Archive for the ‘Podcast’ Category

January 28th, 2011

Robots: Odor Source Localization

In this episode we revisit robot olfaction and take a closer look at the problem of odor source localization. Our first guest, Hiroshi Ishida from the Tokyo University of Agriculture and Technology is an expert in the field, whose sniffing robots range from blimps to ground and underwater robots. Our second guest, Thomas Lochmatter from EPFL talks about tradeoffs between biologically inspired and probabilistic approaches to navigate a gas plume.

Hiroshi Ishida

Hiroshi Ishida is Associate Professor in the Department of Mechanical Systems Engineering, Tokyo University of Agriculture and Technology, Japan.
The focus of his research group is to develop robots that can find sources of airborne gas plumes or underwater chemical plumes. To this end, they developed the Active Stereo Nose (see figure below), a differential gas sampling system inspired by the dog’s nose, and the Crayfish robot that mimics the mechanism used by crayfish in nature to create unidirectional water currents.

Thomas Lochmatter

image credit: SNF


During his PhD at the Distributed Intelligent Systems and Algorithms Lab at EPFL in Switzerland, Thomas Lochmatter developed a modular odor system for the Khepera III robot. His research focused on the pros and cons of biologically-inspired and probabilistic algorithms for odor localization, while dealing with both single and multi-robot systems.

Links:

| More

Related episodes:

December 31st, 2010

Robots: Multi-Robot Teams

Happy New Year from the whole Robots Podcast team! Don’t miss our amazing holiday robot videos!

For this last episode of 2010 we look at multi-robot teams and impressive quadrotor capabilities with Vijay Kumar from the GRASP Lab at the University of Pennsylvania and one of his PhD students, Daniel Mellinger.

Vijay Kumar

Vijay Kumar is Professor at the GRASP Lab and Associate Dean for Academic Affairs in the School of Engineering and Applied Science at UPenn.

As an expert in networked multi-agent systems, he’ll be telling us how he sees robot teams of tomorrow being deployed in real-world missions. Challenges include keeping the robots networked, selecting the right level of autonomy and figuring out how to deal with large swarms of heterogenous robots.

Among the robots he sees doing team work are flying robots including quadrotors that use SLAM to get around or that can perform impressive aggressive maneuvers.

Daniel Mellinger

Daniel Mellinger is a PhD student at the GRASP Lab. He’s a rising star in dynamic control and has been featured in the media and on YouTube for his work with quadrotors passing through hoops and performing amazing motions.

We also speak about his latest work on transporting large objects using cooperative teams of quadrotors. This work got him the best paper award during the recent International Symposium on Distributed Autonomous Robotic Systems at EPFL.

Links:


Latest News:
For a review of 2010 in robotics, have a look at the Robots Forum!

| More

Related episodes:

November 19th, 2010

Robots: Dependable Swarms

In today’s episode we take a close look at swarm robotics and its potential use in real-world applications with expert Alan Winfield, co-founder of the Bristol Robotics Lab in the UK.


Alan Winfield
Alan Winfield is professor at the University of Bristol where he conducts research at the Bristol Robotics Lab. As a pioneer in the field of Swarm Robotics, he’s been looking at how large numbers of simple robots with local information can self-organize to achieve seemingly complex tasks. In such systems, inspired from social insects such as ants and termites, the intelligence of the group emerges from the actions of the individual robots.

Beyond simply showing that swarm systems can work in the lab, Winfield has been considering ways to make swarming a reality in applications. The main challenge is that it is typically difficult to prove that swarm systems will work, which is something desirable for real-world tasks. Furthermore, increasing the number of robots in the system also increases the probability of robot failures that might be disruptive to the swarm’s purpose. To overcome these challenges, he created a new field of research called “Swarm Engineering”.

Winfield is also a follower of the Open Science movement, meaning that he often publishes data, source code and project updates as science progresses. In his effort for transparency, Winfield is also Senior Media Fellow for the EPSRC which has gotten him blogging and tweeting about robotics.

Finally, we also discuss ethics with a focus on robots that make us believe they have feelings and the emphasis on ethical roboticists rather than ethical robots.


Links:


Latest News:
For videos of the robotic sailboat and sponge-wielding robot arm, have a look at the Robots Forum.

| More

Related episodes:

August 13th, 2010

Robots: Distributed Flight Array

In this episode, we discover an aerial modular robot called the Distributed Flight Array. To talk about this, we have Raymond Oung from the Swiss Federal Institute of Technology in Zürich.

Then, to celebrate aerial robotics, we’re holding a contest on flying robot noises for a chance to win a WowWee Bladestar.

Raymond Oung

Raymond Oung is lead researcher of the Distributed Flight Array project at the Swiss Federal Institute of Technology in Zürich under the supervision of Prof. Raffaello D’Andrea (see previous ROBOTS interview).

The idea behind this project is to design a set of vehicles equipped with a single propeller and wheels that can drive around in search for fellow modules with whom to dock. Single modules are not stable but once assembled, the flight array is able to take-off and achieve coordinated flight. Modules then detach in-air, fall to the floor and repeat their search for other propellers.

The main challenge in this system is to come up with a distributed controller that can allow modules to work together to achieve coordinated flight. Because of its endless number of configurations, the distributed flight array is the perfect research and pedagogical testbed to study control theory for complex systems.

Contest

We were trying to imagine the sound of all of these propellers and then realized it would be fun to record the sound of some of the flying objects here at EPFL. If you manage to match the sound with the correct robot picture, we’ll be sending you a Wowwee Bladestar. If multiple correct answers are received, the winner will be selected randomly. The contest ends on the 27th of August and answers can be sent via email to info@robotspodcast.com or can be posted below this episode in the comments section.

WowWee Bladestar

Audio:

Noises of Flying Robots

Images:

1: Eyebot

2: Airburr

3: SMAV

4: Blimp

5: WowWee DragonFly

6: Eyebot



The correct answer was:
1 -> F
2 -> B
3 -> A
4 -> C
5 -> E
6 -> D

Links:


Latest News:
For more information and videos of Ishiguro’s Telenoid R1 and the F1 Robocoaster in action, have a look at our forum.

| More

Related episodes:

February 26th, 2010

Robots: Swarming Satellites

In today’s episode we speak with the lead scientist of the SPHERES project, Dr. Alvar Saenz-Otero from MIT, which aims at developing autonomous formation flight and docking control algorithms for nano-satellites. We then dissect a well known definition of a robot dating back to 1979.

Alvar Saenz-Otero

Alvar Saenz-Otero is lead scientist of the SPHERES project at the MIT Space Systems Laboratory in the US. SPHERES (Synchronized Position Hold Engage and Reorient Experimental Satellites) fulfill all the normal requirements of satellites in a small basketball-sized shape. This small size is what has allowed these robots to be tested in the lab, during parabolic flights and even on board the International Space Station (ISS).



The research question is how to make these satellites work together by flying in formations and physically connecting, or docking. Such swarms of satellites could be used to create giant telescope mirrors in space with nano-meter precision and assemble future space stations without the need for human spacewalks.

Saenz-Otero also describes more generally how you get your robot into the ISS and his plans to motivate students about science or pursue his dream of large swarms in space.

What is a Robot?

This week we look at a traditional definition of a robot, coming straight from the Robot Institute of America.  According to their 1979 definition, a robot is:

“A reprogrammable, multifunctional manipulator designed to move material, parts, tools, or specialized devices through various programmed motions for the performance of a variety of tasks”

What’s interesting about this definition is how far we’ve come in the last 30 years in the development of robotics. In 1979 a robot was simply a manipulator used to move parts in pre-programmed motions, which brings to mind industrial robots used in factories. 30 years later robots are no longer simply manipulators, but can propel themselves in their environment, understand their surroundings and act accord to their particular situation and analysis of surroundings instead of simply enacting pre-programmed motions. This 30-year-old official definition no longer applies, so let’s try to figure out what robots mean to us today! Keep sending us your answers by email at info@robotspodcast.com and let’s get closer to an all-encompassing definition of a robot for the 21st century.

Links:


Latest News:

As always, more information on this episode’s news including the Flyfire project, the artificial foot and Korea’s plans for R-learning can be found on the Robots Forum.

View and post comments on this episode in the forum

| More

Related episodes: