Uncertainty-aware autonomous sensing with deep reinforcement learning

Abdulmajid Murad, Frank Alexander Kraemer, Kerstin Bach, Gavin Taylor

Abstract: Highlights•Framework for autonomous sensing controlled by deep reinforcement learning and guided by an uncertainty-aware prediction model.•Novel reward function based on measured peaks-over-threshold.•Detecting and dense-sensing critical periods of interest.•Representing a phenomenon with an estimation model informed by sparse measurements while reducing energy expenditure.•Spatio-temporal generalization of learned sensing policies.