Miniature Quadruped Robot Is Blazingly Fast

This robot, which still doesn't have a name, is very compact (which measures just 6.5 x 5.5 x 1 centimeter), and according to its creators it is quite possibly "the fastest legged robot of its size." Whether or not this really is a legged robot (or a quadruped) is perhaps debatable: these are wheel-legs, more commonly known as whegs. They're wheels in that there's rotary motion going on, but they're also legs in that there are discrete points of contact with the ground. To some extend, whegs offer the best of both worlds: they can be directly driven with conventional motors and allow for high speed and efficiency, while simultaneously providing traction over rough terrain and obstacles. Plus, you can easily swap them out, and by making them out of springy materials, you can give your robot some compliance. 

What makes the robot wicked fast is the fact that it's got four independent drive motors, each one of which has a power to weight ratio that's absolutely bananas. Only 6 millimeters in size each, the motors output 1.5 watts of power at 40,000 RPM, driving the individual whegs through 16:1 planetary gearheads. They're not cheap (hundreds of dollars each), but they make for one crazy little robot. And of course, independently driven whegs make the robot smaller, lighter, simpler to steer, and generally more efficient overall.

The current generation of this robot isn't capable of taking advantage of all of the power that the motors offer: even at top speed, it's only using about 0.60 watt, less than half of what the motors can output, since increasing wheel speed causes the robot to bounce along the ground, decreasing its actual speed. But, there's a lot of potential for swapping in some new whegs up to 35 mm in length (about twice as long as those currently on the robot), "which might produce even faster running speeds and the ability to navigate very large obstacles or challenging terrain, with a robot that still fits in your hand."

"Efficiency and Effectiveness Analysis of a New Direct Drive Miniature Quadruped Robot," by Christopher Y. Brown, from the Johns Hopkins University, and Dana E. Vogtmann and Prof. Sarah Bergbreiter, from University of Maryland's Microrobotics Lab, was presented last month atIEEE International Conference on Robotics and Automation (ICRA) in Germany.

Update 11 June 2013: The post originally said the Johns Hopkins/University of Maryland quadruped was "much quicker" than the UC Berkeley legged robots. Not so fast, the Berkeley team told us in an email. One of their robots, VelociRoach can reach 2.7 m/s, which is faster than the JHU/UMD quadruped's top speed of 2.2 m/s. On the other hand, the JHU/UMD quadruped achieves 33 body lengths/s, which makes it quicker relative to its size than VelociRoach (27 body lengths/s). So, clearly, comparing robots of different sizes is tricky (in fact, if the criteria is top speed, the winner would be another Berkeley robot, STAR, which can run at 5.2 m/s.) Maybe it's time for a race at the next ICRA conference?

Read more »

Smart technologies : New types humanoid robots can feel

Humanoid Robots can feel now

With the development of smart technologies, new types of Humanoid Robots are having the ability to feel materials . Even this types of Humanoid robots can be more touch sensitive then humans . This types Of Humanoid Robots can feel with the help of sensors, software and actuators.  This smart technologies have the abilities to identify different materials by touching them. Already IBM researchers predicted that we will be able to touch real object faraway from us by our mobile phone or computer screen within five years(Smart future technology – virtual touch of real objects ).This invention is a step to that prediction.
Researchers of the University of Southern California’s Viterbi School of Engineering published a article in Frontiers in Neurorobotics related to this types of Humanoid robots .In this article they wrote that how  specially designed Humanoid Robots can surpass humans in identifying natural materials by their texture.

This sensitive types Humanoid robots have special  types of sensors in their hand. This are tactile sensors built to act as human fingertip. This tactile sensors are Biotechnology sensors with produce vibration when it explore a material like human fingertips. This types of sensors have a soft, flexible skin. This skin is over a liquid filling. There is even fingerprints on the surface of the skin. This fingerprints enhance the sensitivity of sensors to vibration. When the finger of Humanoid Robots slides over any material, the skin starts to vibrate. This vibration I detected by a hydrophone. This hydrophone is in the finger of  Humanoid robots , as like as the bone of the finger. Our human fingers use similar types of pattern to identify materials, but this types Humanoid robots are more sensitive.
But to identify any material, still required an algorithm. This types Humanoid robots use a newly designed algorithm to explore a material. This algorithm is based on human strategies. While exploring a material humans concentrate on it’s   traction, roughness, and fineness. Humans also take informative decisions based on the sequence of the  exploratory movements. This exploratory movements find out the most information for the task. They depends on the past knowledge about those exploratory movements. This way they guess what a material may be. According to a famous theorem by Thomas Bayes ,decisions might be made from the information obtained during these exploratory movements. This exploratory movements done by human based on their prior experiences.  But until now there was no way to decide robots have to made which exploratory movement to identify a material.  Gerald Loeb  Professor of Biomedical Engineering and recently Jeremy Fishel a graduated doctoral student, the writers of that article describes their new theorem for solving this general problem as “Bayesian Exploration.”

Their specialized humanoid robot was trained to identify 117 different types of material such as fabrics, stationary, and hardware stones. The Humanoid robot has been able to  correctly identify the material 95% of times , when it was given materials at random.  It was only rarely confused by pairs of similar textures what human could not distinguish at all by their exploratory movements.
This types of  humanoid robots can be used in companies to identify different materials.
Content source:  http://www.sciencedaily.com/releases/2012/06/120618194952.htm

Read more »

Robot Lull?

It's almost a month since IREX 2011 wrapped up. Going back over my notes, photos, and videos from what is usually the high point of the robotics calendar, I have to say that the show this year was a bit lackluster.  There were a few new robots and applications that surfaced, but not near as many as we've seen in previous years.
Some companies were noticeable by their absence. Either they have been absorbed into other robotics and automation companies, or they couldn't afford to participate. It could also be that they felt that the show isn't as relevant to their target customers as it has been in the past. Whatever their reasons, it isn't a good sign, and seems to indicate that the industry is in a lull, waiting for some new breakthrough or resurgence in investment capital.
The companies that did participate were, by and large, demonstrating slightly enhanced models of the same robots and technology that they've been promoting for several years, in some cases for almost a decade. I will grant that there has been some incremental improvement, but not enough to light a fire under the market and stimulate exciting new growth.
The high point for IREX appears to have been in 2007 when the energy that filled the exhibition center was palpable, almost like the ionized air right after a lightening storm. Two years later, IREX 2009, was struggling in the aftermath of the Lehman Shock, as it's known here in Japan. Attendance and exhibitors were down considerable, though the organizers made a strong attempt to keep up appearances by installing false walls blocking off exhibit booths that were unsold.
This year, IREX 2011, I was surprised to find quite a few small companies with limited financial resources, occupying multiple booths. In better years there is no way that companies of that size would have had the budget to show off their products in two or three distinct, and physically separate, booths at the show. Yet, there they were this year, doing a great job of marketing. I'm very happy for the companies that were able to do that. At the same time, it's a sign that booths weren't commanding the high prices they have in the past.
Hopefully, it's a temporary condition and one that will be just a memory by the time that IREX 2013 rolls around. There are some signs of life in the industry, and some promising applications emerging. Cloud robotics is going to make a major impact on the industry, manufacturing, and our overall quality of life, in the near future. It's still too early to predict when we'll start seeing measurable results, but my guess is that it will be within years, not decades.

Read more »

Latest Robotic Technology: The Possible Roles of Robots

In the current scenario, the actual deployment of force—launching a missile, firing at a target, deploying a mine—is done by human operators remotely controlling unmanned vehicle systems, whether terrestrial, aerial or naval. But as robotic technology advances and makes truly autonomous robots possible, will we or won’t we allow military robots the ability to acquire targets and take suitable offensive or defensive decisions on their own? Although artificial intelligence systems are still in a relatively nascent stage to make that a possibility in the near future, there are also some ethical and legal areas of concern that we must address in order to ascertain the practical and moral viability of developing  armed military robots for use in  warfare in an autonomous role. There are possibilities that include unintentional weapon discharge, incorrect target bias, and the possibility of the opposition compromising weapon systems controls.

Why Greater Autonomy for Military Robots is Important (and Why We Need to be Careful)

Pentagon representatives and researchers working on military projects have often asserted that using armed robots while maintaining a remote human controller works on one level: it limits loss of human life. However, it fails to reach another objective that military robots should achieve; that of cost cutting. Not only do you deploy a costly piece of hardware, but by having a human operate it, you must spend on paying this operator too. Several theories of operation, some on the lines of Rules of Engagement (RoE) that apply to human soldiers, have been suggested that could allow autonomous armed military robots to operate on the battlefield with other manned and unmanned systems. One of these concepts is to develop autonomous armed military robots that they can automatically identify, target, and neutralize or destroy the weapons used by adversaries, but not the people using the weapons.

This runs the risk of leading to a situation where a robot not targeting human opposition leading to damage to self or other military robots and/ or loss of life on the side of humans. In this way, it would appear that the laws of robotics is unbroken, as the robot eliminated human threats to self and human soldiers on the same side as the military robots mentioned. The twisted part, however, is that in order to avoid harming self or other humans through inaction, the military robots must deliberately break the first law to begin with, that robots would not harm human beings. Researchers in this regard propose a fall-back option that faced with such a situation, the robot may be taken over by human operators, thus changing it from intelligent robot status to a mere unmanned fighting vehicle, albeit temporarily. Another proposal is to arm military robots with lethal weaponry for destroying structures and opposing unmanned military robots, while also equipping them with non-lethal incapacitating weapons for use on human targets.

The machines could be designed with various quickly programmable levels of autonomy so that they can switch among operational modes in accordance with the situation at hand.

Read more »