Autonomic Computing Towards Resource Management in Embedded Mobile Robots

• Renan Maidana PUCRS
• Aurélio Salton UFR
• Alexandre Amory PUCRS

Resumo

Small and lightweight robots are becoming a reality due to the increasing pervasiveness of embedded computers. However, much of the research in mobile robotics is focused on increasing the robots' autonomy, disregarding the computational limitations of embedded computers. The area of autonomic computing defines self-managing systems, also known as autonomic systems, which manage computational resources (e.g. CPU load) and reduce the effect of processing bottlenecks, as well as alleviating energy consumption in embedded mobile robots. In this paper, we propose the use of autonomic systems with control theory for resource management in embedded mobile robots. Our expected contribution is to use autonomic computing to develop self-contained embedded robots, capable of operating with few embedded computers and without the aid of external servers. We implement an autonomic system which successfully controls the latency and CPU load of a dense optical flow application, running on a Raspberry Pi 2 embedded computer, which validates the proposed idea.

Palavras-chave: Autonomic systems, Computers, Robot sensing systems, Mobile robots, Boats