Investing

Kickstarter Project Of The Day

by EricWind

Team 696, the robotics team at Clark Magnet High School in La Crescenta, CA, is asking for your help on Kickstarter:

We are Team 696, aka The Circuit Breakers, the Robotics Team from Clark Magnet High School in La Crescenta, CA. We are hard-working, enthusiastic students devoted to creating robots for the purpose of spreading awareness of science and technology-based education. Our team comes together to build a new robot each year. In 2013, we plan to build a robot to participate in two FIRST Regional Competitions: one in Long Beach, CA, and the other in San Bernardino, CA.

Our team strives for excellence in all areas; unfortunately, our California school’s budget limits our ability to attain full robotic perfection, and we are always thankful for additional funding to achieve our goals. By supporting us, you will help fund our official 2013 robot, improvements to our engineering lab, and most importantly, the expansion of mentoring and outreach programs within our community.

They have 37 days to go and are almost to their goal of $2,000. The perks range from personalized dog-tags all the way up to being flown out to the robotics competition with The Circuit Breakers.

Related articles
Enhanced by Zemanta

LineBot Launched Drink Runner On Kickstarter

by EricWind

Robotics start-up LineBot started a Kickstarter campaign to get their Drink Runner bot out there. What it is:

On December 10, 2012, LineBot will launch a Kickstarter project to raise funds for Drink Runner, a line-following service robot that delivers individual drinks in total darkness non-stop for up to 4 hours per charge. Swarms of them act like a conveyor belt, shuttling drinks along a closed-loop course.

The full press release here and the Kickstarter page here.

Three Questions for Chetan Kapoor – Agile Planet

by JonasLamis

Chetan Kapoor is a roboticist at The University of Texas and CEO of Agile Planet.  We spoke about how his firm’s control software might change the world in the years ahead, what is driving robotics growth, and why business should take note.

Three Questions for Colin Angle – iRobot

by JonasLamis

SciVestor spoke with Colin Angle, co-founder and CEO of iRobot Corporation. We asked him about how iRobot will change the world, the importance of Moore’s Law, and why the business community should be interested. We also got his take on the state of the emerging robotics industry.

MIT’s Technology Awakens Mobile Devices–Robots Next?

by RayRenteria

The SynTactic Analysis using Reversible Transformations (“START”) system is a super information-finding technology that was developed by Boris Katz and his associates of the InfoLab Group at the MIT Computer Science and Artificial Intelligence Laboratory. It’s a system that’s been running since 1993–long before Google came along.

The system has two syntactic modules: One for learning things and one for answering questions. The learning module is based on START’s ability to parse the English language, annotate it, and store it for later retrieval. Theoretically, START could be pointed to Wikipedia to learn everything that is in there. The catch (for now) is that START requires well-formed English sentences from which to learn.

The retrieval part is what is most impressive. Instead of giving the user a bunch of search results, it provides a highly precise answer:

I asked, “What is a Euro?” START answered,

EURO

Noun

  • S: (n) euro (the basic monetary unit of most members of the European Union (introduced in 1999); in 2002 twelve European nations (Germany, France, Belgium, Luxembourg, the Netherlands, Italy, Spain, Portugal, Ireland, Greece, Austria, Finland) adopted the euro as their basic unit of money and abandoned their traditional currencies)

Source: WordNet

Perhaps more impressively, START says “I don’t know” when it can’t provide a succinct answer instead of giving me a bunch of useless links that I have to pore through. I asked, “How many feathers in a pound?” To which START responded,

Unfortunately, I don’t know how many feathers there are in a pound.

Finally, I gave it a simple command. I typed in, “Convert 100 dollars to Euros.” START replied,

As of Thursday, April 24, 2008; 11:00:53pm, there are 63.69 Euros in 100 United States Dollars.

Source: XE.com

START was able to perform a basic mathematical operation in this case on the results of the data. It is with this operation that START begins to cross the threshold from super knowledge-finder to a machine-enabling technology.

Katz and his team have developed a prototype system called StartMobile in which the technology is applied to mobile devices. It leverages a knowledge base that is maintained partially on the mobile device and partially on the internet. While the START system on the internet knows a little about Art, Science, Culture, and Geography, StartMobile understands concepts of a mobile device such as calendar, contacts, camera, settings, and so on. This knowledge allows the user to make commands like, “Take a high-res picture using the flash in 10 seconds.”

In a more advanced scenario, StartMobile was told, “Remind my mother to take her medicine at 3PM.” The device was able to understand who “my mother” is, what “remind” means, and the implication that the method to remind my mother is to create a reminder in my mother’s cell phone. It contrived a sentence for her reminder that reads “Take your medicine at 3pm” changing “her” to “your.” Here’s rendering of the flow:

startmobile1.jpg

This project is supported in part by Nokia which is a good investment on Nokia’s part. StartMobile is a prime and marketable candidate for licensing by vendors to place into everyday things like cars, phones, and computers and into not so everyday things like robots. The ability to understand and apply natural language reduces the friction of mainstream personal robot adoption by helping them to successfully interact with humans. Here we see a bridge that connects the two.

—

Boris Katz is a Principal Research Scientist and Head of the InfoLab Group at MIT’s Computer Science and Artificial Intelligence Laboratory. His research interests include natural language understanding and generation, intelligent multimedia information access, knowledge representation, human computer interaction, and machine learning. He has authored more than 50 publications and 2 U.S. Patents.

Source:

Colin Angle on the Business(es) of Robotics

by JonasLamis

Blogging from RoboBusiness in Pittsburgh this week. Much to my disappointment, the Convention Center does not have wireless access, so blog postings are coming out when possible.

Colin Angle, Co Founder and CEO of iRobot spent 45 minutes romping through robotics market business concepts. He presented 21 (I only captured 20 for some reason) – with a detailed to very quick take on each as to its viability.
He ranked each of his concepts on a 2 by 2 chart with the horizontal access representing Revenue Potential (can you see the concept), and the vertical access representing Margin Opportunity (can you make money at the concept?)

Case 1: Industrial cleaning machines

$80B in 2000 in North America cleaning tile (hard floors)
It’s hard to beat the labor costs. You need huge space to be cost effective. Colin spent a year on this space and documented a 42% savings over conventional labor, but business models around replacing human labor are fraught with risk. There is huge gaming by cleaning companies, and lack of objectivity by building supervisors. iRobot is not involved in this market, but Intellibot is working in it.

Case 2: Aggressively priced home cleaning robots

The Roomba and it’s kin dominate this market. The price point anchored to price of vacuum. There are lots of iRobot copies – especially in Korea. The marketplace is competitive and margin challenged.
Total 2007 domestic vacuum market – $2.4 B. Robots have captured 5%. (Homeworld business review)

Case 3: Expensive home robots

Popular in Korea. Much skepticism. In some markets, consumers enter with a base model (Cars, watches, grills etc) and over time some upgrade to a premium product. This has not yet been demonstrated in robot vacuums.

Case 4: Robot Pets

American consumers spend billions on their pets. Is there a market for expensive robot replacements? Furby (cheap) sold 40 million units. 20M of which were sold to adults for adult use. However the expensive Quiro and Aibo were not successful, and Pleo just out.
[As an aside, Bob Christopher, the CEO of Ugobe – maker of Pleo – scratched his appearance at RoboBusiness. And thank goodness as I was up against him in the same time slot ?]

Case 5: Robot Toys

Treat robot like toy industry. You do not have special treatment because you are a robot. New companies succeed by pushing costs to the factory. If you can show up at the factory in China with a design on the back of a napkin, and convince them to build it for you and you will buy it from them if they are successful, you might be able to be profitable. The average cost to develop a new toy is less than $40,000, and 85% don’t make it past their first anniversary.

Case 6: Above Ground Oil Storage Tank Inspection

Here is another area where Colin claims to have spent way too much time. He built a product that should have been successful. After all, regulates mandate regular self-inspection. However he discovered that the regulations are not taken seriously. Companies that operate Above oil tanks rely on “self insurance”. They build in safety berms if there is a tank rupture, and decommission the tank prior to the end of it’s known good lifespan. If there is a rupture, they pay the fine.

Case 7: Military UAV

Unmanned aerial vehicles is a growing market. It has a high price point expectation and over 100 companies involved.

Case 8: Unmanned Underwater Vehicle

Also a high price point expectation, this market is small but growing. A key is the relative lack of obstacles to run into. There are less than 10 companies in this market.

Case 9: Military UGV

The military unmanned ground vehicle market is a key market for iRobot. There are moderate barriers to entry and a relatively low price point compared to UAV and UUV. Challenges exist in both ruggedness and obstacle avoidance. This market has moderate growth opportunities according to Colin.

Case 10: Knock off Military robots

Not a good strategy. Colin had to vent about the RFx controversy. His observations: Steal iRobot IP, get sued, RIP.

Case 11: Oil Well bore robots

Colin was very excited about this market. The biggest, oil reserves are in deep water and Robot boring and exploring robots enable something new. I’d expect to see iRobot make a play in this market.
[I missed case 12]

Case 13: Vending machines

The most successful robots of all time.

Case 14: Material Handling

Material Handling robots are making a comeback. The offer concrete measures of performance. Very interesting implementations by KIVA systems, Seegrid, and Atheon.

Case 15 – Virtual presence robots

An emerging market. Builds on the notion of personal video conferencing. Can be applied to both home (jump into your kids robot to talk to them) and business (doctor interviews a patient and checks vitals from 1000 miles away).

[Aside – I saw a presentation later in the day about TraumaPod – the operating room of the future. It had a heavy telepresence capability. I’d look for Intuitive Surgical to make noise here.]

Case 16: Animatronics robots

Disney themeparks have it as a central theme. How about blending animatronics and vending?

Case 17: Medical Robots

A growing market. Look at the success of Intuitive Surgical and others.

Case 18: Exploration and Recovery robots

A fun concept, but not much money in it for the hobby or academic explorer market. But…

Case 19: Commercial Exploration and Recovery

Has been very lucrative for the few teams that have deployed robots to go after ship wrecks. But definitely not a commercial market.

Case 20: Audio / Video robot

Do you want your ipod and speakers rolling around after you? [Aside – perhaps it as a role in LifeLogging?]

Case 21: Asteroid mining robots

A scifi vision. But Google is on the way to the moon.

Conclusions and a couple of prescient comments from the man who co-founded the personal and military robotics industries:
• Don’t be seduced by the cool side of Robotics? Who gets paid more, the salesman or the janitor? Value tied to what they produce. (Sex robots).
• There are very interesting things happening in the manufacturing space, and iRobot will be talking about that in the future.
• Medical / Vending / UAV highest current revenue vs. margin opportunity.
• Is your goal to create a robot that does new things or does things differently?
• Incrementalism leads to very slow adoption.
• Is it an industry? Seems like a technology enabler for existing and new industries.

Trossen RoboDevelopment Speech Now Available Online

by RayRenteria

At this year’s RoboDevelopment conference, Matt Trossen of Trossen Robotics spoke about robot adoption and industry standards.

Robotics and computers are merging, there is no doubt about it, but questions remain- Why is it taking so long? Where do the standards belong? What tools are in the marketplace which assist integration? What are the necessary steps to take?

One thing is clear, we need a roadmap to help guide us through this winding road of merging worlds. Join us as we dig deep into this discussion of integrating modern computers into our robotics platforms.”

The speech can be heard online now:

Robodevelopment 2007 Talk by Matt Trossen, Creating a Roadmap for the Merging World of Robotics and Computers

Sensor Technology May Optimize Driving

by RayRenteria

The DARPA Urban Challenge created an unprecedented concentration of ambition, intelligence, technology, and money in an ecosystem that advanced autonomous mobility and related technologies by decades.

Holger Salow

One such advancement was with the ibeo Lux Laser Scanning system. ibeo is a subsidiary of SICK, veteran makers of basic-but-pervasive laser scanning technology. ibeo sponsored Team Lux which was led by Director of Technology, Holger Salow with whom Robot Central had an opportunity to chat.

What was most striking about Lux’s vehicle was the apparent absence of sensors of any kind. In the photo, Salow is pointing to a translucent shield that is hiding the robot’s primary sensor technology–the ibeo Lux. The team’s Volkswagen Passat has three of these sensors, one on each side of the front of the vehicle and one in the rear. The following graphic from ibeo illustrates the configuration. The only other sensing technology the robot has is a GPS and Intertial Management Unit (IMU).

Three Sensors

Salow expressed how deliberate the team was in making the vehicle consumer friendly. And it was. Except for an emergency stop button on the console, there was no indication within and about the vehicle that revealed the technology that lay within. “We use it to do our shopping and drive around,” Salow said of the car. Computer devices were neatly stowed in hidden compartments, wiring was well hidden, and the ibeo sensors were all but invisible. The robot appeared more than anything to showcase how the ibeo Lux technology could be physically integrated in a 3rd-party product.

Sarlow then began to explain the innards of the technology itself. He explained that the ibeo Lux sensor has four lasers configured vertically to cover a vertical “stripe” of 3.2° that is swept across a 100° angle identifying obstacles up to 200 meters away. With a 3.2° vertical scan, the ibeo Lux has effectively increased the thickness of the planar scan of its parent company’s (SICK) flagship scanning products.

The primary differentiator between the ibeo Lux and other mobile laser scanning systems is the software system architecture that builds on the primitive data returned by the sensor. The sensor merely provides a list of points within the sensor’s angular field of view. ibeo has built on this information to provide higher levels of abstraction to application developers.

The first software incarnation of the sensor is the data it returns which represents a set of points–each an indicator of the distance and angle from the sensor as was reflected by an object. Darker objects return a weaker signal while brighter or more reflective objects yield stronger signals. It is this principle that allows the software to identify stripes in the road. ibeo has an algorithm that evaluates four measurements per sample in order to “see through” rain, fog, or dust.

 

SciVestor Ibeo Architecture

The second layer is an object recognition layer. This layer clusters points returned by the sensor and identifies types of objects such as other vehicles or pedestrians. Beyond that is an object tracking layer that monitors up to 64 objects and their location, direction, and speed.

The highest layer is the application layer which leverages the foundation layers to implement high-level behaviors such as Adaptive Cruise Control (ACC), Pedestrian Avoidance, Stop and Go, and other driver-assisting applications.

The ibeo Lux is currently in prototype form and is expected to start production in October 2008. ibeo expects to produce approximately 3,000 units a year at a little over € 9,000 EUR or about $ 13,132 USD. Once the product hits mass production the unit cost is expected to drop to € 180 EUR according to Marketing Director Tanja Veeser.

The predecessor to the ibeo Lux is called the Alasca and was widely used in the DARPA Urban Challenge; however, in a poll of the top four teams (MIT would have been fourth according to Dr. Tony Tether) conducted by Robot Central, none used the higher-level functions made available by ibeo and none except one ranked the sensor in their list of top three most important sensors.

This is likely due to the fact that the implementation of the higher-level software intelligence and data analysis was precisely the point of the Urban Challenge. Each team believed they had the best way of doing it. When the software is removed from the ibeo equation, there is little more than a 3.2° vertical scan across a 100° sweep. ibeo’s biggest competitor, Velodyne, did not invest in a software stack and instead provided a substantially more robust and complete visualization of the world around the robot.

The strongest proponent of the ibeo technology was third-place winner Victor Tango of Virginia Tech. Team Leader Dr. Charles Reinholtz explained that all the systems were critical to the operation of their robot and he was very satisfied with its performance. When pressed to rank the priority of his sensors, he ranked the ibeo as #2 behind GPS.

As with all cutting edge technologies ibeo had its share of quirkiness. Dr. Reinholtz explained that they programmed the robot to stop and reboot the system whenever it started acting up. They experienced one such event during the final race where the robot sat for 90 seconds while it rebooted its ibeo sensors.

The DARPA Urban Challenge it seems wasn’t the best place to showcase the ibeo technology because the value of the ibeo software was undermined by contestants’ desire to implement their own approaches to solving the same problems; however, in the marketplace, ibeo has designed a formidable player in the Digital Horsepower economy. It is easy to envision this technology being embedded in every vehicle that rolls off the assembly line. Things like Adapative Cruise Control and brace-yourself Collision Avoidance systems will be made cheap and commonly available.

 

Central Texas Robot Makes Waves

by RayRenteria

H2rObot On Water

One of the best applications for a robot is one requiring repetitive and consistent behaviors. The mapping of subsurface characteristics of lakes is just such an endeavor.

“These studies are extremely tedious. Detailed mapping requires piloting the boat very slowly back and forth across the same water body. It would be like mowing several thousand acres of lawn,” says Jim Patek, inventor of H2rObot. To make matters worse, OSHA requires at least two people to be in the boat, which causes the boat to be heavier–reducing the amount of water area it can analyze due to the substantial shallow area it has to map. “When trying to map several fringe marsh areas on Matagorda Bay, we were able to navigate into less than 20% of the study area due to depth restrictions of the boat.”

When Environmental Engineer Jim Patek of Cedar Park, TX thought about these problems, he decided to build H2rObot–an autonomous boat ‘bot that can deploy various mapping packages. The robot is a catamaran that carries what looks to be a pair of watertight ice chests. The pontoons are each about the size of a canoe. Both have electric trolling motors controlled by a central computing system contained within the watertight compartments. System central processors run a flavor of Linux called “SLAX,” and several PIC controllers that interface with the actuators.

For navigation, the robot uses GPS, a compass, three accelerometers and three gyros. Patek also devised a system using an advanced predictive algorithm called a “Kalman Filter,” allowing the robot to navigate from waypoint to waypoint in spite of hardware performance.

Jim Patek

The robot stays in contact with a land-side laptop while running in autonomous mode. If it loses contact, it will first try to back-track to reacquire a signal. Failing that, it shuts down, begins blinking a flashing light, and yelps a 900mHz chirp while Patek goes looking for it.

Patek has been educated in water sciences, a self-described “aquatic biologist turned engineer.” He was working on his PhD in Aquatic and Mathematical Ecology when he decided to switch schools and become an Environmental Engineer. However, “Anyone with much field work will know that being a good scientist or engineer is great, but for field work, being a good mechanic, electrician, carpenter, machinist, is better. This worked out well for me since I started working in an automotive shop with mechanics to whom I was related just as soon as I was tall enough to see over the fender.”

Patek built many robots over the last decade and a half but he described H2rObot as his most ambitious.

Benthic Mapping RobotComponent Case

Oshkosh Steadfast on its Path to Military Smartcars

by RayRenteria

Chris YatesIn what seems to be a recurring theme among DARPA Urban Challengers, these guys have greater ambitions than winning the race. “The goal was to be a strong performer, not necessarily win,” Chris Yakes, Director of the Advanced Products Group for Oshkosh, said.

Yakes explained that his team is developing an autonomous system in kit form similar to other kits the company provides to its military customers. Approximately 30% of Oshkosh’s business is in the military sector. The price of the kit will be comparable to the cost of their armored vehicle kit.

The kit still needs a little bit of fine-tuning. “We’re a couple of years from having a final product but we’re ready now to start doing demonstrations. By demonstrations I mean offering a vehicle (retrofitted with the kit) to a customer and letting them see first hand how it performs.” Yates went on to explain that variations of the autonomous system might be ready sooner. A semi-autonomous follower / leader system, for example, might be applied in a convoy of several vehicles in which only one contains a human driver or it is teleoperated.

Oshkosh has made some very overt technical design decisions based on the deep knowledge it has of its customer base. Their systems must be able to perform in very harsh environments and remain highly dependable for decades. “Our trucks live about 20 years. The system we provide must last at least that long.” Yates explained that they bias their technology selection to those without moving parts. Their computers are commercial off-the-shelf (“COTS”) systems that are cheap, easily replaced, and very durable. Oshkosh also depends more on computer vision than any other team that performed in the competition because cameras can be very rugged as they have no moving parts.

Computer vision is the method of identifying characteristics of a scene from digital images. Algorithms vary substantially depending on the desired data. Streoscopic vision, for example, requires two side-by-side cameras each generating an image of the same scene each from a slightly different perspective. The software behind this configuration finds objects of one image in the other. If the software finds the object in the same position on both images, the element is deemed to be far away. If the software finds that the object’s position has “shifted,” the object can be characterized as being closer to the cameras. The greater the shift, the closer the object is. From such information a 3D representation of the scene is constructed. Oshkosh is working with the University of Parma in Italy to develop the software behind their computer vision system.

Biasing towards cameras does not necessarily make them the only sensor in the kit. Oshkosh has also chosen to use ibeo LIDAR scanning technology to supplement their computer vision system. Yates was very reserved when asked details about other systems they evaluated and why they chose ibeo; however, he did explain the role of the ibeo. There are three units on the vehicle–two forward units and one rear. The units are used to identify obstacles.

During a conversation with 3rd-Place winner Virginia Tech Team Leader Charles Reinholtz about their own implementation of the ibeo sensors, Robot Central learned that there were occasions that the ibeo would be blinded and the system would require a reboot. There was at least one time when their robot, Odin, paused itself for approximately 90 seconds while it rebooted its ibeo systems. Team Oshkosh was eliminated when their robot threatened to bring down a building after it ran off the course and met a load-bearing pillar. As of the writing of this article, it was unclear if any contact was made nor was it clear if the failure was due to the ibeo issue described by Virginia Tech.

Nevertheless, Oshkosh is optimistic about its future role in saving the lives of American soldiers. And it should be. The sensor vendors will address their issues and Oshkosh will continue to mature its product. Oshkosh’s performance in the 2005 Grand Challenge and in this year’s 2007 Urban Challenge has very effectively demonstrated its technical and business prowess.

Terra Max

Velodyne Prez talks Mobile Sensing

by RayRenteria

I wrote about the Velodyne sensor a couple of weeks ago touting its commercial viability in the emerging mobile robotics market. I had a chance to talk to one of its inventors today. Velodyne President Bruce Hall made it clear that although his company doesn’t have a vehicular entrant, they are clearly and very visibly participating in the DARPA Urban Challenge.

“We went from being on 12 out of 35 teams at the beginning of the National Qualifying Event to being on 7 out of 11 at the finals. I hope that when vehicles 1, 2, and 3 come in, we’re on 3 out of 3 vehicles,” Hall said.

The Urban Challenge teams are comprised of some of the most brilliant engineers on the planet. Adoption of a technology that can give them a structured view of the world around their robot obviously seems too compelling to resist, considering the adoption rates of Velodyne’s LIDAR within the DARPA Urban Challenge ecosystem. The interpretation of the data is where the scientists shine. Mere mortal application developers, however, require a higher-level abstraction that interprets the data for them–recognizing a blob of points as a vehicle or a pedestrian, for example.

Every team that uses the Velodyne system has built what is likely to be a driver intended to solve the same problem. That seems to be the point. One of those implementations will be the best. “It’s a competition. These guys are trying to solve a common problem better than the rest,” Hall explained. That statement seems to be validated by several teams with whom I spoke. Teams that used a competing technology, ibeo, chose to employ their own object recognition algorithms instead of using the algorithms provided by ibeo, a technology I will cover in a subsequent post.

“Today, our customers require a certain level of sophistication that inherently limits our sales [to scientific programmers that can interpret the data]. We know that. Some customers are asking for (inerntial management) correction, object identification, and so on. Like any good business, we try to stay focused.”

Staying focused is not the same thing as ignoring customer feedback. “Customers wanted to be able to choose density over speed. We gave them the ability to spin the unit at 300 and at 900 RPM.” The slower-spinning mode provides higher density of data points while the higher speeds senses the world faster. Another feature due to customer feedback was the governing of the high data volume being broadcast by the device. The governor reduces the number of points sent by the device to something relatively consumable by most applications. The device is capable of informing the robot of over 2.3 million points every second but for the software that consumes it but it’s like drinking from a full-open fire hose. The device can be governed down to a slower 1 million points per second and the application can choose to ignore data it can’t consume.

In all emerging technologies there are always issues that come up that allow their makers to refine and to strengthen the technology. Velodyne is not immune to this. There is a little bit of tension in the air with regards to a bug that has manifested in the context of the Urban Challenge. Close-proximity RF signals can potentially confuse the sensor. This problem has a simple fix that was offered to all teams. All but one chose to decline the fix. Hendrik Dahlkamp of Stanford Univeristy reflected a classic product development discipline. “We declined on putting in the fix because it’s too risky to make a change this late in the game. Plus, the cost of the bug isn’t that great to us.” For another team, however, it’s life and death. Team Annie Way was almost eliminated when the Velodyne sensor was “blinded” by the radio of a DARPA official. Actually, the team was in fact eliminated until the Team Leader contested the decision and explained the situation. Team Annie Way is now a finalist and Velodyne is working very closely with them.

Visit Scivestor.com or contact me at rrenteria@scivestor.com to learn more about the Velodyne and the Mobile Sensing market.

DARPA Urban Challenge Team Develops Autonomy-In-A-Box Product

by RayRenteria

Team Gray may not have advanced to the finals but they’ve got their sights on bigger targets. With their latest autonomy in-a-box they may be positioned to become the Microsoft and IBM of autonomous vehicles.

Chief Engineer Paul Trepagnier explained that the Autonomous Vehicle System, or AVS for short, is a completely self-contained unit comprised of componentized navigational mobility that can be scripted.

The system effectively boils the entire autonomous navigation down into three major categories:

  • Localization
  • Obstacle Avoidance
  • Actuation

It takes input from a GPS sensor to determine where it is in the world and the desired route to traverse. “The core AVS platform takes the GPS information and drives the vehicle. We’ve driven the car up to eighty miles per hour with less than 5 cm of error,” Trepagnier said. “Eighty two miles per hour,” President and Director of Gray Matter Eric Gray corrected.

The system is designed to accept input from sensors in order to identify obstacles. In the event an obstacle is in the way of the vehicle, the AVS performs the appropriate behavior to avoid the obstacle if possible.

As a testament to the modularity and flexibility of the system, Trepagnier explained that when they acquired the new vehicle for the National Qualifying Event, retiring “the white one,” they were able to integrate all sensors and actuators in less than fifteen minutes.

The modularity sheds a little light into the elegant architecture in the design. First, the system is comprised of two major layers: The Hardware Layer and The Software Layer. Both are proprietary.

Features of The Software Layer include the extensibility of the sensor array. It takes only three days or less to integrate a new sensor. The integration is comprised of a reusable driver that provides a common output that is consumable by the obstacle detection algorithms. Considering the available-now and relatively inexpensive cost of this device it behooves all major sensor manufacturers to develop AVS-compliant sensor drivers for their technologies.

Also within The Software Layer is a high-priority fail-safe system. There are two threads constantly running that monitor the software and the sensors: a Safety Monitor and a Failure Monitor, respectively. If one of the threads identifies something wrong, they have the authority to bring the vehicle to a stop.

The Hardware Layer also has some important responsibilities. “We didn’t like that (the monitor threads) could potentially go down. So we designed hardware to monitor the Safety Monitor and the Failure Monitor. The Hardware Layer can stop the vehicle if it detects that something went wrong with the monitors.” The Hardware Layer is connected directly into the vehicle’s drive-by-wire system. The system has no moving parts and can withstand “way above 10g’s. Most of what we’ve seen in our applications have incurred fewer than 2g’s.”

The team has several prospects with whom demonstrations and negotiations are underway. The base price of the unit is $125,000. This is a tremendously cost-efficient solution to an otherwise expensive problem.

“Our vision is to be a commonplace technology in all vehicular testing environments,” declared Eric Gray. “Universities are talking about using this as a basis for their work. It allows them to focus their efforts on areas of specialization such as computer vision.”

Trepagnier seemed to enjoy his next hypothetical scenario almost a little too much. “We have a computer that can drive within five centimeters of a planned course. We can run the car in record mode, drive the track, and then put the path through a post-processor that will run a genetic algorithm to find the optimal route, and feed it into AVS. ” He said of racing an autonomous vehicle at high speeds. “We would be unbeatable.”

Ugobe CTO John Sosoka: Inside Pleo at Maker Faire 2007

by RayRenteria

Robot Central recently had the opportunity to have a candid one-on-one talk, arranged by Jonas Lamis, Executive Director of SciVestor and contributor to Robot Central, with John Sosoka, CTO of Ugobe, makers of Pleo at Maker Faire 2007. I fully admit that I was sure I would be talking to John about just another fancy toy. I was profoundly wrong.

Just as the doors were opening on the first day of Maker Faire 2007, John grabbed Pleo and we made our way out of the noisy floor to the quiet but windy outside and sat on the fresh-cut grass. Pleo seemed right at home standing in the grass as he appeared to be grazing. It was the first time I had seen Pleo up close. He felt sturdy and he had a mass that a dog his size might have.

Fascinated with Pleo and his convincing fusion of sensory competence and animatronic grace, I blurted out the first question that came to mind. “Do you think people will perceive Pleo as a robot toy or as a pet?” He responded indirectly. “For me, the thing that’s exciting about going down this path is the life-form side.” He went on to say that “you have to do a good job of building a robot in order to pull it off, but it’s almost like that’s the price of admission to get to do the part I’m really intrigued about which is to give that illusion of life.”

Given that the “price of admission,” or the building of the Pleo platform, has already been paid, I asked about Pleo’s openness to 3rd party developers. He said that “[Ugobe is] trying to provide access for the widest group of people we can to modify Pleo.” He explained that Ugobe wanted to make available a platform that would also appeal to creative application developers and not necessarily just to researchers. “I think it would be a lot of fun for a lot of kids to be able to do programming, and algorithms, and robotics experiments with something like this instead of a wheeled car, or an iRobot Create, which are great but not everybody likes those things.”

I dug in a little deeper into the programming aspects of Pleo. I asked John about the VM running in Pleo and the primary language used by Ugobe. When I asked if it was Java-based, he said “Java to me is punishment.” He immediately tried to soften the statement by explaining how he loved Gosling’s work but quipped again that programming with Java is like programming “in handcuffs.” “I love virtual machines, though, I’ve used them a lot. We didn’t write a proprietary language, as much fun as that would have been.” After briefly consulting with Ugobe’s publicist whether he could tell us the language Pleo’s written in, he told us it was written in Pawn.

While on the topic of programmability, I had to ask John about his opinion regarding standards on personal robots. He gave a good and seemingly well-rehearsed response. “One of the things that we find is that without standards we usually don’t find strong commercial interest because there’s too much overhead” in developing software for the robots. He deepened his analysis, however and acknowledged that standard interfaces to actuators and sensors would make life much easier for him, in particular during the prototyping phase; however, he believes that the large variations between robot applications and actuator configurations constitute a major inhibitor to the establishment of a standard. “I think it’s a good idea but it won’t solve all of my problems right now which is what I’d like.”

It was about that time that I looked down at the grass and noticed that Pleo was looking right at me. His tail was wagging and his eyes blinked. I was taken aback at how effective the combination of a wagging tail, friendly eyes, and a small growl can be when combined with sensors that respond to the world. Next to the animated Pleo was the carcass of a relative which depicted the inanimate collection of innards comprised of plastic, touch sensors, a camera, microphones, an infra red transmitter and receiver, motors, springs, joints, and wires all organized in a very deliberate and compact configuration. And those were the elements of Pleo I was able to see from the outside. Inside was an equally dense microprocessor and battery ecosystem that continued the artistry of Pleo’s design. This hardware was the price of admission to which John referred earlier in our conversation.

After playing with Pleo for the first time, I felt compelled to ask, “Why a dinosaur?” John lit up as if though he had been waiting for that question. He started with a good business answer. We chose a dinosaur “because it’s a great brand. There’s no (sic) royalties.” Once he got that out of the way, he got a little more interesting. “Everybody’s fascinated with dinosaurs.” He went on to explain how deliberate the decision was to make Pleo a dinosaur. Not just a dinosaur, but a particular kind of dinosaur–a Camarasaur. “A Camarasaur after a couple of weeks of hatching would be right about this size,” gesturing to Pleo. “We didn’t want to make a scaled-down dinosaur because it would feel like you were playing with a toy. As a Camarasaur, you feel as though you’re playing with a dinosaur.” John also explained that a dinosaur had certain elements that could be used in expressing or emoting, such as a tail or a long neck.

Before wrapping up the interview I had to fulfill my promise to my 8 year-old daughter, asking if Pleo could sing. John picked Pleo up and showed me the SIM card slot underneath and explained that all of Pleo’s sounds could be shadowed on the card if named properly. So, the purring sound Pleo makes when he’s content might be replaced with a singing sound–thus, making John’s answer, “Yes.”

John Sosoka’s responses made it clear that Ugobe was trying to do more than make a toy. Every decision Ugobe made about Pleo’s design aligned with the desire to create an artificial life form, from the decision to make Pleo a dinosaur to the behaviors chosen to be coded within him. Like the Roomba, I believe Pleo has the capacity to evoke emotional responses in humans. Unlike the Roomba, Pleo’s purpose is to target and appeal to our emotions. I believe Pleo will do just that.

We recorded the interview and broke it out into different topics:

Observations of Adoption Trends in Robotics

by RayRenteria

Selling products to businesses is a relatively quantifiable endeavor. A business customer calculates the cost of his investment and compares it to his projected increased revenue. If his revenue exceeds his cost then the purchase is deemed to have been worth the investment. The higher and sooner the return, the better the investment. Most of the work in selling to businesses is in determining and negotiating the formula for calculating the return on investment because the currency of a business deal is money.

Such is the case with selling manufacturing robots. Manufacturing robots provide automation. With automation, a business can produce more products more quickly and with consistent quality. Typically, the cost of the investment is overtaken by profit at some point in time.

So where does that leave personal robots? Personal robots represent a new class of robots that are front runners for becoming man’s new partner. What’s that worth? Our findings indicate that robots with no specific function still have a ways to go before consumers will pay for animatronic companionship.  Conversely, robots with capabilities that allow us to justify the investment will be quicker to get adopted. Capitalism, it seems, will be the primary driver for adoption of personal robots into society.

Evidence

iRobot stumbled upon human-robot attachment when its home service robot, Roomba, began evoking emotional responses in its owners. A great article by Joel Garreau depicted multiple instances of soldiers bonding with their Military robots in life and death contexts. In the following 2×2 matrix, we’ve plotted the relative position of various classes of robots based on their functional and emotional value.

Robot 2×2 : Emotion vs. Function

Robots with higher pure functional value tend to be consumed by businesses. Classes with less function and less emotional value typically fall into the hands of hobbyists who would take issue with the assertion that these robots evoke little emotional attachment.

Projected on the chart is the belief that anthrobots, or humanoid robots, will yield relatively high emotional value with limited functional value. David Levy, a brand new PhD., defended his dissertation on October 11th where he contended that humans and robots will be sharing intimate relationships in the near future. He projected the adoption curve of these robots in the following graph:

Levy Graph

Observations like the ones we’ve made about human responses to the child-like robot named CB2 concur with Dr. Levy’s prediction that humans may generally become profoundly disgusted with life-like anthrobots before they’re accepted into mainstream culture.

Military and Home Service robots provide both function and emotional value. Although toys are doing well financially, they don’t offer much function. Their relatively low cost is an investment parents are willing to make. My dad once told me that buying one of his bikes actually meant the purchase of countless smiles and games of cops and robbers for the children of his customers. A child’s happiness, he went on, was neatly packaged on two wheels for $79. That’s a deal.

Two Predictions

  • We believe that robots who have both a function and can tap into or facilitate human emotional situations will perform the best. Telepresence robots are first in line because their price point has gone down, mitigating the risk of investment we should realize good performance in Christmas of ’07.
  • Competitors in the Home Service robotics sector will emerge in the wake of iRobot’s floor and gutter cleaning robots and have a particularly strong showing Christmas of ’08.

This Time, It’s For Real

Humans have been saying that robot culture is right around the corner for decades. Our observations of trends, and falling prices indicate that this time, it’s really happening.

References:

DARPA Urban Challenge Yields its first Commercial-Grade Innovation

by RayRenteria

In 2004 the guys from Velodyne who make subwoofers entered a robot into the 2004 DARPA Grand Challenge under the name of Team DAD (Digital Auto Drive). Back then, their primary obstacle avoidance system was their proprietary stereoscopic vision system mounted on the roof of their Toyota’s cab.Team DAD

In 2005 they invented a new kind of a scanning technology. Team DAD’s scanning invention appeared exotic to say the least. It looked like a fast-spinning UFO hovering a few inches above the team’s Toyota pickup truck. SICK Ladar systems were the obstacle-detection sensor of choice at the time but they only scan a thin slice of the world. Competitors who used SICK had to build their own contraptions that would swivel the scanners in order to sweep a scan or they would install multiple scanners, each scanning a different plane.

DAD’s innovation solved that problem by rotating 64 lasers 600 times a minute. The placement of the lasers also yields a 20-degree vertical scan. Here’s the description they wrote up in their technical brief for DARPA:

A unique LADAR terrain mapping and obstacle detection system is employed as the single sensor. The 64-element LADAR system has a 360-degree field of view and a 20-degree vertical range. It is mounted on top center of the cab, giving it a clear view in all directions, and rotates at a rate of 600 RPM. The camera is shock mounted, and has an INS sensor system mounted on it to report exact pitch and roll of the unit that is used by navigational computers to correct for these forces. The unit generates its own light and uses a proprietary filter to reject sunlight, so it works well under all lighting conditions. Since the whole camera spins, dust and rain are spun off the unit as it rotates. The LADAR unit is capable of seeing through fog and heavy rain by ignoring early reflections. The unit has a dynamic power feature that allows it to increase the intensity of the laser emitters if a clear terrain reflection is not obtained.New LADAR

They have since commercialized the technology and changed its design. It now resembles a silver can spins at 1200 RPM and it has its own web site from which you can order the product at a meager $75,000.

DAD was the only competitor with this sensor in the 2005 DARPA Grand Challenge. There are no fewer than six teams with the technology competing in this year’s 2007 DARPA Urban Challenge. Big budget teams such as Stanford and Carnegie Mellon (Team Tartan) are included in that list:

  1. Team Annie WayVelodyne on Junior
  2. Austin Robot Technology
  3. Ben Franklin Racing Team
  4. Golem Group
  5. Stanford Racing Group
  6. Team Tartan

As technological innovations such as Velodyne’s begins to mature, the future of autonomous driving gets closer. Robot Central will be discussing two other technologies that are prime innovative contenders for commercialization and accelerating autonomous transportation in civilian life.

References:

Matt Trossen: Robotics Must Become a Software Problem for Mainstream Adoption

by RayRenteria

Humanoid Crossing

Robot Central had an opportunity to speak with Trossen Robotics CEO Matt Trossen about everything from his days as an IT Analyst to the maturing of the personal robotics industry. It was clear after just a few minutes that Matt Trossen wasn’t interested in opening an average on-line robot store. Trossen is a man on a mission.

“Making applications for robots needs to be a software problem. Most robot manufacturers require their programmers to understand too much detail down in the hardware implementation layers,” Trossen asserts. By requiring intimate knowledge of the hardware layers in robotics creates a major roadblock in the progression of personal robots into the mainstream. “Programming is the most common technical skill of the day but today you have to be an Electrical Engineer to program most personal robots.”

As an IT Analyst three years ago, Trossen found himself using and being impressed with Phidgets–a library of Active-X based abstraction software controls to PC-based hardware controllers. “You could just drag and drop a switch into a UI and turn things on and off. These guys were on to something.” Trossen was so impressed with Phidgets and their philosophy that he started selling them. “We built the world’s first hardware store for software programmers.” He eventually upped the inventory to include robots and changed the name to Trossen Robotics.

Trossen Robotics System

Trossen is acting on his philosophy. He’s introduced a proposal to the industry that aims to standardize personal robotics platforms by dileniating abstraction points while leveraging XML and a taxonomy that allows for a common definition for arbitrary inputs and outputs be they software or hardware. The Trossen Robotics System is the basis for an industry leading standard.

“Standards” such as OpenJAUS, Player / Stage, OROCOS and many others are actual implementations of operating software that make an attempt at getting developers to adopt them. (Read Microsoft and Tmsuk Create Robotic Alliance: Showtime? for more analysis on these platforms.) Trossen doesn’t fall into that trap and instead endorses the adoption of TRS while soliciting feedback from the Trossen Robotics community.

Robot Central did a survey of thirty four (34) robot vendors to identify other players in the industry that may be pushing a standard. We found a lot of great robot catalogs with varying degrees of customer focus. What we didn’t find is another proposal for an industry standard or a vision for standardization. Matt Trossen is paving the way.

You can listen to Matt Trossen in person at RoboDevelopment as he will be a speaker at the event.

Supplemental resources:

  • RoboRealm for some robot shop destinations.
  • We visited the following sites to search for industry vision:

… → Read More

Texas Company’s Technology Makes Robots Teachable

by RayRenteria

Skilligent TrainerRobot Central had the opportunity to speak with CTO Sergey Popov of Skilligent about their robot learning technology. We were intrigued with the prospect of a robot that could be taught skills and behaviors without any conventional programming. I’ve asserted that Human Robot Interaction is about to reach a tipping point that will make personal and service robots available to the masses and the technology Skilligent has developed is yet another validation of that assertion.

“We believe that trainable robots utilizing Skilligent will be much more flexible than today’s robots which are either remotely controlled or pre-programmed to perform a few functions,” Popov said. “You could train a robot to do simple palletizing in your garden to assembling furniture in a workshop.”

This range is enabled by the fact that it does not require a specific control system, calling language, and or hardware platform to operate.

Although Skilligent does not yet officially support Microsoft Robotics Studio (“MSRS”), they provide a C# version of their library which is easily consumable by MSRS. Additionally, Skilligent has a professional services group that will work with its customers to enable its technology with MSRS if necessary.

The software does have some explicit dependencies, however. “Our technology requires methods of observing the world and getting positive and negative feedback by the trainer–just as a human would.” As such, a camera is required to observe the world. The only actuators required are those required to perform the desired behavior. If you wanted to teach a robot to open a door it would need a gripper, for example.

In a sequence of three videos created by Skilligent, a robot is taught a basic set of behaviors. In all videos the trainer gains the robot’s attention by shaking an object in front of it. The robot will instinctively follow it until the trainer shifts attention to another object. In video #1, the trainer leads the robot to object #1, a random poster. He repeats this exercise in video #2. In video #3, the robot performs the trained behavior. To the untrained eye, this is a simple matter of record and playback. Upon closer scrutiny, however, we observe that the robot knows when to perform the task and has stored a symbolic representation of the target objects in its database which it can recall later. Had the target objects been moved or rearranged, the robot would have still performed the desired behavior.

Besides task-level behaviors, Skilligent can be trained to execute low-level control policies called “skills.” Task-level behaviors combine low-level skills in a hierarchical structure.

The software provides a skills database abstraction that can be shared with other robots running the software. Furthermore, skills can be used as building blocks to create more complex behaviors. For example, you could teach a robot to fill a watering can with water as the “Fill Watering Can” skill. You could later teach a robot to water the plants which can utilize the “Fill Watering Can” skill.

This opens a new and unknown space. Will the human trainers want to share their robot’s library of behaviors with other trainers? If so, then we’ll have the same problem of requiring homogeneity in actuators, sensors, and mobility in order that the behaviors could be performed consistently across robots. It may be so cheap, easy, and fun to train robots, however, that this space won’t have the same limitations that control systems do.

Skilligent may be on the verge of inventing the killer application for robotics.

Additional links:

Google puts $30M on the Table to First who Lands and Drives Robot on Moon

by RayRenteria

The world is abuzz with today’s news that Google will back the Lunar X Prize. The $30M prize money is broken out into a $20M Grand Prize, $5M Second Prize, and $5M in bonus prizes. The offer is good through 2012 after which the grand prize will drop to $15M through 2014. After that, the bet’s off unless Google wants to extend it.

To win the Grand Prize, a team needs to land a spacecraft on the moon and drive around for at least 500 meters, and send back “specific” video and images to earth. For second place, a team needs to land a robot, drive around on the moon, and send back some data. Bonus prize money will be given based on the performance of some bonus tasks such as finding water ice or man-made artifacts such as Apollo stuff.

Reference:

Microsoft and Tmsuk Create Robotic Alliance: Showtime?

by RayRenteria

Several days ago we learned that Microsoft is creating a technical alliance with Japan robot player Tmsuk (pronounced “tim suck”) to establish a standard robotics platform. It represents the first time a strategic relationship has been established between a major software platform maker and a robot manufacturer. At stake is the effort to bring robots into the mainstream and fulfill Bill Gates’s vision of a A Robot in Every Home.

It’s easy to map parallels between the evolution of the personal computer and the progress of robotics. In their early days, both were used mostly to impress friends with the engineering prowess required to make the machines do cool tricks. When Apple introduced its version of the personal computer it was a spreadsheet application that caused the explosion of mass adoption by consumers. That big bang has yet to occur in robotics primarily due to the lack of standard platform for application developers. The guys who designed the first spreadsheet application didn’t know squat about building computers. The robotics industry must reach the same panacea where robotics application developers don’t have to know squat about robots in order to build a killer app.

It isn’t for lack of trying, though. Several platforms have been made available over the last decade but none thus far has been established as the standard platform for robotics. On the contrary, the continued addition of a new robotics platforms has further compounded the problem.

  • 1996 Webots is developed by Microcomputing and Interface Lab to be spun out as Cyberbotics in 1998.
  • 2001 Version 1.0 of Player, Stage, and Player Tools is introduced.
  • 2002 Sony introduces its OPEN-R architecture with the popularity of the now-defunct Aibo robot dog.
  • 2002 Evolution Robotics is founded and later introduces its platform and now flagship product, ERSP.
  • 2002 The first version of OROCOS is released.
  • 2005 MIT introduces YARP [pdf] (“Yet Another Robot Platform”) which encapsulates OROCOS.
  • 2005 (?) OpenJAUS made available. Touts self as “Military-Ready.”
  • 2006 Gostai is spun out of Ensta’s Cognitive Robotics Lab in Paris. They introduce Universal Real-time Behavior Interface, or URBI, which started development as early as 2004.
  • 2006 iRobot makes AWARE 2.0 available to 3rd party developers.
  • 2006 Microsoft releases Microsoft Robotics Studio.
  • 2007 Skilligent goes GA with its module. Note: Skilligent says they’re an add-on and not a platform. According to their webiste, “Skilligent! is a software component that can be integrated with various robotic platforms.“
  • 2007 CLARAty “reusable robot software” is made available by NASA.

What makes Microsoft Robotics Studio stand out of this crowd is the platform abstraction experience the company has with its operating system and, most importantly, Microsoft’s desire to evangelize the technology and establish a standard in the industry.

Now that Microsoft has this opportunity with Tmsuk, it must be successful. This initiative has white-hot spotlight of attention on it. Tmsuk is a member of the Japan Robot Association which is comprised of 48 members with household names such as Fuji, Mitsubishi, and Yamaha. Success with Tmsuk surely will follow by a push to spread into peer companies within the association which it will need to do in order to gain critical mass.

By American standards, Tmsuk is a relatively small company. As of March of this year, Tmsuk had a little over 1045 million yen (~$9M) in capital with around 30 employees. It’s unclear to me how they make their money. They have some humanoid robots designed as receptionists and most recently have introduced robots that look like the exoskeleton Sigourney Weaver wore in Alien. The latter represents a more practical and likely marketable technology that the company says it’s having trouble selling. Still, the company obviously has staying power. It’s been around since 2000 and is showing no signs of slowing down.

So, what now?

Either Evolution Robots and iRobot drop their own software and adopt Microsoft’s platform (good luck with that) or the Japanese robots whose brand names are already familiar to us will be tomorrow’s application base for the next generation of American software developer. The Japanese listened to Deming because they got his message. They became masters of efficiency and quality in the automotive industry. Tmsuk’s willingness to drop it’s proprietary software from its robots and invest in Microsoft is a modern-day sign that they get the value of having a homogeneous platform. It’s possible that Tmsuk will become the Toyota of robotics and Microsoft will sell tons and tons of robotic platforms.

References:

Harvard Makes Progress on “Biohybrid Materials for Soft Robotics”

by RayRenteria

Muscle FilmA team of scientists in the Disease Biophysics Group at the Harvard School of Engineering and Applied Sciences has been working on interfacing biological material such as heart muscle tissue with man-made polymers. The team has figured out how to grow muscle tissue in a structured way so as to be able to begin applying it. Their method allows them to simply cut pieces of a thin film coated with microscopic stripes of living muscle tissue into whatever shape they want.

In a video published on their site yesterday, the team shows a successful micro-scale biomechanical muscle twitching semi-autonomously. They can control the kinematics by shaping the piece of the material in some deliberate way. The same video shows pieces of the material that were cut into triangular strips “swimming.” Another portion of the video shows a microscopic gripper opening and closing.

This technology certainly begets interesting conversation regarding the continued convergence of humans with technology.

Principal investigators on the Harvard Team are Kevin Kit Parker and George M. Whitesides. Co-Investigators are Adam W. Feinberg, Alex Feigel, Sergey S. Shevkoplyas and Sean Sheehy. The video was made by Adam W. Feinberg

Sources: Harvard School of Engineering and Applied Sciences (video), Additional material: NewScientistTech

Previous Posts
Proud Member of the Robohub Community

Enter your email address to subscribe to Robot Central and receive notifications of new posts by email.

Our Tweets