Safe Parallelism for Behavioral Control

Abstract

During the Spring 2008 semester at Olin College, we introduced the programming language occam-pi to undergraduates as part of their first course in robotics. Students were able to explore image processing and autonomous behavioral control in a parallel programming language on a small mobile robotics platform with just two weeks of tutorial instruction. Our experiences to date suggest that the language and tools we have developed allow the concise expression of complex robotic control systems, and enable the integration of events from the environment in a consistent and safe model for parallel control that is directly expressed in software.