Extending the Service Lifetime of a Battery-Powered System Supporting Multiple Active Modes

reducing the power consumption of a system while maintaining an acceptable performance degradation level. This paper presents a Reinforcement Learning-based dynamic power management framework for extending the battery service lifetime of a system with multiple active modes. In the proposed framework, a Power Manager adapts the system operating mode to the actual battery state of charge based on the observation of the battery output voltage. Moreover, it uses the reinforcement learning technique to accurately define the optimal battery voltage threshold value and use it to control the system active mode. Additionally, the power management policy is automatically adapted to the optimal timeout value. The proposed algorithms are experimented on both single and dual-battery powered systems and the obtained results confirm their excellent performance for defining the optimal power management policy particularly for high latency constrained systems. Experimental results show an improvement in battery State of Charge (SoC) savings up to 42.83% and 55.38 % respectively for single battery and dual battery systems in comparison with classic pre-specified timeout policies using a given preset battery threshold voltage