Single-ended loop-makeup Identification-part II: improved algorithms and performance results

Digital subscriber lines (DSLs) transmit over twisted-pair telephone loops, and knowledge of the loop makeup is vital for determining what types of DSLs may be deployed and at what bit rates they may transmit. This paper expands and applies the algorithm presented in Part I of this study for identifying a loop makeup, including cable gauges, lengths, and bridged taps, via single-ended measurements based on the use of an enhanced time domain reflectometer. Techniques based on the exploitation of the information of the sign of the echo for reducing identification time are described. Loop identification is further improved by retaining multiple hypothesized loop topologies, as the loop estimate is built out constructing multiple paths in a tree search via branch-and-bound techniques. Furthermore, statistical data is used to improve overall loop identification results by accounting for the different likelihood of different cable gauges at different distances from a central office. Test loops derived from a loop database are described, and results of automated loop-makeup identification on these loops indicate the accuracy of the loop identification as a function of the percentage of loops at a wire center.