Energy generation technologies that use piezoelectric materials as uninterrupted power supplies are one of the most practical solutions of low-power wireless sensor network. The piezoelectric generator collects mechanical energy from the environment and transforms it into electricity to supply to microelectronic devices. Thus, these alternative energy sources can reduce the consumption of batteries, thereby reducing environmental pollution. Piezoelectric materials can work in the bending, compression, and shear modes, which are named as d31, d33, and d15 modes, respectively. In this study, a piezo stack which worked in d31 mode has been designed and integrated into an energy harvesting pedal. A novel compliant amplifying mechanism has to be designed to amplify the input load so that the high-stiffness piezoelectric stack can achieve a large energy output at a lower input force. This compliant mechanism has been designed by the pseudo-rigid-body and topology optimization methods. The amplification ratios of different sized flexible amplification mechanisms are calculated through the finite element analysis and validated by experiments. Finally, a pedal generator has been made and the test results show that the collected electricity can effectively drive a low-power microcontroller, sensor, and other devices of these kinds.
Low-Power FPGA Implementation of Convolution Neural Network Accelerator for Pulse Waveform Classification
