Miniature autonomous systems are being developed under ARL's Micro Autonomous Systems and Technology (MAST). These systems can only be fitted with a small-size processor, and their motion behavior is inherently uncertain due to manufacturing and platform-ground interactions. One way to capture this uncertainty is through a stochastic model. This paper deals with stochastic motion control design and implementation for MAST- specific eight-legged miniature crawling robots, which have been kinematically modeled as systems exhibiting the behavior of a Dubin's car with stochastic noise. The control design takes the form of stochastic receding horizon control, and is implemented on a Gumstix Overo Fire COM with 720 MHz processor and 512 MB RAM, weighing 5.5 g. The experimental results show the effectiveness of this control law for miniature autonomous systems perturbed by stochastic noise.

Date Published: 3 June 2013
PDF: 9 pages
Proc. SPIE 8725, Micro- and Nanotechnology Sensors, Systems, and Applications V, 87251D (3 June 2013); doi: 10.1117/12.2016320

Published in SPIE Proceedings Vol. 8725:
Micro- and Nanotechnology Sensors, Systems, and Applications V
Thomas George; M. Saif Islam; Achyut K. Dutta, Editor(s)