Interests

Synthesis of Physically Simulated Virtual Organism


... ...


My primary research interest lies in the evolutionary synthesis of bio-inspired robots and their controllers, for the purpose of providing more insight into how complex systems and biological evolution behave. I am also very keen to investigate the dynamical system's perspectives of neural networks to deal with the controllability of intrinsically unstable robotic systems. Further interests include computational neuroscience, evolutionary computation, chaos theory, and virtual reality.
Currently, I am especially excited by the possibilities of interactions between dynamical systems and their environments. I want to look more closely at the occurrence of transitions between different behaviours prompted by highly sensitive environmental changes, and to model them using physically simulated agents and neural networks.
As the media and computer arts, I am also interested in the evolutionary synthesis of virtual creatures. Yet, I do not agree that the current stage of this field has fully addressed technical challenges for practical applications. Working with dynamic simulations and evolutionary framework is a complicated task because of their complexity and computational cost. I am still seeking a way to improve physical and biological validities of evolved creatures. Nevertheless, it is worth in itself to see and enjoy these digital children.


Artificial Life
    Implication?

Physically Simulated Virtual Creatures
    underwater/terrestiral
    flying


Movies of early studies ( MPEG4 CODEC is needed )

Reimplementation of "BLOCKIES"
       land-1
       land-2
       land-3

       water-1
       water-2








. Two step evolution - Path following
       Aqua1 straight
       Aqua1 S-shape
       Aqua1 Spiral
       Aqua1 Heart

       Aqua2 straight
       Aqua2 S-shape
       Aqua2 Spiral
       Aqua2 Heart

       Aqua3 straight
       Aqua3 S-shape
       Aqua3 Spiral
       Aqua3 Heart
. Extention to the "Aero-Cylinders"
       Flyers by nested di-graphs

Two step evolution
       Cruise A-1
       Cruise A-2
       Cruise A-3

       Cruise B-1
       Cruise B-2
       Cruise B-3

       Follower-1
       Follower-2
       Follower-3



Recent study
     Flyers for Efficient Cruising
     Neural Controlled Flapping Robot with RPG [NEW]

Software in development
     NeuronomicoN - The Realtime Neural Network Simulator




Publications


Shim, Y. S. and Husbands, P. (To appear) Feathered Flyer: Integrating Morphological Computation and Sensory Reflexes into a Physically Simulated Flapping-Wing Robot for Robust Flight Manoeuvre. In 9th European Conference on Artificial Life (ECAL 2007), Lisbon, Portugal, September 10-14. [PDF]

Shim, Y. S. and Husbands, P. (2007) A Simple Evolutionary Framework for Generating Robust Flight Manoeuvre in a Physically Simulated Flapping Wing Robot. In Flying Robots and Insects (FIR 2007), Ascona, Switzerland, August 12-17. [PDF]

Shim, Y. S. and Kim, C. H. (2006) Evolving Physically Simulated Flying Creatures for Efficient Cruising. Artificial Life 12(4), MIT Press. [PDF]

Shim, Y. S., Kim, S. J. and Kim, C. H. (2004) Evolving Flying Creatures with Path Following Behaviour. In Proceedings of 9th International Conference on the Simulation and Synthesis of Living Systems (ALIFE IX), Boston, Massachusetts, September, 2004. [PDF]

Shim, Y. S., Shin, S. Y. and Kim, C. H. (2004) Two-step Evolution Process for Path-Following Virtual Creatures, In Proceedings of Computer Animation and Social Agents(CASA 2004), Geneva, Switzerland, July 7-9. [PDF]

Shim, Y. S. and Kim, C. H. (2003) Generating Flying Creatures using Body-Brain Co-Evolution. In Proceedings of ACM SIGGRAPH/Eurographics Symposium on Computer Animation (SCA'03), July, 2003, San Diego, Ca. [PDF]






Return to home

¨Ï 2002, Jeff Shim, All rights reserved