US 6,490,327 B1 · System and method for self-adaptive maximum likelihood sequence detection using a T-algorithm

Abstract

A telecommunications system and method is disclosed for employing an improved self-adaptive maximum likelihood sequence detection technique using a T-algorithm and its variants in order to reduce signal distortion in cellular networks. If the noise variance is known, the T-algorithm computes the metric for all the branches in the stack and computes an expected value of the desired metric. Each branch not within a small, predefined deviation from the desired metric is discarded. However, if the noise variance cannot be reliably estimated, the T-algorithm used is similar to the previously described T-algorithm, except that the comparison is with the overall best path with the minimum metric. In other words, the path with the minimum value of the metric is computed and all paths within the small, predetermined deviation of this metric are retained and the remaining are deleted from the stack.

Claims17

Dependent claims

A method for detecting a maximum likelihood sequence of symbols within a transmitted input data stream, said symbols being subject to intersymbol interference during transmission, said method comprising the steps of:

constructing an enumeration tree of nodes for a given sequence of said input data stream containing said symbols, said enumeration tree having at least a root one of said nodes and a given number of end ones of said nodes grown from said root node, each said end node being included within a first stack having a size equivalent to said given number;
modifying said enumeration tree from said end nodes to include additional ones of said nodes, each said end node having at least two of said additional nodes branching therefrom;
computing a desired metric value of a metric;
computing, at each said respective additional node, a respective metric value of said metric associated with said given sequence of said symbols;
and retaining select ones of said additional nodes having said respective metric value within a predetermined deviation from said desired metric value.

2 more elements

A telecommunications system, comprising:

a transmitter for transmitting a data stream of discrete symbols;
a receiver, in communication with said transmitter, for receiving said data stream from said transmitter, said symbols being subject to intersymbol interference during transmission from said transmitter;
and
a maximum likelihood sequence detector, attached to said receiver, for detecting a maximum likelihood sequence of said symbols for a given sequence of said received data stream, said maximum likelihood sequence detector comprising:
means for constructing an enumeration tree of nodes for said given sequence of said input data stream, said enumeration tree having at least a root one of said nodes and a given number of end ones of said nodes grown from said root node, each said end node being included within a first stack having a size equivalent to said given number,
means for modifying said enumeration tree from said end nodes to include additional ones of said nodes, each said end node having at least two of said additional nodes branching therefrom,
means for computing a desired metric value of a metric, and at each said respective additional node, a respective metric value of said metric associated with said given sequence of said symbols, and
means for retaining select ones of said additional nodes having said respective metric value within a predetermined deviation from said desired metric value.

5 more elements

Dependencies are inferred from the claim text.

System and method for self-adaptive maximum likelihood sequence detection using a T-algorithm

Life of this patent

System and method for self-adaptive maximum likelihood sequence detection using a T-algorithm

Life of this patent

Abstract

Claims17