IEEE WF-IoT Session: Leveraging Human Gait Characteristics Towards Self-Sustained Operation of Low-Power Mobile Devices
Vishwa Goudar, University of California, Los Angeles, USA; James B Wendt, UCLA, USA; Miodrag Potkonjak, University of California at Los Angeles, USA; Zhi Ren, University of California, Los Angeles, USA; Paul Brochu, University of California, Los Angeles, USA; Qibing Pei, University of California, Los Angeles, USA
The proliferation of mobile ubiquitous devices faces a hurdle in the form of high resource consumption rates that restrict longevity. Several low-power devices and application designs and optimization techniques have been proposed. Simultaneously, energy harvesting technologies are increasingly being viewed as a complementary technique to drive down resource consumption rates and even achieve self-sustenance. Towards this end, we propose a foot-strike powered harvester array composed of a novel high-energy density material called Dielectric Elastomers. To compensate for their control parameter sensitivity, we propose an adaptive closed-loop control algorithm based on general characteristics of human gait. From experimentally collected datasets of human plantar pressure and detailed characterization of DE behavior, we show that our algorithm yields enough accuracy to produce upwards of 85\% of the maximum energy harvestable by the DE array. We also show that, in many cases, this is sufficient to fully drive low-power mobile ubiquitous applications.