Proposal by Lucas De Sordi
The 4th Revolution is profoundly changing the way we live, our idea of property. The family car, the bicycle, workspaces, tools, all those things are becoming services. Things are radically evolving, yet little is said about how distributed robotics and AI will change the way we occupy spaces, allowing them to rapidly adapt for different uses, and this way, create abundance and democratize the spaces.
ENVOLVE proposes an investigation of self-conforming spaces by building an AI driven volume that may bring us a rare opportunity to play with spatiality by changing its usually static characteristics, such as: shape, size and height, and then investigate in real-time the aesthetic unfolding in individuals.
The name comes by mixing two words: “involve” refers to the most important component in architecture, and maybe in our lives, the sense of sheltering and belonging; and “evolve” as referring to the IA, continuous learning, pursuing harmony on human interactions.
“Architecture domesticates limitless space and enables us to inhabit it, but it should likewise domesticate endless time and enable us to inhabit the continuum of time.”
― Juhani Pallasmaa
DESCRIPTION / TECHNICAL DETAILS
ENVOLVE is a kinetics and light art installation that invites people to interact with it by moving around, entering, dancing and allowing the machine learning algorithm to improve its behavior and response capability. It consists of an array of kinetic vertical light rods arranged radially around a focal space - a stage -forming a shape-shifting light mandala.
Each light rod module consists of a RGB addressable led bar that can perform very complex light sequences. The bar is hanged by wires and controlled by two motors in line, that synchronously move each bar in an individual plane in two-axis: height and distance from center. By combining this set of rods, the whole system can arrange it´s cage in infinite sizes and shapes.
One quality of this art piece is that its dimensions and number of light elements can be easily altered and adapted for different exhibition spaces and budgets.
The installation has two different interaction inputs: visual, by tracking in real-time the spectators’ position, movement and gestures using a camera and Open Pose deep learning framework; and an audio channel, that can be connected to a microphone registering and analyzing ambience sound.
MACHINE LEARNING & PERFORMANCE
The sensors are connected to a main computer that runs the algorithm and triggers the whole pack of actuators (motors, motor drivers, led stripes, etc.), performing all the magic.
The system will search for harmony with the performer and ambient music, search for occupation and respect some parameters such as, avoiding collision by keeping some distance from spectators, and controlling movement speed, especially in the horizontal vector, to avoid the swinging of the lightbars, and in case of collision, retreating the light to the dock position in the ceiling, stopping its movement.
Besides that, it is necessary to create standby or demo mode, also to create some level of engagement parameter to the installation, so in case the spectator is too calm, the installation can approach them, involve and invite them to interact. The camera detects and avoids spectators but focuses its attention to the person on the focal center of the space, the stage. So, the installation can be contemplated by many spectators at once, but the interaction dwell time is expected to take 5-10 minutes.
Despite the complex light possibilities, the idea is to focus on movement and create an austere space, using the light temperature to set an atmosphere, it can be used - and I expect to do it - in a more agitated artistic performance as part of the installation demonstration, allowing to create an AI driven scenographic kinetic presentation.
As I see it, the production process of the piece, can be accomplished by three distinct phases, that can be arranged like this:
1. Conceptual Studies - at Deutsches Museum in Munich for 1 month
2. Physical Assembling - at ZKM | Karlsruhe for 1 month,
Workshops and a big set-up area will be necessary.
3. Programing the AI - at ZKM | Karlsruhe for 1 month,
The support of the developers’ team will be necessary.
Green screen room and the volumetric capturing system
may be helpful for experimenting with spectators tracking.
Covid: In circumstance of the Covid pandemic becoming any kind of obstacle or a mobility issue, I believe building and coding it remotely with the developers’ support is still completely manageable.
I am Lucas De Sordi,
an architect from Brasília, the Brazilian capital.
I hold a Master’s degree in Interactive Medias from IaaC - Institute of Advanced Architecture of Catalonia, institute connected to Fab Lab Barcelona, Spain, and Bachelor in Architecture and Urban Planning.
My passion for technology and an experimental approach to design, I believe, encouraged me to develop a wide range of skills, such as 3D modeling, Parametric Design and basic coding skills with Processing and Python. I am also proficient with Digital Fabrication tools, like 3D Printing, CNC Machines and Robotics.
I am currently living in São Paulo, where I came to join the team of, one of the finest scenography studios in Brazil, and where I had the opportunity to work in projects focused on immersive and sensible experiences.
At the moment I am looking to explore an artistic bias and develop authorial projects on non-traditional spatiality and ephemeral experiences, especially using of media art and kinetic installations.
+55 61 98333-9933