Hybrid Sentient Canopy: An implementation and visualization of proprioreceptive curiosity-based machine learning (2016)
article⁄Hybrid Sentient Canopy: An implementation and visualization of proprioreceptive curiosity-based machine learning (2016)
abstract⁄This paper describes the development of a sentient canopy that interacts with human visitors by using its own internal motivation. Modular curiositybased machine learning behaviour is supported by a highly distributed system of microprocessor hardware integrated within interlinked cellular arrays of sound, light, kinetic actuators and proprioreceptive sensors in a resilient physical scaffolding system. The curiositybased system involves exploration by employing an expert system composed of archives of information from preceding behaviours, calculating potential behaviours together with locations and applications, executing behaviour and comparing result to prediction. Prototype architectural structures entitled Sentient Canopy and Sentient Chamber developed during 2015 and 2016 were developed to support this interactive behaviour, integrating new communications protocols and firmware, and a hybrid proprioreceptive system that configured new electronics with sound, light, and motion sensing capable of internal machine sensing and externally oriented sensing for human interaction. Proprioreception was implemented by producing custom electronics serving photoresistors, pitchsensing microphones, and accelerometers for motion and position, coupled to sound, light and motionbased actuators and additional infrared sensors designed for sensing of human gestures. This configuration provided the machine system with the ability to calculate and detect realtime behaviour and to compare this to models of behaviour predicted within scripted routines. Testbeds located at the Living Architecture Systems GroupPhilip Beesley Architect Inc. LASGPBAI, WaterlooToronto, Centre for Information Technology CITA, Copenhagen National Academy of Sciences NAS in Washington DC are illustrated.
|
|
| Year |
2016 |
| Authors |
Beesley, Philip; Ilgun, Asya Zeliha; Bouron, Giselle; Kadish, David; Prosser, Jordan; Gorbet, Robert; Kulic, Dana; Nicholas, Paul; Zwierzycki, Mateusz. |
| Issue |
ACADIA 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines |
| Pages |
362-371 |
| Library link |
N/A |
| Entry filename |
hybrid-sentient-canopy |
