobot to work on the PackBot; we retained the Pioneer name and started building our own robots, starting with the Pioneer2-DX in 1998, which we rapidly followed by the P2-AT all-terrain in 1999.
A year or two later, the other major robot supplier, Nomadics, quit the business; Nomadics was the only company that made a robot large enough to carry a Mitsubishi industrial arm, so we started building the PowerBot(tm), with its 100kg payload. And we started adding an upper extension to the Pioneer-DXs, for people doing human interface research; we called that system the PeopleBot(tm). PeopleBot is very popular as a tour guide in museum exhibits. The next robot was a tiny one for researchers in collaboration, called AmigoBot(tm); this is the robot used by two teams in the DARPA challenge in which teams had to keep 100 robots performing a collaborative task for 24 hours.
Then we turned our attention to software. We'd been using some licensed software that was very buggy and difficult to use, so we decided to create our own mid-level API, with we call the Advanced Robotics Interface for Applications (ARIA). We wanted to keep the maintenance down, so we provide it in source under the GPL terms. People raved about it.
Then we set our sights on autonmous mapping, navigation and GUI interface, starting about four years ago, while we were building an intelligent wheelchair through National Institutes of Health. Again, we began by licensing the best software available, but much of it was buggy and unreliable, so we had to replace all the localization and navigation. It was grueling, but we finally had a research version, which we spent a few more years on. For commercial applications, it had to build maps and have the robot navigate reliably under myriad conditions. Then we released our MobileEyes(tm) control GUI, which lets you connect to the robot, and download its maps. You can see images from its camera, click on the floor plan to send it places, etc. MobileEyes was soon followed by MobilePlanner(tm) software, which lets you edit maps, set up goals and routes and assign tasks to the robot. Then we started communicating among robots so that fleets of them would operate efficiently together.
With these capabilities, we could actually sell a useful robot application, which we have just begun doing. We designed a self-charging robot, robust enough to run 24X7, called PatrolBot. PatrolBot can monitor buildings' safety and security and make deliveries. We also took the brains of the system and put it in a box we call ARCSinside(tm); this we sell to manufacturers of large factory equipment -- since we don't want to build forklifts and so forth.
Now we're just releasing our commercial developer software layers. This includes SetNetGo(tm) for network and system set-up, and our ARCSinside Developer Modules for Systems Integration, Onboard Integration and Interface Development. These modules enable commercial value-added resellers to integrate sensors, effectors, third-party software, external controls and other extensibility to the robots. We're also expanding the multi-robot capabilities and attacking the issue of highly dynamic spaces in buildings. Of course, those are only the announced projects under development; there are many others. And, speaking of announcements... watch April 10th something that's big and bold and loves the out-of-doors!”
You might also enjoy: