SPIKING NEURAL NETWORKS ON SELF-UPDATING SYSTEM-ON-CHIP FOR AUTONOMOUS CONTROL

摘要

The artificial intelligence (AI) technique has suffered in solving its computationally hard problems in recent years. In this paper, a self-upgrading autonomous system is designed to tackle end-to-end AI-hard problems and to achieve self-adapting communication via modular and hierarchical extension from linguistic and semiotic constructs. A system-on-a-chip (SoC) self-adaptive control system can learn arbitrary shape of the robot body or machine parts. Simulation results have proved the effectiveness of learning abilities of the proposed autonomous system.