Adaptive algorithms for target tracking.

The proposed control law only requires the relative positions of the vehicles with respect to the target and the velocity of the time-varying target both expressed in their

R´ ESUM´ E: Dans ce travail sur le d´ebruitage d’image, nous pr´esentons un nouveau type d’estimateur par patchs inspir´e de la m´ethode NL-Means propos´ee par Buades, Coll

A classical approach to ground robot path planning is the construction of a traversability map [Singh et al., 2000] [Gennery, 1999] followed by a graph search method to find the

Sequential Monte Carlo Methods : particule filters Particle methods are based on the principle of Monte Carlo to estimate and predict online dynamic systems.. They enable to

This contextual appearance information is extracted from moving image regions and used to train on-line a discriminative binary classifier that, in each video frame, detects the

We present a quadratic program whose globally optimal solution yields an optimal tracking joint policy, one that maximizes the expected targets detected over the given

Channel representations can be considered as regularly sampled kernel den- sity estimators [12], implying that channel-based tracking is related to kernel- based prediction for

insects form a “modality-independent internal representations of object shapes”, an