next up previous contents
Next: Other Artificial Animals Up: Conclusion and Future Work Previous: Impact in Computer Vision and

Potential Applications in Ethology

tex2html_wrap3948

Ethologists have long hoped to simulate animals in computers in order to facilitate the systematic study of animal behavior. As early as the seventies, ethologists Rémy and Bernadette Chauvin expressed their keen interest in such an approach to the analysis of animal behavior [Chauvin and Muckensturm-Chauvin1977]. More recently, McFarland McFarland93a and Roiblat Roiblat94 also emphasized the importance of such an approach:

``It is possible that ethologists will profit to a greater extent from the possibilities offered by simulation models... The availability of simulation programs makes it possible to identify which instructions devolve from a purely ethological interpretation and which from a cognitive one. A comparison of the program's behavior with the behavior actually displayed by the animal should enable the validity of the corresponding interpretation to be assessed.'' [Roiblat1994]
To this end, our artificial fish can act as a prototype model animal that provides a novel and potentially useful investigative tool. Although in several ways crude at this stage, our model can be furnished with more sophisticated behavioral mechanisms, more elaborate muscle models, etc.

Prior simulation models only allowed the modeling of one aspect of the problem to be investigated in isolation. Our model of artificial animals exemplifies a unified and complete system that allows a whole animal to be studied, from the biomechanical motor system to the perception-driven behavior system within a physics-based virtual world. In this way our approach allows the investigation of complex interactions between these different levels.

Moreover, the results of our simulation are displayed in a convenient and easy to interpret form. In effect what we have produced can be viewed as a sophisticated scientific visualization tool for ethologists that may enable new insights to be gleaned and new relationships discovered. Our work demonstrates the potential of realistic computer animation in applications to ethology.


next up previous contents
Next: Other Artificial Animals Up: Conclusion and Future Work Previous: Impact in Computer Vision and
Xiaoyuan TuJanuary 1996