Inverse modeling with SVMs-based dynamically reconfigurable systems

In this paper we discuss the performances of an FPGA-based architecture which solves an inverse modeling problem. The case study is a typical telecommunication task where the hardware implementation, on the receiver side, is of paramount importance. A reconfigurable device is used in order to build a complete system on chip implementing an efficient learning by examples technique.