Poster Competiton

Presented my PhD work (which is mostly conceptual at this stage) at a showcase of current research at University of Bristol's faculty of engineering. The judges were suitably impressed and awarded me 2nd prize for my efforts (not bad considering there were 65 competitors from all branches of engineering and computer science).

Elumotion Arm

Started work a few weeks ago on controlling an anthropomorphic upper limb manufactured by Elumotion. At the moment I'm just using it as a hardware test bed in anticipation of working on a full humanoid torso however the anthropomorphic nature of the device means that certain motions can make you feel rather uneasy. This is the Uncanny Valley principle in action as one could say the arm looks like part of a 'moving corpse'.

Insect Braitenberg Vehicle

A few days ago I observed an insect behaving like a simple mobile robot (Braitenberg vehicle) with a line-avoiding algorithm. I ran and got my camera and the video below is the result.

Obviously Braitenberg designed his thought-experiment vehicles with insect behaviors in mind, a light stimulus to the right eye drives the left locomotor (wing / set of legs) to result in a moth that is attracted to light (or a 'photovore' in the case of the robotic vehicles).

By simple modification of this algorithm you can produce a creature that prefers darkness and can be made to follow a dark line on a light background (the dark line reflects less light into its sensors).

This insect appeared to be avoiding dark lines which makes sense as insects like light. But do insects really apply these rules to their perception of the environment on such a local level? The geek in me cries out for formalised experiments!

Unfortunately the best example (where it stayed in a circle for around 20 seconds) isn't on film (as it was what inspired me to get the camera).

It has crossed my mind that perhaps it doesn't like wet ink... some of the occasions where the insect stays in a circle for a while might disprove this but then it walks over some of the older lines quite happily.

I'd be very interested to hear people's opinions about this video...

Ad in a book (new and used)

Have just discovered that one of my papers has been published in a book available from Amazon for $180 -$539 (here is the link). The editor was the chair at a conference on Augmented Cognition that I presented at in 2005.

Here is the blurb:

"Bringing together a comprehensive and diverse collection of research, theory, and thought, this volume builds a foundation for the new field of Augmented Cognition research and development. The first section introduces general Augmented Cognition methods and techniques, including physiological and neurophysiological measures such as EEG and fNIR; adaptive techniques; and sensors and algorithms for cognitive state estimation. The second section discusses Augmented Cognition applications such as simulation and training, intent-driven user interfaces, closed-loop command and control systems, then goes on to explore lessons learned to date, and future directions in Augmented Cognition-enabled HCI."

Animatronics

Attended a very interesting meeting regarding a new EPSRC funded robotics outreach project. At the meeting were some Hollywood animatronics guys who among other things were responsible for the ABC-Warrior (below) from Judge Dread and the robot (Robbie?) from Lost in Space.



Though my work on human motion deals with asthetics to some degree it is quite a new twist to think of robots as only having to look good on camera. That requirement aside these guys are great at what they do (though I think the massive budgets help somewhat!).

BRL Project Description

A description of my PhD research project has been added to the Bristol Robotics Lab project pages website.

I like to keep this blog (mostly) seperated from my PhD so head to 'ANDy' to read about what I'm lucky enough to get paid to work on.

Keyboard Hack

Chopped up a broken M-Audio Keystation USB keyboard to make an effects box.

The keyboard had been dumped at the local tip and after taking it home I discovered it was because the keys didn't work. The pitch, modulation and volume controls all provided midi signals though and make a nice physical interface for music software. Much more enjoyable that moving sliders around with a mouse. Thanks to Pete for software advice.