This paper presents an educational experience carried out in robotics undergraduate courses from two different degrees: Computer Science and Industrial Engineering, having students with diverse capabilities and motivations. The experience compares two learning strategies for the practical lessons of such courses: one relies on code snippets in Matlab to cope with typical robotic problems like robot motion, localization, and mapping, while the second strategy opts for using the ROS framework for the development of algorithms facing a competitive challenge, e.g. exploration algorithms. The obtained students’ opinions were instructive, reporting, for example, that although they consider harder to master ROS when compared to Matlab, it might be more useful in their (robotic related) professional careers, which enhanced their disposition to study it. They also considered that the challenge-exercises, in addition to motivate them, helped to develop their skills as engineers to a greater extent than the skeleton-code based ones. These and other conclusions will be useful in posterior courses to boost the interest and motivation of the students.

The suitable operation of mobile robots when providing Ambient Assisted Living (AAL) services calls for robust object recognition capabilities. Probabilistic Graphical Models (PGMs) have become the de-facto choice in recognition systems aiming to efficiently exploit contextual relations among objects, also dealing with the uncertainty inherent to the robot workspace. However, these models can perform in an incoherent way when operating in a long-term fashion out of the laboratory, e.g. while recognizing objects in peculiar configurations or belonging to new types. In this work we propose a recognition system that resorts to PGMs and common-sense knowledge, represented in the form of an ontology, to detect those inconsistencies and learn from them. The utilization of the ontology carries additional advantages, e.g. the possibility to verbalize the robot's knowledge. A primary demonstration of the system capabilities has been carried out with very promising results.

In general, the problems of objects' and rooms' categorizations for robotic applications have been addressed separately. The current trend is, however, towards a joint modelling of both issues in order to leverage their mutual contextual relations: object → room (e.g. the detection of a microwave indicates that the room is likely to be a kitchen), and room → object (e.g. if the robot is in a bathroom, it is probable to find a toilet). Probabilistic Graphical Models (PGMs) are typically employed to conveniently cope with such relations, relying on inference processes to hypothesize about objects' and rooms' categories. In this work we present a Conditional Random Field (CRF) model, a particular type of PGM, to jointly categorize objects and rooms from RGBD images exploiting object-object and object-room relations. The learning phase of the proposed CRF uses Human Knowledge (HK) to eliminate the necessity of gathering real training data. Concretely, HK is acquired through elicitation and codified into an ontology, which is exploited to effortless generate an arbitrary number of representative synthetic samples for training. The performance of the proposed CRF model has been assessed using the NYU2 dataset, achieving a success of ~ 70% categorizing both, objects and rooms.

This chapter presents a telehealth system called GiraffPlus supporting independent living of elderly in their own home. GiraffPlus system is a complex system which monitors activities and physiological parameters in the home using a network of sensors. The elaborated information is presented to the primary user, the elderly, and to secondary users like health care and home care providers and possibly to family members as a help to assess possible health and wellbeing deterioration, provide acute alarms, and support health procedure. The secondary users can also visit the elderly via the Giraff, a teleoperated robot that can communicate and move in the home under the control of the secondary user. The chapter focusses in particular on the deployment of the system in six real homes in Sweden, Italy and Spain. The chapter outlines the technological various components used, the expectations of the users and the evaluation method.

Despite Kinect has been quickly integrated into a variety of robotic systems and applications in substitution to traditional sensors like range scanner lasers, not too much attention has been paid to its characterization for typical mobile robots’ environments. This characterization is basic to understand how the sensor behaviour depends on the nature of the environment, which allows the robot to tackle possible sensor errors due to such a nature. In this report we present an experimental study of the performance of the Kinect range camera, which differs from other studies in addressing a thorough evaluation for assessing the sensor performance and reliability under diverse environmental conditions and factors, including radiometry and reflectivity of objects, lighting conditions, repeatability, interference with other sensors, and coverage. The results and the conclusions of the experiences presented here provide useful insights about the performance of the Kinect RGB-D camera that help to understand and exploit the characteristics of this device.

This paper presents H, a minimalistic specification language for designing heterogeneous software applications, particularly in the realms of robotics and industria, which takes advantage of a Component-Based Software Engineering (CBSE) approach. H copes with some of the most outstanding characteristics of these systems, like diversity at different levels (hardware platforms, programming languages, programmer skills), network distribution, real time and fault-tolerance. The H specification covers the life-cycle of any heterogeneous application. Its development system offers to the designer and/or builder a set of tools for specifying modules, generating code semiautomatically, debugging, maintenance, and a real time analysis of the system.

This paper deals with the use of semantic knowledge to improve the intelligence and autonomous behavior of a mobile robot. A robot can exploit the semantics of its environment to infer new, implicit information. Another interesting possibility is to use semantics for detecting deviations between the real world and what is supposed to be “normal”. For instance, normative semantic knowledge may state that towels should stay in the bathroom. If a robot detects a towel in the kitchen, it can react and decide to solve this inconsistency by bringing it to the bathroom. However not all ways to solve an inconsistency are acceptable: for instance, if the robot put the towel temporarily on a dirty sink in order to re-grasp it with the other arm, it would violate another norm – namely, that towels should always stay on a clean surface. In this work we present an algorithm that detects and recovers from norm violations, according to a semantic representation of norms, and ensures the normative acceptability of the robot actions throughout execution.

LEGO Mindstorms NXT robots are being increasingly used in undergraduate courses, mostly in robotics-related subjects. But other engineering topics, like the ones found in data acquisition, control and real-time subjects, also have difficult concepts that can be well understood only with good lab exercises. Such exercises require physical educational tools that should be low cost, easy to configure and use, multi-purpose and motivational for the students, being all of this hard to achieve with a single device. The ''classical'' solution has been to acquire specific commercial kits for each subject, or even topic, usually proprietary and expensive. Our work extends the already existing alternative of using the LEGO Mindstorms NXT robots as a training platform, but not by imitating the same approach of commercial kits (e.g., to isolate some part of the robot for teaching a particular topic); we rather aim at accomplishing all the mentioned requirements simultaneously. For that purpose, we have used only one out-of-the-box, complete robot configuration, to be shared among different subjects without hardware/software/firmware modifications. This has reduced significantly the effort of a group of professors when preparing exercises, and encouraged the reuse of their work among several topics and subjects. Also, we have collected a number of surveys on students and the professors' experiences. In this paper we describe our approach and present in detail the results, which assess the higher motivational adequacy of using a complete robot in these subjects and also the real fulfillment of the other requirements along several academic years.

Small computers used for assisting drivers (mostly in finding routes) have been growing in popularity in the past years. These systems are inherently interactive, but up to now this interaction is tackled under rather simple approaches. For example, current routing computer assistants consider only the shortest or the quickest route to a destination, although in certain situations it could be interesting for the driver to take into consideration other factors, such as the criminal rate, land value, or the beauty of the areas to be traversed. On the other hand, the interactive processes between the driver and the routing assistant are still very limited: They only enable the user to discard (or suggest) particular locations through a fixed set of names, i.e. street's names. This paper proposes a novel interactive mechanism for in-vehicle routing that uses topological information at different levels of detail and a multihierarchical representation of urban maps. These hierarchical representations permit the system not only to plan routes efficiently but also to report them at different levels in detail in a human-like set of symbols adapted to each user. This enhances the human–computer interaction during the routing process, increasing driver satisfaction. We illustrate our technique through a case of study in the city of Málaga (Spain). © 2010 Wiley Periodicals, Inc.

This paper presents a procedure to robustly distribute control point (CP) pairs in high-resolution satellite images as a preliminary step for accurate image registration. The proper distribution of the CPs is achieved by means of a quadtree decomposition of a coarse digital terrain model (DTM) of the sensed region. This technique parcels up the image according to its relief variance yielding almost planar pieces of land. A corner detector is then employed to identify key points in the reference image and an affinity-based feature tracker that searches for their corresponding corner in the target one. This search is executed in every parcel, selecting (at-least) one CP, ensuring thus denser distributions in rugged regions than in flat ones. Additionally, robustness to mismatches is attained by exploiting the intrinsic affine epipolar geometry of the two images. The proposed method has been successfully tested with a broad variety of panchromatic high-resolution images of the city of the Rincon de la Victoria (Malaga, Spain).

This article addresses the problem of mobile robot localization using Ultra-Wide-Band (UWB) range measurements. UWB is a radio technology widely used for communications, that is recently receiving increasing attention for positioning applications. In these cases, the position of a mobile transceiver is determined from the distances to a set of fixed, well-localized beacons. Though this is a well-known problem in the scientific literature (the trilateration problem), the peculiarities of UWB range measurements (basically, distance errors and multipath effects) demand a different treatment to other similar solutions, as for example, those based on laser. This work presents a thorough experimental characterization of UWB ranges within a variety of environments and situations. From these experiments, we derive a probabilistic model which is then used by a particle filter to combine different readings from UWB beacons as well as the vehicle odometry. To account for the possible offset error due to multipath effects, the state tracked by the particle filter includes the offset of each beacon in addition to the planar robot pose (x,y,ϕ), both estimated sequentially. We show navigation results for a robot moving in indoor scenarios covered by three UWB beacons that validate our proposal.

Image registration is the process of spatially fitting two images of the same scene acquired on different dates, from different viewpoints, and/or using different sensors. Image registration is required in a variety of applications, such as, in image fusion, 3D scene reconstruction, and multi-temporal analysis (i. e. natural disaster monitoring, urban change detection, etc.). See [1] for a comprehensive survey. A typical image registration process is accomplished by identifying so-called control points (CP) pairs, in the involved images. Through such control points it is possible to estimate the underlying geometrical transformations between the considered images. Accuracy in image registration is strongly tied to: a) the geometrical transformation considered, which should account for the relative geometric (possibly non-rigid) differences between the images, and b) the distribution of CPs over the images. The correct selection of the latter is primordial not only to gain in accuracy but also in efficiency: while two pair of CPs would suffice to perfectly overlap images of a flat terrain (since they may only differ in shift, scale and rotation), a large number of them will be necessary to capture the relative geometric difference between images of high-relief surfaces acquired from different viewing angles, requiring, thus, complex elastic transformations. In the absence of information about the type of terrain, the best (thought not efficient) solution is to regularly distribute CPs all over the images. However, when some information about the terrain profile is available, a more elaborated algorithm can help us to decide the density of CPs on each region of the image. In this paper we present an automatic method to extract control points for accurate registration of high-resolution images, such as Quickbird, Ikonos, etc. Our approach generates a minimal distribution of control points based on the relief information provided by digital terrain models (DTM), achieving significant speedup in the process without sacrificing accuracy.

This paper describes SANCHO, a mobile robot intended to perform within crowded areas as a servant, for instance as a fair or congress host. This robot has been constructed upon a commercial platform on which a number of sensors and devices have been integrated. A software control architecture has been implemented and adapted to this particular robot, enabling it to perform in human scenarios. Among the different subsystems of the control architecture developed for SANCHO, we highlight in this paper two of the most relevant ones: the navigation component which permits the robot to navigate in a safe and robust manner within crowded and dynamic environments, and the communication component which provides different possibilities for human-robot interaction. We illustrate the performance of SANCHO through a number of experiences carried out in public shows.

El número de personas con problemas de movilidad está creciendo en los últimos años en los países desarrollados, como consecuencia de una mayor esperanza de vida y del gran número de accidentes de tráfico. La posibilidad de transferir y adaptar resultados del campo de la robótica a este colectivo resultaría de enorme interés toda vez que repercuta en una mejor calidad de vida y una mayor integración social. En este artículo presentamos la silla de ruedas robotizada SENA, que es el resultado de un proyecto de investigación desarrollado en el Dpto. de Ingeniería de Sistemas y Automática de la Universidad de Málaga (España). SENA se basa en una silla eléctrica comercial a la que se ha integrado una serie de sensores y dispositivos que son gestionados mediante un ordenador portátil. Este prototipo se caracteriza por su capacidad para la navegación autónoma en espacios interiores, así como por su facilidad para interactuar y cooperar con el usuario o personas del entorno gracias, entre otros, a una arquitectura software desarrollada específicamente para robots asistentes. En este artículo se describen los principales elementos hardware y software de SENA, se ilustra su funcionamiento y se exponen algunas de las reflexiones y conclusiones derivadas del desarrollo de este prototipo.