The 7 Things I Know About Cyber Security After 25 Years | April 2024
ArtsIT 2013_hoyer
1. Giving Robots a Voice:
MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
2. Giving Robots a Voice:
MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Ralf Hoyer Composer & Sound Artist, Berlin/Germany
André Bartetzki Audioinformatic, Berlin/Germany
Kurt Geihs Distributed Systems, Universität Kassel/Germany
Dominik Kirchner Distributed Systems, Universität Kassel/Germany
Andreas Witsch Distributed Systems, Universität Kassel/Germany
M.J.G van de Molengraft Control Systems Technology, TU Eindhoven/Netherlands
Harrie van de Loo Control Systems Technology, TU Eindhoven/Netherlands
3. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Artistic Concept
- every sound is a result of motion
- feedback as a special case in sound generation
Loudspeaker Microphon
Amplifier
4. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Artistic Concept
- every sound is a result of motion
- feedback as a special case in sound generation
- sound modules for producing feedbacks
- additional data from the board computers
the movements of the robots, determined by the
soccer game, are transformed into a piece of music.
5. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Robot Technology
- autonomous system behavior
- team coordination
- mobility
6. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Robot Technology
- autonomous system behavior
- team coordination
- mobility
Rotation
Slide
7. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Robot Technology
omnidirectional drive system
in 120 degree configuration
1 2 3 4
8. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Sound Processing Technology
9. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Sound Processing Technology
- Hardware Architecture of the Audio Modules
10. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Sound Processing Technology
- Hardware Architecture of the Audio Modules
microphone
11. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
Sound Processing Technology
- Hardware Architecture of the Audio Modules
- signal processing Software with SuperCollider
- communication with Open Source Control
12. Giving Robots a Voice: MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
13. Giving Robots a Voice:
MID SIZE ROBO SOCCER MUSIC
A Kineto-Acoustic Performance
by Ralf Hoyer
Audioinformatic: André Bartetzki
Team CARPE NOCTEM, Kassel
Team TECHUNITED, Eindhoven
Festival
inhuman music –
compositions by machines, animals and randoms
House of World Cultures in Berlin 2013 February 21st- 24th
Notas del editor
Dear ladies and gentlemen, dear colleagues, my name is Ralf Hoyer, I am a composer and sound artist living in Berlin, Germany. It´s a great pleasure for me to speak at the conference “ArtsIT 2013” in Milano, March 21-23, 2013. I would like to introduce a kineto-acoustic performance, created by autonomous soccer robots of the Robo Cup Mid Size League, which are equipped with specially developed sound modules for creating a complex sound structure. The project is entitled MID SIZE ROBO SOCCER MUSIC and was performed in the festival „ inh uman music – compositions by machines, animals and randoms in the House of World Cultures in Berlin, February 21-24, 2013. I first began thinking about this project 4 years ago, and the practical realization has taken about half a year.
The project could not have been realized without the powerful help of both the Robo Soccer Team TECHUNITED from the Technical University Eindhoven /Netherlands and the Team CARPE NOCTEM of the University of Kassel / Germany. Both teams participate annually in the Mid Size League competitions of the Robo Cup, and provided the expertise for the technical infrastructure. I also want to give warm thanks to the audio-information expert André Bartetzki as well as to Kurt Geihs, Dominik Kirchner, Andreas Witsch, René van de Molengraft and Harrie van de Loo. First, I will say something regarding the artistic conception. Secondly, I want to give you some information about the technology, and then I want to speak about the performances at the end of February in Berlin, including a few video impressions. (You can find the video at https://vimeo.com/64014768)
The starting point was the consideration, that every sound is a result of motion. Whether you knock two little stones in your hands, or you blow into a trumpet, or you give an electric impulse to a loudspeaker, it makes the air vibrate, so that we hear a sound. The type of sound generator is quite unimportant. A special case in electro-acoustic sound generation is the feedback between a microphone and loudspeaker. Depending on the distance between them, the amplitudes of several frequencies increase, and the tone grows to an unbearable level very fast. It´s a well known situation in every sound-check. The resonances in the room and the frequency response of all devices in the electro-acoustic chain also influence the result. Nevertheless, this effect is often used in sound art.
Watching a Mid Size Robo Soccer game a few years ago the idea was born to equip the soccer-playing robots with special feedback sound modules for generating a complex polyphonic sound structure. The emerging sound structure is audibly linked to the motion of the various autonomous robots. Their movements as independent players within the game generate the “robots’ voices ” . Additionally, data from the robot computers is evaluated for the sound modulation. For example the distance between the robots in conjunction with their speed, their positions on the playground, or the impulse of the kick. Thus, the movements of the robots are transformed into a piece of music. Due to the unique motions of each game, the sound cannot exist in any other form, and is different in each performance. The acoustical output might be perceived as speech, musical gestures or the wailing of a siren.
Creating a polyphonic sound composition in this way requires very sophisticated robot technology. The initial requirement is autonomous robot behavior, especially their ability to move and act independently. An all-round camera senses the environment and extracts data to measure the current situation, for example the positions of the robots, the ball, the goal, the teammates and, perhaps, obstacles. After a complex data processing the motion commands have to be transformed into motor velocities to execute the selected behavior. All of these quantities are measured in real time, and the number of possible states is quite infinite. Secondly, team coordination is required. The robots are aware not only of their own status, but also of the active states of the other robots and of the entire situation. Based on their “world model” they are able to act together purposefully. This is very important for a successful transformation into music, because music also typically consists of an overlay of multiple sounds into a meaningful arrangement. And last but not least, Mobility is a key requirement for creating sound structures in this way. As velocity and agility are inherent elements in Mid Size League competitions, a lot of effort is exerted by the community to realize and optimize the driving systems of the robots. Systems with 3 or 4 omnidirectional wheels, have been established as a quasi standard in the Mid Size League. Let´s have a short look at this fascinating sophisticated technical solution:
An omnidirectional wheel has a series of slide rollers attached to its lateral surface that are perpendicular to the rotational direction of the wheel. This allows an almost frictionless slide in the direction of the rotational axis of the wheel. The wheel rotation is drawn in green, the rotation of the slide rollers and the direction of movement of least friction is yellow.
Here you can see an omnidirectional drive system in 120 degree configuration. Each wheel velocity is individually controlled. The combination of all wheel velocities defines the global robot movement dependent on its geometry. The white arrows show different global movements. In No. 1 the movement is straight on, No. 2 shows a turn around its own axis, No. 3 shows a lateral movement and in No. 4 you can see a circular arc movement.
Now let´s reflect upon the sound processing technology. The application of networks for sound production, musical composition, and performance has a long tradition in experimental music and sound art. While these are still usually based on a traditional ensemble of individual human musicians, there is an ongoing evolution of concepts and pieces of art in which humans are just partially involved or are totally absent. Without human musicians or organizers, it is necessary to find different ways to structure the music and sound production. So we developed a distributed system for sound processing based on portable audio modules and a communication network.
The audio modules work autonomously but are connected to the hosting platform to allow reception of allocentric position and velocity. In addition, there is a main audio processing unit which can send and receive control data from the audio modules of all mobile robots. The main component of such a module is a single-board computer (BeagleBoard xM) with a 1 GHz ARM CPU. A small electret microphone capsule perceives the sound of the environment. The microphone is connected to a pre-amplifier that includes an adjustable noise gate, helping to separate the sound of the other robots from the background noise. After being processed by the software running on the BeagleBoard the audio signal is amplified and dispensed by the speaker. All audio modules are able to share features of the sound analysis and contextual information of the hosting platform with the central audio process via the wireless network.
Here you can see the the Beagle Board and the loudspeaker. The microphone is fitted on the back. The directional patterns of both, the cardioid microphone and the horn speaker, were chosen to minimize the probability of direct feedback between the audio output and the input of the module itself. For the same reason, the microphone and speaker are aligned to opposite directions. Each team of robots has 4 outfield players and a goal keeper. So there are 10 distinct sound sources. The horn shape of the speaker can be understood as a social and aesthetical reference to the megaphones and Vuvuzelas common for football games spectators in stadiums.
The signal processing software is written in SuperCollider, a programming language and run-time environment for sound synthesis, signal processing and algorithmic musical composition. The data is transferred using the Open Sound Control protocol.
Before we take a look at the video just a few remarks about the performances: In every project there is a difference between the idea and the outcome. That happened in these performances too. In this case, it was evident that the direct feedback sounds were less variant than expected. The robots agreed to a single feedback frequency within a very short time. So we developed several variant signal treatments for making other frequencies possible too. For example: (1) suppression of the dominating frequency with a resonance filter. (2) direct feedback, level dependent on velocity. (3) direct feedback, just active if robots are standing. (4) a random controlled resonance filter enabling short-time feedbacks of several frequencies.(5) short time delay of a single click dependent on robots speed. (6) frequency modulation of a low pitched sinus tone dependent on the position of the ball... etc. These several treatments were available simultaneously for live-mixing. Beyond the humour, this performance has a serious background. We observed that the “voice” of the robots reinforced the impression of human-related individuals. Extensive research exists for example in the field of neuroscience, with the goal to emulate human-like behavior and communication for autonomous robots. This is particularly the case in areas such as service robotics, health care, and search & rescue. Enhanced acceptance of these technologies among potential users needs to be established for robots to take on a role as helpful partners. However, it is also observed that users develop emotional ties to intelligent robots, because their quasi-human behavior also suggests the presence of human feelings. So the problem may arise that someone may be tempted to hand robots too much responsibility. This controversial issue requires our awareness in the development of robot technology.
But I should give you a warning: when I was speaking about music, I did not mean bass & drums or sweet melodies. Thank you for attention. (You can find the video at https://vimeo.com/64014768)