Phonotaxis in crickets and robots

Over the past decade, we have built and tested several robot models to investigate a particular biological behavior, the sound localizing (phonotaxis) ability of the cricket. This work has had several purposes. One is to develop robotic technology, such as novel sensors and control systems, by copying biology. However, the primary motivation is the "reverse"-to use the technology to develop understanding of biological systems-in particular, how neural circuits control sensorimotor behavior. This is effectively a new methodology for biological modeling, discussed in Webb (to appear). In this chapter, the aim is to provide a summary of the problems addressed and the key results to date. (More-detailed presentations of the biological background and the implementations can be found in Lund, Webb, and Hallam, 1997, 1998; Webb and Scutt, 2000; and Webb and Harrison, 2000.)