Adaptive energy harvesting approach for smart wearables towards human-induced stochastic oscillations

Harvesting diverse renewable energy sources from humans offers an effective approach to achieving cleaner production and sustainable development. The integration of self-powered smart wearables enables continuous power supply for long-term operation of outdoor civilian and military electronics, especially in situations with limited access to conventional power sources. Resonant energy harvesting schemes working at the dominant frequencies of human movement can trigger large mechanical oscillations and significantly enhance the generation of electrical power. This study presents the development of a nonlinear two-degree-of-freedom (2DOF) resonant energy harvesting system with a maximum power point tracking (MPPT) algorithm that adapts to individual human differences and stochastic motion patterns. Through an analysis of the system characteristics based on equivalent circuit modeling, it is found that the electrical damping of the harvester plays a crucial role in determining the actual power output, and the relationship between its optimal value and the oscillation frequency is unique. Thereby, an energy harvesting circuit with an adaptive MPPT algorithm is developed to optimize the electrical damping in real time by applying a novel Perturb and Observe technique based on the average power over each time period. In addition, a variable-step search method and a special convergence judgment are proposed to enhance the tracking speed. The sensitivity analysis of algorithm parameters indicates that the time period and power tolerance need to be reasonably selected to strike a balance between search speed and accuracy. The results of constant motion pattern tests show that the proposed algorithm with a variable-step search method exhibits superior search speed and accuracy over fixed-step configurations. The results of variable motion pattern tests reveal an increase in the actual system power output of up to 194% compared to the conventional fractional open-circuit voltage (FOCV) algorithm. Finally, the system adaptability test is conducted to validate the MPPT efficiency in four resonant systems. The results showcase an impressive MPPT efficiency of 93%–99%, demonstrating its reliability and robustness in various application scenarios.