Organic Polymer-Based Photodiodes for Optoelectronic\nReservoir Computing with Time-Based Coding

The\nintegration of optoelectronic devices with reservoir computing\noffers a novel and effective approach to in-sensor computing. This\nwork presents a hybrid digital–physical solution that leverages\nthe high-performance poly[(bithiophene)-alternate-(2,5-di(2-octyldodecyl)-3,6-di(thienyl)-pyrrolyl\npyrrolidone)] (DPPT-TT) organic polymer-based photodiodes for the\nhardware implementation of reservoir computing system. The photodiodes\ndemonstrate nonlinear photoelectric responses, fading memory, and\ncyclical stability, in relation to the temporal information on light\nstimuli. These attributes enable effective mapping, historical context\nsensitivity, and consistent performance, with time-encoded inputs,\nwhich are particularly essential for accurate and continuous processing\nof time series data. The optoelectronic reservoir computing system\nwith pulse width modulation (PWM) coding demonstrates impressive performance\nin the prediction of chaotic sequences, achieving a normalized root-mean-square\nerror as low as 0.095 with optimized parameters. The DPPT-TT-based\nphotodiodes and time-based coding offer a hardware-efficient solution\nfor reservoir computing, significantly advancing Internet of Things\napplications.