Energy-Aware Task Scheduling for Edge IoT Deployments

Abstract

Battery-constrained edge devices must balance responsiveness against energy budgets. This paper introduces an adaptive scheduler that predicts workload bursts and shifts non-critical inference to low-power windows. In a smart- building pilot spanning 240 sensor nodes, the scheduler extended median device uptime by 2.3x without violating latency service-level objectives, showing how applied scheduling policy can materially reduce operational cost at the edge.

Keywords