Searching the optimal sampling instant in a TDMA packet transmission system

A device for searching the optimal sampling instant intended to retrieve a sample, referred to as optimal sample, from input samples and comprising an amplitude-computing member for computing the average amplitudes of the input samples, wherein the computing member uses only one part of the input samples for determining the optimal sample having the maximum average amplitude among all the input samples, the device including a device for taking predetermined input samples to be supplied to the computing member with a view to determining the sample having the maximum average amplitude, referred to as approximate maximum, a selector for selecting a supplementary input sample comprised between the approximate maximum and the neighboring sample whose previously computed average amplitude is the largest and for supplying said sample to the computing member with a view to computing its average amplitude, and a comparator for comparing the average amplitude of the selected sample with the previous approximate maximum and for deriving a new approximate maximum, wherein the supplementary sample is situated midway between the last approximate maximum and the neighboring sample whose previously computed average amplitude is the largest, and wherein the number of input samples is an integral multiple L of an integer N, with N=2n, and the predetermined number of samples among the N input samples is an integer L×M, with M=2m and 2<M<N, and the optimal sample is obtained at the end of a number of average amplitude computations which is equal to M+n−m, where n and m are integers.

A method of recovering an optimal sampling instant from input samples, the method comprising the steps of:

• computing average amplitudes of a predetermined number input samples;
• determining a maximum of the average amplitudes of these samples;
• deriving an approximation of the optimal sample using the maximum,
• repeating the average amplitude computation for input samples between the approximation and a neighboring sample whose previously computed average amplitude is the largest, and
• comparing of the result of each average amplitude computation with a previously computed maximum corresponding to the previous approximation for deriving a new approximation of the optimal sample.