metimes you stumble on the most interesting projects, and you get all inspired to start making stuff yourself. The other day, I came across a really cool DIY attempt at building a voice activated robotic arm.
Okay, to be fair, the robotic arm is already built, but the good news is that it only costs around US$55. You can buy it directly from OWI if you’re based in the US, or you can buy it from Maplin if you’re in the UK.
The rest of the project is built on freely available open-source software, and doesn’t look particularly difficult to set up.
The project that I discovered was developed by UK Aerospace Engineer, Arthur Amarra, who normally works on the structural analysis of composite aircraft wings, but who professes to have been an avid linux geek for as long as he can remember.
Amarra initially purchased the robotic arm as a gadget to play with and admits that the machine is not particularly useful in itself since it is only capable of lifting objects that weigh in at about 100 grams.
Typically, the arm is controlled by issuing a series of commands to it via its USB interface. While the device comes with its own software, some months back, a hardware hacker known as notbrainsurgery reverse engineered the USB protocol that the robotic arm uses.
This opened up a whole world to other hackers who were interested in creating their own uses for the device. When Amarra broke his wrist, he realised that it would be more convenient if he could control the robotic arm by training it to respond to voice commands.
He jokes that he wished that it could help him wash the dishes, but the arm is not really that capable, even with voice-recognition in place.
Amarra tackled the project by using of a variety of external open-source tools to build up the voice recognition capabilities, including the Julius speech recognition engine and the Hidden Markov Model Toolkit, and then went ahead coding the glue to tie all of the pieces together in Python.
As a strong advocate of open-source technology, Amarra has published all of the python source code and step-by-step instructions on how he managed to complete the project.
He makes a point of stating that it is really important that manufacturers of electronics devices keep their protocols and specifications open to the public. After all, hardware hackers ultimately determine the real capabilities of end products.
He gives the example of Johnny Chung Lee, who first became well-known for hacking WiiMotes for head-tracking and who ultimately ended up helping develop Kinect, which has benefited the electronics industry as a whole. In the same breath, he praises Microsoft for keeping the protocol open for the Kinect, saying “if the hardware remains open, then hackers can achieve more cool things as well – and the electronics industry can certainly use hackers’ creativity!”
Amarra is right. Since “notbrainsurgery” posted his reverse-engineered protocol online, other hackers have come up with some other interesting projects, including a brilliant example of how the robotic arm can be combined with a PS3 Eye to implement complete face-tracking device. Not ready to drop the gauntlet, Amarra already has plans to implement PS3 sixaxis controller (+ accelerometers) support for the robot arm soon, which would allow the arm to imitate movements performed through a controller.
Still, Amarra is quite willing to admit that this cheaply bought robotic arm has its limitations. He would love to start making changes to the hardware itself, including powering the unit from a wall-socket and ideally being able to incorporate harmonic drives to improve strength and speed, but these goals are still a long way off.
Projects like these are admittedly very quirky. Amarra candidly says that he tried to use his voice control to get the robotic arm to help him clear a coffee table, but it fared poorly at the task. Nonetheless, they teach us a number of things.
Firstly, programming can be a lot of fun. It is also very obvious that by making it possible for hackers to access hardware easily, manufacturers of devices will not only have better selling potential but will also see potential usage of their hardware increase through innovation. Finally, these projects give life to the dreams of our future technologies.
I’m thinking of buying a robotic arm.