A Self-organizing Learning Array and Its Hardware-software Co-simulation

− In this paper, a self-organizing learning array (SOLAR) and its hardware-software (HW/SW) co-simulation are presented. In SOLAR, every neuron maximizes its information index during feed forward self-organizing learning. SOLAR was simulated on benchmark examples and showed good ability to learn exceeding the performance of many traditional classifier algorithms. A simple HW/SW co-simulation method was adopted to avoid the use of two simulators and complex inter-process communication. In this co-simulation method, software is modelled using behavioral, and hardware using structural hardware description language (VHDL) models. Their interface is also modelled in VHDL.