Stochastic approximation with correlated data

New almost sure convergence results for a special form of the multidimensional Robbins-Monro stochastic approximation procedure are developed. The results are applicable to cases where the "training data" is heavily correlated. No conditional expectation properties or boundedness assumptions are required to apply the new results. For example, when the data sequence is normal and i) M -dependent, ii) autoregressive moving average, or iii) "band-limited", the results can be used to establish the almost sure convergence of each algorithm treated. The special form of the Robbins-Monro procedure considered is motivated by a consideration of several algorithms that have been proposed for discrete-time adaptive signal-processing applications. Most of these algorithms can also be viewed as stochastic gradient-following algorithm. The ease with which the new results can be applied is illustrated.

[1]  D. Sakrison.  A Continuous Kiefer-Wolfowitz Procedure for Random Processes , 1964 .

[2]  M. T. Wasan.  Stochastic Approximation , 1969 .

[3]  Mischa Schwartz,et al.  A Rapidly Converging First-Order Training Algorithm for an Adaptive Equalizer , 1971, IEEE Trans. Inf. Theory.

[4]  Alan Weiss,et al.  Digital adaptive filters: Conditions for convergence, rates of convergence, effects of noise and errors arising from the implementation , 1979, IEEE Trans. Inf. Theory.

[5]  THOMAS P. DANIELL.  Adaptive Estimation with Mutually Correlated Training Sequences , 1970, IEEE Trans. Syst. Sci. Cybern..

[6]  B. Widrow,et al.  Adaptive antenna systems , 1967 .

[7]  D. Farden,et al.  Tracking properties of adaptive signal processing algorithms , 1980, ICASSP.

[8]  L. J. Griffiths,et al.  A simple adaptive algorithm for real-time processing in antenna arrays , 1969 .

[9]  O. L. Frost,et al.  An algorithm for linearly constrained adaptive array processing , 1972 .

[10]  C. Niessen,et al.  Adaptive Equalizer for Pulse Transmission , 1970 .

[11]  M. Sondhi,et al.  New results on the performance of a well-known class of adaptive filters , 1976, Proceedings of the IEEE.

[12]  Richard T. Lacoss,et al.  Adaptive Combining of Wideband Array Data for Optimal Reception , 1968 .

[13]  J. Kiefer,et al.  Stochastic Estimation of the Maximum of a Regression Function , 1952 .

[14]  Jae-Kyoon Kim,et al.  Adaptive linear estimation for stationary M-dependent processes , 1975, IEEE Trans. Inf. Theory.

[15]  Louis L. Scharf,et al.  On stochastic approximation and the hierarchy of adaptive array algorithms , 1972, CDC 1972.

[16]  D. Sakrison.  Stochastic Approximation: A Recursive Method for Solving Regression Problems1 , 1966 .

[17]  A. Gersho.  Adaptive equalization of highly dispersive channels for data transmission , 1969 .

[18]  Khalid Sayood,et al.  On the "Desired behavior" of adaptive signal processing algorithms , 1979, ICASSP.

[19]  R. Serfling,et al.  Convergence Properties of $S_n$ Under Moment Restrictions , 1970 .

[20]  Robert Serfling.  Moment Inequalities for the Maximum Cumulative Sum , 1970 .

[21]  Lennart Ljung,et al.  Analysis of recursive stochastic algorithms , 1977 .