On redefining the optimal least squares filter under floating point operations | IEEE Conference Publication | IEEE Xplore