Emotional contribution process implementations on parallel processors

Abstract

An emotional agent software architecture for real-time mobile robotic applications has been developed. In order to allow the agent to undertake more dynamically constrained application problem solving, the processor computation time should be reduced and the gained time is used for executing more complex processes. In this paper, the response time of the operating processes, in each attention cycle of the agent, is decreased by parallelizing the highly parallel processes of the architecture, namely, emotional contribution processes. The implementation of these processes has been evaluated in Field Programmable Gate Array (FPGA) and multicore processors. © 2011 Springer-Verlag.