1: \begin{abstract}

2: In this paper sequential monitoring schemes to detect nonparametric drifts are

3: studied for the random walk case. The procedure is based on a kernel smoother.

4: As a by-product we obtain the asymptotics of the Nadaraya-Watson estimator and

5: its associated sequential partial sum process under non-standard sampling.

6: The asymptotic behavior differs substantially from the stationary situation,

7: if there is a unit root (random walk component). To obtain meaningful asymptotic

8: results we consider local nonparametric alternatives for the drift component.

9: It turns out that the rate of convergence at which the drift vanishes determines

10: whether the asymptotic properties of the monitoring procedure are determined by

11: a deterministic or random function. Further, we provide a theoretical result about 

12: the optimal kernel for a given alternative.

13: \\[1cm]

14: {\bf Keywords:} Control chart, nonparametric smoothing, sequential analysis, 

15:   unit roots, weighted partial sum process.

16: \end{abstract}

17: