On redefining the optimal least squares filter under floating point operations

A novel solution on approximation to the least squares filter problem under floating point arithmetic is presented for a linear stochastic model.