作者: Andrey Velichko , Maksim Belyaev , Vadim Putrolaynen , Valentin Perminov , Alexander Pergament
DOI: 10.1016/J.SSE.2017.12.003
关键词:
摘要: Abstract We explore a prototype of an oscillatory neural network (ONN) based on vanadium dioxide switching devices. The model system under study represents two oscillators thermally coupled VO2 switches. Numerical simulation shows that the effective action radius RTC coupling depends both total energy released during and average power. It is experimentally numerically proved temperature change ΔT commences almost synchronously with power peak T-coupling reveals itself up to frequency about 10 kHz. For studied structure configuration, value varies over wide range from 4 45 μm, depending external circuit capacitance C resistance Ri, but variation Ri more promising practical viewpoint. In case “weak” coupling, synchronization accompanied by attraction effect decrease main spectra harmonics width. “strong” number effects increases, can occur subharmonics resulting in multilevel stable oscillators. An advanced algorithm for efficiency subharmonic ratio calculation proposed. shown leading one higher frequency, and, addition, stabilization observed. Also, strong thermal limit supply current parameters, which oscillations exist, expands ∼10%. obtained results have universal character open new kind ONNs, namely, T-coupling, allows easy transition 2D 3D integration. hold promise application classification pattern recognition.