Design of optimum Nyquist signals based on generalized sampling theory for data communications

A novel algorithm for the optimal design of bandlimited Nyquist-type signal shapes for data communications has been developed. The proposed approach maximizes its energy in a given time interval and is based on the periodically nonuniform sampling theory that utilizes the linear splines. The computation is simple and the constraint for intersymbol interference is easy to include in the problem. It is also shown that the optimal signal shapes are almost shielded from small offsets at the sampling instants.