Optimizing Robot Control Systems for Low Power and Reconfigurable Architectures

Abstract

Robot control systems rely on Finite State Machine (FSM)-based architectures for efficient decision-making and task execution. However, as robotic applications grow in complexity, optimizing FSM-based control systems for power efficiency, area, and reconfigurability becomes crucial. This paper presents a novel FSM decomposition strategy tailored for robot control systems, reducing power consumption and reconfiguration overhead while maintaining performance. By decomposing a given FSM into two sub-machines, the number of essential transitions is minimized, enabling selective activation of sub-machines to conserve power. The proposed method introduces an enhanced decomposition technique that significantly reduces redundant transitions, leading to improved efficiency. Experimental validation using FPGA-based synthesis demonstrates a 50% reduction in essential transitions, a 40% decrease in reconfiguration time, and an increase in clock frequency up to 92.7 MHz with reduced resource utilization. Additionally, we compare our approach with existing FSM decomposition techniques, demonstrating its advantages in terms of power efficiency and scalability.