DocumentCode :
3325282
Title :
Neural preprocessing of auditory-wind sensory signals and modular neural control for auditory- and wind-evoked escape responses of walking machines
Author :
Manoonpong, Poramate ; Wörgötter, Florentin ; Pasemann, Frank
Author_Institution :
Bernstein Center for Computational Neuroscience, University of Göttingen, Bunsenstr. 10, 37073, Germany
fYear :
2009
fDate :
22-25 Feb. 2009
Firstpage :
786
Lastpage :
793
Abstract :
The flying crickets Teleogryllus oceanicus have sound sensitive organs to elicit an “acoustic startle response”. Another kind of the startle response is also evident in the cockroaches Periplaneta and the crickets Gryllus bimaculatus where they use their cercal filiform hairs (wind sensitive hairs) to trigger so called “wind-evoked escape behavior”. Similarly, in this paper a setup is described where an auditory-wind detector sensor and a neural preprocessing system together with a modular neural controller are used to simulate such animal behaviors in an abstract form on a six-legged walking machine. The neural preprocessing network acts as a low-pass filter and sensory shaping unit. In addition, the modular neural controller then generates the desired responses such that the machine performs a fast movement away from abrupt, intense, and unexpected auditory and/or wind stimulus in a physical environment.
Keywords :
Acoustic sensors; Animal behavior; Detectors; Hair; Legged locomotion; Personal digital assistants; Robot sensing systems; Sense organs; Sensor systems; Ultrasonic imaging; Autonomous hexapod robots; Biologically-inspired robots; Neural control; Recurrent neural networks;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Robotics and Biomimetics, 2008. ROBIO 2008. IEEE International Conference on
Conference_Location :
Bangkok
Print_ISBN :
978-1-4244-2678-2
Electronic_ISBN :
978-1-4244-2679-9
Type :
conf
DOI :
10.1109/ROBIO.2009.4913100
Filename :
4913100
Link To Document :
بازگشت