L. Molina-Tanco

Pyramid segmentation algorithms revisited

The main goal of this work is to compare pyramidal structures proposed to solve segmentation tasks. Segmentation algorithms based on regular and irregular pyramids are described, together with the data structures and decimation procedures which encode and manage the information in the pyramid. In order to compare the different segmentation algorithms, we have employed three types of quality measurements: the shift variance measure, the F function and the Q function.

A SURVEY OF VISION-BASED ARCHITECTURES FOR ROBOT LEARNING BY IMITATION

Learning by imitation is a natural and intuitive way to teach social robots new behaviors. While these learning systems can use different sensory inputs, vision is often their main or even their only source of input data. However, while many vision-based robot learning by imitation (RLbI) architectures have been proposed in the last decade, they may be difficult to compare due to the absence of a common, structured description. The first contribution of this survey is the definition of a set of standard components that can be used to describe any RLbI architecture. Once these components have been defined, the second contribution of the survey is an analysis of how different vision-based architectures implement and connect them. This bottom–up, structural analysis of architectures allows to compare different solutions, highlighting their main advantages and drawbacks, from a more flexible perspective than the comparison of monolithic systems.

Fast gesture recognition based on a two-level representation

Towards developing an interface for human-robot interaction, this paper proposes a two-level approach to recognise gestures which are composed of trajectories followed by different body parts. In a first level, individual trajectories are described by a set of key-points. These points are chosen as the corners of the curvature function associated to the trajectory, which will be estimated using and adaptive, non-iterative scheme. This adaptive representation allows removing noise while preserving detail in curvature at different scales. In a second level, gestures are characterised through global properties of the trajectories that compose them. Gesture recognition is performed using a confidence value that integrates both levels. Experimental results show that the performance of the proposed method is high in terms of computational cost and memory consumption, and gesture recognition ability.

Real-time human motion analysis for human-robot interaction

This paper introduces a novel real-time human motion analysis system based on hierarchical tracking and inverse kinematics. This work constitutes a first step towards our goal of implementing a mechanism of human-machine interaction that allows a robot to provide feedback to a teacher in an imitation learning framework. In particular, we have developed a computer-vision based, upper-body motion analysis system that works without special devices or markers. Since such system is unstable and can only acquire partial information because of self-occlusions and depth ambiguity, we have employed a model-based pose estimation method based on inverse kinematics. The resulting system can estimate upper-body human postures with limited perceptual cues, such as centroid coordinates and disparity of head and hands.

Real-time object tracking using bounded irregular pyramids

Target representation and localization is a central component in visual object tracking. In this paper a new approach for target representation and localization is presented. This approach tackles two of the most important causes of failure in object tracking: changes of object appearance and occlusions. We propose a modified template matching approach which does not require a priori learning of object views. This method allows to track non-rigid objects in real-time by employing a weighted template which is dynamically updated, and a hierarchical framework that integrates all the components of the tracker. Our hierarchical tracker allows tracking of multiple objects with low increase of computational time. The capability of the tracking system to handle occlusions and target distortions is demonstrated for several video sequences.

The construction of bounded irregular pyramids with a union-find decimation process

The Bounded Irregular Pyramid (BIP) is a mixture of regular and irregular pyramids whose goal is to combine their advantages. Thus, its data structure combines a regular decimation process with a union-find strategy to build the successive levels of the structure. The irregular part of the BIP allows to solve the main problems of regular structures: their inability to preserve connectivity or to represent elongated objects. On the other hand, the BIP is computationally efficient because its height is constrained by its regular part. In this paper the features of the Bounded Irregular Pyramid are discussed, presenting a comparison with the main pyramids present in the literature when applied to a colour segmentation task.

Robot Learning by Active Imitation

A key area of robotics research is concerned with developing social robots for assisting humans in everyday tasks. Many of the motion skills required by the robot to perform such tasks can be pre-programmed. However, it is normally agreed that a truly useful robotic companion should be equipped with some learning capabilities, in order to adapt to unknown environments, or, what is most difficult, learn to perform new tasks. Many learning algorithms have been proposed for robotics applications. However, these learning algorithms are often task specific, and only work if the learning task is predefined in a delicate representation, and a set of pre-collected training samples is available. Besides, the distributions of training and test samples have to be identical and the world model is totally or partially given (Tan et al., 2005). In a human world, these conditions are commonly impossible to achieve. Therefore, these learning algorithms involve a process of optimization in a large search space in order to find the best behaviour fitting the observed samples, as well as some prior knowledge. If the task becomes more complicated or multiple tasks are involved, the search process is often incapable of satisfying real-time responses. Learning by observation and imitation constitute two important mechanisms for learning behaviours socially in humans and other animal species, e.g. dolphins, chimpanzees and other apes (Dautenhahn & Nehaniv, 2002). Inspired by nature, and in order to speed up the learning process in complex motor systems, Stefan Schaal appealed for a pragmatic view of imitation (Schaal, 1999) as a tool to improve the learning process. Current work has demonstrated that learning by observation and imitation is a powerful tool to acquire new abilities, which encourages social interaction and cultural transfer. It permits robots to quickly learn new skills and tasks from natural human instructions and few demonstrations (Alissandrakis et al., 2002, Breazeal et al., 2005, Demiris & Hayes, 2002, Sauser & Billard, 2005). In robotics, the ability to imitate relies upon the robot having many perceptual, cognitive and motor capabilities. The impressive advance of research and development in robotics over the past few years has led to the development of this type of robots, e.g. Sarcos (Ijspeert et al., 2002) or Kenta (Inaba et al., 2003). However, even if a robot has the necessary skills to imitate the human movement, most published work focus on specific components of an imitation system (Lopes & Santos-Victor, 2005). The development of a complete imitation architecture is difficult. Some of the main challenges are: how to identify which features of an action are important; how to reproduce such action; and how to evaluate the performance of the imitation process (Breazeal & Scassellati, 2002).

Robot learning of upper-body human motion by active imitation

This paper presents a general architecture that allows a humanoid robot to imitate upper-body movements of a human demonstrator. This architecture integrates a mechanism to memorize novel behaviours executed by a human demonstrator, with a module to recognize and generate its own interpretation of already observed behaviours. Our imitator includes three biologically plausible components: i) an attention mechanism to autonomously extract relevant information from the visual input; ii) a supra-modal representation of the motion of observed body parts to map visual and motor domains; and iii) an active imitation module which involves the motor systems in the behaviour recognition process. Experimental results with a real humanoid robot demonstrate the ability of the proposed architecture to acquire novel behaviours and to recognize and reproduce previously memorized ones

Online absorption of mediolateral balance disturbances for a small humanoid robot using accelerometer and force-sensor feedback

Sociable humanoid robots should be able to coexist with humans in human environments, where they will very likely have to face unexpected disturbance forces. In this paper, an online ankle-hip control strategy is presented that allows a standing biped robot to withstand a force in the mediolateral direction, using foot force-sensor and accelerometer feedback. The control algorithm is simple and can be implemented on any biped robot equipped with these sensors. Experimental results are presented for a small humanoid robot Fujitsu HOAP-1 showing that our method allows the robot to bear considerable perturbation forces.

A model-based humanoid perception system for real-time human motion imitation

This paper presents real-time human motion analysis based on hierarchical tracking and inverse kinematics. Our goal is to implement a mechanism of human-machine interaction that permits a robot to learn from human gestures, and, as a first stage, we have developed a computer-vision based human upperbody motion analysis system. This application requires developing a real-time human motion capturing system that works without special devices or markers. Since such a system is unstable and can only acquire, partial information because of self-occlusions, we have introduced a pose estimation method based on inverse kinematics. This system can estimate upper-body human postures with limited perceptual cues such as position of head and hands. The method has been tested using a HOAP-I humanoid robot.