Electronic device for signal processing with embedded optimization of electrical energy consumption and a corresponding method

Abstract

This electronic device for signal processing with embedded optimization of electrical energy consumption includes: at least one sensor and at least one actuator; hardware for storing at least one correspondence table assigning a quality value to each pair formed by a state value measurable by each sensor and an action value adjustable by each actuator; processing hardware comprising an electrical energy consumption manager with a reinforcement learning engine designed to selectively choose between an exploitation and a random exploration in each correspondence table so as to select a new action and transmit it to each actuator. The learning reinforcement engine has a random exploration limit parameter, comprising an exploration distance value in each correspondence table, around a pre-selectable action value in each correspondence table.