For underwater robotics applications involving monitoring and inspection tasks, it is important to capture quality color images in real time. In this paper, we propose a statistically learning method with an automatic selection of the training set for restoring the color of underwater images. Our statistical model is a Markov Random Field with Belief Propagation (MRF-BP). The quality of the results depends strongly on the trained correlations between the degraded image and its corresponding color image. However, it is not possible to have color ground truth data given the inherent conditions of underwater environments. Thus, we build a color adaptive training set by applying a multiple color space analysis to those frames that present a high change in its distribution from the previous frame and use only those frames for training. Experimental results in real underwater video sequences demonstrate that our approach is feasible, even when visibility conditions are poor, as our method can recover and discriminate between different colors in objects that may seem similar to the human eye.

An omnidirectional mobile manipulator (OMM) provides extended capabilities to a robotic system since it enlarges its workspace at the expense of an omnidirectional mobile base, which is nowadays a reality. When various OMM transport an object, the resultant constrained OMM dynamics is quite complex that demands effective control schemes without neglecting dominant nonlinear effects nor interaction forces. In this paper, it is proposed a decentralized cooperative control for a set of k OMMs (k-OMM) subject to homogeneous holonomic constraints imposed by the object being manipulated. A chatterless integral sliding mode force-position control guarantees exponential tracking whose implementation does not require any knowledge of dynamics. This controller is decentralized in a sense that neither dynamics nor interaction forces of the i-OMM are required for the j-controller, and it is cooperative by correcting collectively the velocity deviation due to the pushing of the other j-OMM. Each omnidirectional base provides object mobility along passive smooth velocity fields to avoid obstacles on the floor, which is synthesized as the secondary task when resolving redundancy, and the force control stabilizes internal forces to ensure object grasp, which are declared together with end-effector position, as the primary task. Simulations are discussed for two 8-DoF Kuka youBot OMM system navigating in a cluttered environment.