Festo shows operating concepts for human machine interaction
Although Festo primarily represents a supplier of automation technology, our loyal readers have been introduced to them with various interesting biomimetic projects we covered through the years. On a recent Hanover Trade Fair 2012, Festo exhibited a couple of their new concepts, and in this article we’re going to write about human-machine interaction, music being reinterpreted by artificial intelligence and an upgraded version of a robotic gripper we wrote about earlier.
CogniGame is a reinterpretation of Pong – a well-known video game that was launched on the market in the 1970s, in which players used a joystick to move a paddle up and down on the screen to return the ball to their opponent. Festo made a twist on the game by bringing it into reality. For CogniGame, they created a real court built with Festo components. Two linear axes whose drives move to the left and right along the baselines move the racquets to return the ball and keep it in play.
One player controls his racquet via a brain-computer interface (BCI) which measures voltage fluctuations on the surface of the head. The number of electrodes and general looks of the device resembles Emotiv Epoc, and it operates on a principle similar to electroencephalography (EEG) via attached electrodes. On the opposite side of the table, the second player moves the racquet by pressing a lever using muscular power.
Festo worked with CogniWare to develop a software solution for controlling a racquet using thoughts and biosignals. It establishes a communication channel between the brain and hardware without any interaction from the user via voice or input devices.
Although the beginning of the video reminds me of a scene from the 1993 movie named Demolition Man, Festo’s work on CogniGame is related to other practical applications where people could interact with machines in a more effective way. It would enable faster and broader communication, but you would also need to be far more concentrated on your work to avoid misinterpretations.
Sound Machines 2.0 – artistic human-machine interaction
Eealized in cooperation with the artist Roland Olbeter, Sound Machines 2.0 is an intelligent, robot-controlled sound installation that records a melody, uses it to compose a new piece of music and plays it live on two violins, a viola, a cello and a double bass. Unlike normal instruments, each of these instruments contains only one string and they are suspended freely to guarantee a high level of sound quality.
The quintet is operated by electric drives and pneumatic cylinders controlled by the automation platform CPX. While the complexity of this system isn’t that impressive, the interesting part of the whole system the algorithm it uses to create music. The individual sound robots playing the instruments react to any melody played to them according to stored composition rules.
The sequence starts when a user defines a starting position, and the robots interact with the original idea and continue the process with continuous variations. The robots are networked together in a way that allows them to “listen” to each other and maintain the essence of the composition.
While it may not seem useful aside entertainment applications, the idea behind this project is to create networks of decentralized components which are able develop autonomous group behavior patterns and independently make decisions in order to ease the work of a human operator.
Bionic Handling Assistant 3.0 – now with voice and image recognition
We already wrote about Festo Bionic Handling Assistant, but as the title says, the version presented at the Hanover Trade Fair 2012 shows how its interaction with the surrounding and among humans could be improved. Unlike conventional factory robots, Bionic Handling Assistant does not pose any danger in the event of a collision with a person.
Since that ability allows the Bionic Handling Assistant to assist in various areas where tools are used, Festo developed the logical next step of the system with voice control and image recognition. The assistance system can now grip objects independently and without the need for any programming or manual operation.
A small camera located in the gripper module registers the working space and targets objects. Voice detection is used to provide commands to the system which is able to reach the desired object and moves it according to the defined collection of commands.
As you can see in the video above, current system isn’t advanced enough to recognize objects without markers, but we can expect better results with the advance of object recognition algorithms. There are also many other hurdles to solve before we witness the use if this kind of technology in near future robotic assistants.
Don’t get me wrong, I like the idea behind this project, and there are many teams around the world which are trying to solve the problems I’m about to mention, but I doubt the folks from Festo have plans to commercialize this technology for use among humans anytime soon.
Some of the problems I can think of are related to portability of the system since it relies on pneumatics, the speed of its operation and how much faster it can get while still operating safely around humans, sound recognition in more noisy areas, as well as distinguishing a single person’s voice that would reduce the number of wrong command interpretations.