Architecture of a predictive power management system for an autonomous mobile robot under stochastic pulse load

Abstract

The issue of ensuring the energy stability of autonomous mobile robots equipped with electric pulse actuators operating under stochastically distributed pulse loads is discussed. A unique architecture for an energy redistribution system is proposed based on the concept of a single DC bus with predictive digital control of traction electric drives. A mathematical framework was developed including a stochastic pulse load model, a system energy survivability criterion, and a law for adaptive traction power damping. It was found that the stochastic generation of high-voltage discharges (up to 5 kW) with limited power supply capacity led to a constant-power load mode which caused critical voltage sags in the DC link, accelerated degradation of lithium-ion batteries, and emergency shutdowns of the protection system. Differential equations were derived for the DC bus state equation and for the traction motor angular velocity correction. The research target was an on-board DC network with a nominal voltage of 48 V. A method for predictive power supply management of an autonomous robot was proposed using the adaptive traction damping law. A comparative analysis of two modes was made: a hard-wired load and an active stabilization mode. It was found that the application of the developed algorithm reduced the maximum voltage drop from 12.42 V to 4.56 V (a 3-fold increase in stability) and decreased the peak current consumption from the battery from 114.6 A to 76.3 A (a 33% reduction). A slight increase in the dynamic directional stability error (from 0.5 cm to 1.2 cm) was also detected which did not critically affect the quality of agricultural operations.