Determination of pulse locations by a multipulse coder is formulated as the minimisation of an error function in P-dimensional discrete space. A suboptimum solution is presented which offers better performance than conventional designs.
Read full abstract