Field Studies Version 2
Field Studies Version 2 at Oboro in Montreal showing robot pens and Anthropod Tales video on wall

Jessica Field

Field Studies Version 2, 2009

Field Studies V2
Field Studies V1
Robot Zoo Project
SICB
Early Robots
Allegory of Pulse
Automata Films
Tantalus Synthesis
Performances
Mechanical Sculptures
Youth Workshops
Adult Workshops
Bio
Exhibitions
Contact Info

Field Studies on Walking Robots from Jessica Field on Vimeo.

Installation Overview

The Field Studies project is a large installation comprised of 19 robots of varying levels of intelligence. The first level are 10 Protozoan Flagellates, single celled robots, that are shown on a metal table. Next are a group of 4 Hydrozoan robots that uses sensory inputs to drive their outputs. The last group of robots are 5 Anthropods that are displayed in 3 pens on the floor. These Anthropods are modeled from the same AI system as the robot vacuum cleaners and use behaviour based programming.

The installation also has 3 videos presenting the robots. Each video takes on the assumption that the robots are real animals and explains the robot's habitat from this bias. The video Like a robot give the audience in inside view of what it is like to see like the Anthropods exhibited. The other video Ideological Ecologies is a documentary describing the Hydrozoan's ecology. It is a stop motion animation designed to make the audience skeptical about the truth of the video's content. The last video, Anthropod Tales is a narrative much like the nature films from the National Film Board of Canada's Hinterland's Who's Who humanizes the robots and shows their real nature has none of the qualities represented in the video.

The last major element the this installtion are 5 book works that contain the Evolutionary History of each Anthropod in the form of the code that was written for each version of the robot’s AI that lead to its current evolution. The code is written in Assembler with my notes on why each new version of code had to be written in order to improve the robots' abilities to interact with the other robot's and their environment. On looking at these books, the reader gets the full sense of how simple the robots are and how complicated it is to make robot do the most simple activity. The books also play on the idea that the robots are evolving. Each has gone through less than a hundred variations and explains their utter simplicity. This begs the question: How advanced could they become if the creator kept trying to evolve them for millions of years?



The still images in the all the videos of the gallery space were taken by Paul Litherland.


Canada council for the arts logo
This project was realised with the support of the Canada Council for the Arts

Concept

This project addresses social issues concerning the commercial creation of artificial intelligence robots that attempt to fulfill human expectations and desires. Field Studies uses video works to create a facade of expectation in the viewer. The videos are idealistic, depicting the robots in a virtual manner where everything they do is harmonious, orderly and fulfills the desires of the artist by doing everything she wants them to do. The real robots in the exhibition do not embody these behaviours; they have their own intelligence and shortcomings and are nearly always out of sync with the video’s expectations. The robots instead have behaviours of their own and their lack of evolution towards the artist’s expectations creates a character that is their own. more and less

The idea of evolution is explored by having three different classifications of robots where each robot interacts within its class in an ecosystem. Each class of robot has a different level of sophistication. The first, Protozoan Flagellates, carries the intelligence of a single cell organism. The second classification, Hydrozoans, use a nerve-net intelligence structure; it uses its senses to directly control the whole body’s reaction without any process of thought interpreting the experience. These robots are static and incapable of any form of evolution. They have been created from conception to reality without having any way of reprogramming them to improve their behaviours. The last class of robots, Anthropods, can evolve. They can be modified indefinitely to work in their environment as efficiently as possible by being able to remember their experiences and use logic to interpret their sensory data. With all this going for them, they still do not quite fulfill the grandeur of what they can do in their video. It then comes down to the question of how many times do these robots have to evolve before they work perfectly to the artist’s expectations.

The mainstream idea of a robot is a stereotype offering immediate gratification and is designed to fulfill human expectations and desires. Yet the robot itself has many hardware limitations in what it can actually do. In working with these limitations, the robot evolves very slowly and when finished, only fulfils the smallest shadow of the ideal we desire for them. Our imagination of what we want the robot to become fills in the rest of the details.



The still images in the all the videos of the gallery space were taken by Paul Litherland.


Canada council for the arts logo
This project was realised with the support of the Canada Council for the Arts

Anthropod Tales (Video Work)

Anthropod Tales from Jessica Field on Vimeo.

Anthropod Tales is a collection of videos based on the five Walking Anthropod Robots and presents itself like a Fairy Tale. These fairy tales are based on programmed robots that have each been given a tragic flaw. They are all autonomous and capable of acting out their part as described in the story. The robots were built and programmed first with the video serving as an interpretation of their final behaviours. Robot designs were copied from all three of Karl Williams How to build a robot books (Insectronics, Ambhibionics, and Build Your Own Humanoid Robots). They were then modified. All programming, electronics, and circuit board designs are from Jessica Field.

Anthropods

This is a series of five robots that are separated by different pens. The video Anthropod Tales is about what I want the robots to appear like to the public. What they actually do is another story. These robots are programmable, meaning they can evolve. Each program I write is the next phase of evolution. So the question becomes how many phases of evolution will it take for these robots to perfectly embody the ideal I have given them in the videos? In the meantime, are the robots successfully behaving like artificial organisms? Even more important, do they appear more like an artificial life form in the process of being created then what they will be when the project is completed? By meddling with the robots, will the artist loose the dynamic behaviour of the robots slowly over time molding them to her ideals or will their character's improve as they are modified and "perfected?"

The Number of Generations that Each Anthropod has Completed

The Bi Magnetotransis has completed 15 phases of evolution. The Quadra Magnetoconsumous has finished 13 generations of evolution. The Hexa Infraseekozoid, being the most complex, had to undergo 23 phases of evolution to have it begin to be able to mate with other robots. The Bounci Vicarioustransis is the simplest robot and has only had 3 generational evolutions. The Sidewind Externalreactozoid has had 10 phases of evolution to date.



The still images in the all the videos of the gallery space were taken by Paul Litherland.


Canada council for the arts logo
This project was realised with the support of the Canada Council for the Arts

Ideological Ecosystem (Video Work)

Ideological Ecologies from Jessica Field on Vimeo.

This animation depicts how I want my Hydrozoans to interact in an ideal ecology. This video is part of an installation where a juxtaposition between the real robots, which do not have the ability to evolve, and the artist’s unattainable social ecology are shown in tandem. The Hydrozoans have no brain and it is unknown whether they could ever fulfill their expected behaviours from the narrator in the video. They function like a nerve-net. Sensor data is intermediated to control the motors that drive the robot. The robots were influenced by Tamiya toys for walking gaits but redesigned to have full autonomous behaviours.

Hydrozoens - The Robots

The two legged robot on the left is Bi Thermotaxis and on the right is Macropod Motivetaxis.
These robots are a species of Cnideria, meaning that they have a nerve-net central nervous system. These robots have no chance of evolving. They have sensors, intermediate circuitry and motor outputs. They have no brain, nor can they be reprogrammed.
There exists a facade where humans are creating a fantasy of what they think artificial intelligence should be rather than looking at the reality before them to see what artificial intelligence could be. The video, Ideological Ecologies, is about these Hydrozoans. It shows the facade of stereotypical roles I would like my robots' to adopt in their behaviours. These robots do not share the stereotypes that the video claims they have as I have no ability to control or change them.
The rules used to create these robots are as follows: 1) I will build the robots so they function individually; 2) The moment they work I will stop the adjusting them, as this is where the manipulations begin in order to create a robot that suits the creator's mental image. The video will be used to show how the robots are meant to behave and will provide a juxtaposition against their actual performance. Thus the unexpected results of my first attempt can be regarded as a robot behaving beyond human expectation. In seeing them as they are, one can begin to wonder if these robots are a form of artificial life.
The surface of their environment is covered with iron filings that shows the traces of their movements over the course of the installation. A six legged robot can draw lines so its activities can be traced. The other robots behaviours can only be interpreted by their foot prints.



The still images in the all the videos of the gallery space were taken by Paul Litherland.


Canada council for the arts logo
This project was realised with the support of the Canada Council for the Arts

Protozoen Flagellates

Protozoan Flagellates from Jessica Field on Vimeo.

Inspired by a toy from Tamiya (using vibration propulsion), I created a single cell robot, using a solar panel, a solar engine and a vibration motor. To add a dynamic to the society, I made all of the single cell robots slightly different. Some have more absorbent solar cells than others and each has a different energy storage capacity. What you see in the video is the result of these differences.



The still images in the all the videos of the gallery space were taken by Paul Litherland.


Canada council for the arts logo
This project was realised with the support of the Canada Council for the Arts

Like a Robot (Video Work)

Like a Robot from Jessica Field on Vimeo.

This video is a representation of what it is like to see like an Anthropod.



The still images in the all the videos of the gallery space were taken by Paul Litherland.


Canada council for the arts logo
This project was realised with the support of the Canada Council for the Arts

Technical Video on how Field Studies Version 2 robots work

Field Studies Technical Video from Jessica Field on Vimeo.

This video offers a technical explanation on how the robots in the Field Studies Project Version 2 work. It explains the theory behind how the each robot type was made using simple animations.

The designs of the Anthropod robots were researched from Karl Williams book series on how to build your own biologically inspired robots. The designs used to make a walking robot were found in these books: Amphibionics, Insectronics and Humanoid Robots. The designs were then modified to improve robustness and to suit all the programming and power requirements of the actual robot species.

The designs for the Hydrozoens and the Protozoans came from the inspiration of Tamiya robot kits and then were distorted and improved from there by making them fully mobile objects that could move in all directions instead of simply forward. The protozoans were influenced by a vibrating fox.

Originally the Field's Study project was intended to make comments on the world of robot kits but the piece moved away from these ideas to the ones described earlier. I am fascinated with kits as they are useful to teach you how to make something specific but in practice these kits do not function very well or for very long. This is of course to keep the learning process cheep and affordable but important knowledge in how to make something robust must be figured out through tinkering. Thus to make a kit very robust, you have to spend long hours making improvements. I am interested in the ideas of what is good enough for the learning process and why is it okay that they don't have to be made better? The intention of kits seem to be designed to give pleasure in the making, then you can play with them for a bit, then they sit on a shelf and collect dust which is what most of my Tamiya robot toys are doing now.



The still images in the all the videos of the gallery space were taken by Paul Litherland.


Canada council for the arts logo
This project was realised with the support of the Canada Council for the Arts