References

Alternating projection onto convex sets is powerful tool for signal and image restoration. The extensions of von Neumann’s [3] alternating projection method by Dykstra and Han [1, 2] permit the computation of proximity projection onto certain convex sets. This paper exploits this fact in constructing a globally con-vergent method for computing the closest positive definite symmetric Toeplitz matrix to a specified matrix. Some applications to signal processing and control problems are discussed. Comparative numerical results are also reported