The conventional change-point detection problem aims to detect distribution changes at some unknown time point in a sequence of multivariate observations. Such problem is hardly addressed when the data are functional and both the pre-change and post-change distributions are unknown. In this paper, we propose an online sequential procedure based on a Generalized Likelihood Ratio (GLR) testing to address these issues. This procedure aims to minimize the expected detection delay subject to a false alarm constraint, and is designed to detect multiple change-points in a stream of multivariate curves. The methodology relies upon a specific multivariate regression model that takes into account prior information about the curve segmentation. This generative model can be fitted using a dedicated Expectation-Maximization (EM) algorithm presented in a semi-supervised framework. The monitoring strategy is applied to a sequence of real data collected from a door system operating in a transit bus. The experimental results allow to highlight the effectiveness of the proposed approach. Transit buses door system, Sequential Testing, Change point detection