Human beings make tools. Animals make tools. A tool gives an advantage over adversity and guarantees survival. Will the same happen if a robot makes a tool?

Researchers at ETH Zurich are on a mission to not only have robots make tools, but they are teaching the bots how to use them. Their goal is to simplify the complexity of this type of robot, while at the same time giving it extraordinary capabilities like creating its own parts, fixing itself, and even creating other robots from printed parts and integrated motors.

So far, their robot consists of a mechanical arm with Hot Melt Adhesive (HMA) capabilities. This means that at the end of the mechanical arm is just a hot glue gun that can be used to create parts layer by layer. In their effort to avoid complexity, it has no manipulator mechanisms, instead it only uses the stickiness of the printed parts to move them and place them.

The project has only just been started. In a demonstration of their progress, ETH researchers showed that their robot is capable of creating tools and thus performing tasks that it would not be able to do otherwise. It took the robot about an hour, but it was able to create a cup with a handle to transport water, an impressive task for a robot with no hands.

Next, the team or researchers are working to give the robot perception capabilities to eventually achieve a completely autonomous robot. At the moment, the robot can only create simple shapes in about an hour’s time; a long way away from creating an entire mechanism, but even bots need to take it one HMA layer at a time.

Eavesdropper

Wilfried Stoll and an engineering team from Festo in Germany have designed a robot that can compose and perform its own music after ‘listening’ to a melody. The robotic system, called Sound Machines 2.0, simulates two violins, a cello, a double bass and a viola with each using only one string. The string is loosened to change notes by an electric DGE drive unit that runs parallel to the string with a pneumatic cylinder that moves a wooden hammer which strikes the string to play the note.

To make music the robots first listens to a musician playing a melody over a MIDI capable synthesizer or xylophone that’s connected to a computer where modular synthesizer software processes the signal and sends it to the robots in real-time. The computer actually composes the music in its interpretation with the help of open-sourced software that’s programmed with special algorithms that are derived from John Conway’s Game of Life cellular automaton. Once the computer writes the music it is then pieced out to the robots accordingly. The music is heard through the robots own amplifiers and 40 watt speakers which makes the music sound more spatial with a better sense of depth, as opposed to using one speaker for all 5 robots. While the robots don’t exactly play like the string section of the New York Philharmonic, they do play pretty well even if it sounds like the music from TV show Buck Rogers.

The band (via Festo)

Cabe

http://twitter.com/Cabe_e14

Building projects with microcontrollers (especially Arduinos as of late) opens minds to an unlimited amount of innovations. Whether the projects are created for fun, or to serve specific purposes, they can always motivate others to be creative. The latest Arduino powered creation comes from Ekaggrat, an inspired individual with a strong interest in science.  

Ekaggrat has created a robotic arm that writes down the time by the minute on a dry-erase board. After a minute is up it erases the digit and rewrites the new time. It is driven by four servos that control the arm and hand movements. Two 9G metal gear RC servos move the arm while two 4.5G ultralight servos control the hand movements. Using servos makes it a little loud and noisy and limits the accuracy of the robotic hand writing. However, there still can be improvements to be made such as a change to stepper motors to improve accuracy.

Nonetheless, the creation is truly inspiring and the programming that it takes to complete a project such as this is difficult. Keeping track of the timing, 4 servos, and 7-segment display writing orientation in sync with one another is impressive for a hobby project. This goes to show that the amount of creations possible with technology is limitless.

Eavesdropper

A group of researchers from Cornell Creative Machine Labs are currently developing an autonomous robot inspired from metabolisms of biological organisms. The goal is to be able to break down larger complex structures into smaller simpler pieces that may be reassembled to create a different type of structure.

The robot clamps onto a specially designed truss using rotational robot-lockable connectors and can traverse the three dimensional structure using three basic motions. Bi-directional gears on the robot allow it move between perpendicular planes, horizontally and vertically along a truss, and 180 degrees around a truss to move from the top to the bottom. In addition, reflectivity sensors are implemented into the robot which give it a sense of location relative to the truss structure, ultimately allowing it to function autonomously.

The sensors and movements are all controlled from on board the robot. The robot carries with it a sensor I/O board, a microcontroller, servo motors, and an on board battery. These simple electronics allow it detach truss beams with 100 percent success rate and reattach them with a 70 percent success rate.

While they would like to use the robot for building deconstruction and repair, that does not stop the researchers from thinking big. One day they hope that it could help repair and construct the space station or any other dangerous tasks associated with the outside of the space station or a shuttle. Meanwhile, the researchers will be working on improving the robot so it will be cable of carrying multiple truss beams at once, and they plan on re-working the algorithms that will allow it to traverse a path of most efficiency based on the final design. 

Eavesdropper

Nano-quads (via KMEL Robotics)

Kmel Robotics may have one of the best promotions for their future flagship product, the mini quad-copter, via a collaboration with the University of Pennsylvania's GRASP Lab. The demonstration shows a series of synchronized flying patterns. The video gives the illusion of autonomous swam behavior. At the moment, both Kmel and GRASP are not letting out any information on the project.

The University of Pennsylvania researchers Alex Kushleyev, Daniel Mellinger, and Vijay Kumar put these quad-copters to the test. As the video shows, it was a very successful project. Like small remote control helicopters, we all received one Christmas in the past, these quad-copters will have a very limited battery power source. This is the only information released at the moment. As more develops, I will report back immediately.

Where can all this lead? To a better understanding of autonomous swarm control, rescue purposes, or continue to work together and impress like the Swarmanoids. Either way, these are sure to make an impact.

Cabe

http://twitter.com/Cabe_e14

University of Albany's robot in action

It was proven in 1997 that a computer’s computational abilities could outsmart a human when super-computer Deep Blue beat professional chess player Gary Kasparov at his own game. This match put the human brain against the software programming capabilities of a team of developers, but Deep Blue still needed human assistance in physically moving the chess pieces.

14 years later, roboticists hope to grant computers like Deep Blue the ability of playing an entire game of chess independent of human assistance. In 2010, a competition was held at the annual Association for the Advancement of Artificial Intelligence Conference in San Francisco. Roboticists from many places and universities presented robots that resembled those on an automotive assembly line.

Their robots ran into trouble trying to identify and then properly move the game pieces in accordance to the game rules. Some robots used cameras to locate pieces,  but none were programmed to identify them. Instead, memory of the initial position of each piece indicated which piece it was and how it could legally be moved around the chessboard. Despite all methods of movement, all had a tough time clearly identifying what moves were made and where exactly the pieces were placed. Furthermore, they were slow in making their moves, which only took them milliseconds to process and decide.

The winning robot was "Maxwell" from the University of Albany. The robot moved along its side and probably made generous use of its mobility to clearly see moves and piece placement from different angles.

While a robot’s arm-camera coordination still pales in comparison to a human’s hand-eye coordination, it is pertinent to note we developed those skills over hundreds of thousands of years. Roboticists and programmers have only been working on developing these skills in robots for only a few decades.

Cabe

http://twitter.com/Cabe_e14

TI has hit a homerun with the $4.30 TI Launchpad.  This low-cost, professional, easy to use tool gives the beginner everything he/she

needs to get up and running in the world of microcontrollers and robotics.

EMGRobotics MSP430G2553 Robot Controller Board

$15.00 Without CPU (Board is stuffed with socket only)

$17.00 shipped with socketed MSP430G2553

Features:

• Board ships fully assembled
• Board ships with MSP430G2553 in socket ( 16K flash, 512 Bytes RAM, A/D converter, ... ) $17.00 version
• Runs on 3 volts ( 2 AAA batteries )
• FAN8200 dual DC motor drive in socket - Control up to two 3 volt DC motors
• Control up to 4 RC servos
• Built-in IR Range sensor to detect objects up to 12 inches away
• 3 mounting holes for easy mounting
• Separate RC servo power supply connection
• Programmable using the low cost TI Launchpad

MSP430G2553 Pin      Function

P1.0                                User3      Digital/Analog

P1.1                                IR analog in

P1.2                                M2CE

P1.3                                M1CE

P1.4                                M1IN

P1.5                                Servo0

P1.6                                User8        Digital/Analog

P1.7                                User7        Digital/Analog

P2.0                                Servo2

P2.1                                Servo3

P2.2                                M2IN

P2.3                                User12      Digital

P2.4                                Servo4

P2.5                                User11      Digital

P2.6                                IR LED

P2.7                                User4         Digital

The EMGRobotics MSP430G MSP430G2553 Robot Controller Board (RCB) is based on the powerful TI MSP430G2553 16bit microcontroller.  The MSP430G2553 has 16K of flash and 512 bytes of RAM, more than enough to build some very interesting robots.  The MSP430G 10bit A/D converter and analog comparator really set this TI microcontroller apart from the others.  These analog capabilities make the MSP430G the perfect heart of a high tech robot controller.  The robot  controller works with voltages from 2.7 to 3.3 volts, perfect for 2 AA or 2 AAA operation.  The servo's are powered from a separate 6 volt power supply.  If you are not using servos, you do not need the 6 volt power supply.

The EMGRobotics MSP430G2553 Robot Controller Board (RCB) can individually and bidirectionally control two 400ma DC motors.  3 volt DC motors are perfect for small desktop robots, and this robot controller can control two of them simultaneously.  Each motor can be run forward, backward, or stopped independently.  This makes it easy to build any desktop two wheeled robot.

A robot needs a sensor to interact with its environment.  The EMGRobotics MSP430G2553 Robot Controller Board (RCB) ships with a built-in Infra Red (IR) range sensor.  The IR LED sends out pulses of light that is reflected off objects and detected by the IR transistor.  Using the high performance 10 bit A/D in the MSP430G2553 the range of the object can be estimated.  The IR range sensor provides a range measurement from 50 to 1000 depending on a object distance of 1 inch to approximately 12 inches.

The EMGRobotics MSP430G2553RCB is designed to be used with the   TI Launchpad debugger/programmer.  The TI Launchpad is a $4.30 ( that is not a typo ) full featured professional debugger and programmer.  The TI Launchpad can be ordered directly from TI here:

http://e2e.ti.com/group/msp430launchpad/w/default.aspx

The TI Launchpad is an amazing device for the price.  With the include C compiler and debugger ( Code Composer Studio ) you can be up and programming in minutes.  These are easy to use professional tools that let you program the chip, and debug your program using breakpoints and single stepping.  If you are new to embedded programming TI provides a series of tutorials to help you get started.

The EMGRobotics MSP430G2553RCB is supported by www.buildsmartrobots.com, where you can find projects based on the board, example code, and answers to questions. Our goal is to help you build smarter robots.

wiring the MSP430G2553SBC for DC motor only operation

wiring the MSP430G2553 for DC motor and Servo operation

The EMGRobotics MSP430GRCB Controlling a HEXBUG Spider

The EMGRobotics MSP430GRCB Controlling a HEXBUG Spider

The EMGRobotics MSP430GRCB Controlling a HEXBUG Spider

To Purchase a controller goto: Buy Here

For instruction on how to Hack a Spider goto: http://buildsmartrobots.ning.com/profiles/blogs/hacking-the-hexbug-tm-spider-with-the-emgrobotics-spider-hack-kit

The Nao Robot, part of the MyRobot connected site (Via Nao)

A social network for robots has gone live. MyRobots.com hopes to connect bots and smart web-enabled objects together in a type of "cloud robotics." The goal is to augment devices capabilities by off-loading computation or other task to the "cloud." This pushes robots beyond their physical and software limitations.

Users of the service can create apps to aid in adding to a bot's function. For example, a robot has a camera but does nothing with it but stream video. A cloud based app would give this bot image recognition, navigation, object tracking and the like to that image.

Once the owner's bot is connected to MyRobots.com, it can be monitored via the web, sent commands, return updates, and run custom applications. At the moment only a handful of robots are compatible with the social network. However, Arduino based devices and PCs can connect to the system.

A serial to Ethernet gateway allows  most non-connected bots to hop on the network. At $77 USD a piece, connecting various old "dumb-bots" may have to wait.

My question, can a virtual robot connect to MyRobots? If so, can improvements be made to the virtual-bot?

Eavesdropper