LIGHT & SHADOW

by Tairan hao

Project Brief: This work demonstrates the ability of the most basic AI - based on time and time of learning and database analysis, trying to use artificial light and sound to predict some unknown possibility. The machine understands the information that is returned to the person's machine based on the behavior of the observer (information entered by the person).

Deep interpretation of the emergence of AL, AI can not change the natural laws of the human universe - everything is constantly changing, there is a certain degree of uncertainty. When this change of uncertainty is reflected in AL/AI, this change becomes gradually predictable, showing a vague and predictable prediction.

The high fidelity machine of this project should be displayed in a dark space with 4 sides of mirror, which can reflect the laser light into different angles and direction that creating various visual effect. For prototype, the main idea is to make the machine work and describe the concept. Therefore, demonstration videos are shot in the daylight environment.

Tool: Wekinator, TouchDesigner, Arduino, Kinect, Leap Motion, Processing

Duration: 12 weeks

Credits: Low Tech Art Lab(Technical support)

INTRODUCTION

According to the behavior of human input, the three laser lights produce a random visual effect (the machine remains stationary if there is no input in one space. This also gives a feeling of communication between humans and machines.

MACHINE INTERACTION SYSTEM

STEP 1: Kinect Sensors read inputs from people

STEP 2: Wekinator receives input data(human skeleton) and starts to run algorithm in order to run the servos

STEP 3: TouchDesigner receive data from Wekinator and simulate the output data.

REVISED CULTURAL AI DESIGN TOOL

RULES AND DATASET

I build a database that collects differents motion of human. I also build an independent database that has different x,y,z position of the machine. The rules I created are trying to give a feeling that the machine is responding a similar move to the behavior of people.

CONSTRUCTION & MEASUREMENTS

Size: 40cm * 40cm * 40cm (can be changed according to hardware size or other requirements)

TECHNICAL KEY POINTS

Installed Motion Interactive System Design:

3 Robotic Arms; 6 joints total

Based on the position of people to determine the reaction of robotic arm, in the x, y, z coordinate of robotic arms in the physical space

extract the coordinate data of people’s motion and calculate, reflect on the physical machine

Challenging Parts:

Calculation, machine learning and algorithm through the simulation of people’s motion

Building the independent database of both machine’s motion and people’s motion

The data transmission between Kinect, TouchDesigner, Arduino and Wekinator

Future Iterations:

Finalize, enrich and expand the interactive system

The machine can imitate what the input does(human’s behavior) as a mirror

Adding more spreading medium in the space to create and emphasize the visual effect

PHYSICAL OBJECT

PROCESS

IDEA GENERATION & SKETCHES

Cultural AI design tools help me to clarify my concept and how i want the machine to work. After dozens of iteration, i finalize my inputs and outputs, which are motions and laser light reflection in a space. The visualization of laser light displayed in a space can express the communication between human and machine.
After figure out what to do and how the machine operate, i started to draw sketches, material lists and made a detailed plan of the creating process. It includes what software do I use, what materials, 3D printing, laser cutting as well as calculating the measurements.

RAPID PROTOTYPING

Included Materials:

Software: Arduino CC, Wekinator, Processing

Hardware: Arduino Uno, Leap Motion, Bread Board, Jumping Wires, Servos, Brackets

Processes and Challenges:

In the prototyping process, instead of using human motion as input, I chose to use leap motion as input since it is easier to build and control. The basic idea was to use finger gesture to control (affect) the servos running. The challenges i faced was to connect the Arduino and processing software in order to make the code react to the input responses.

The First time made servos running, controlled by hand gestures

The prototype connects with a laser light pen