Small-footprint mobile ground robots, such as the popular Turtlebot and Kobuki platforms, are by necessity equipped with sensors which lie close to ground. Reliably detecting tracking people from this viewpoint is a challenging problem, whose solution key requirement for many applications involving sharing of common spaces human-robot interaction. We present robust cluttered indoor environments, using an inexpensive RGB-D sensor Microsoft Kinect or Asus Xtion. Even in scenarios multiple view...

Once a multi-robot system is implemented on real hardware and tested in the world, analyzing its evolution debugging unexpected behaviors often very difficult task. We present tool for aiding this activity, by visualizing an Augmented Reality overlay live video feed acquired fixed camera overlooking robot environment. Such displays information exposed each robot, which may be textual (state messages), symbolic (graphs, charts), or, most importantly, spatially-situated; spatially-situated...

In this paper, we consider the problem of localization in a multi-robot system. We present new approach focused on distribution, scalability, and minimum-uncertainty perception. An Extended Kalman Filter (EKF) is used to update an estimate robot poses correspondence each sensor measurement. entropic criterion used, order select optimal measurements that reduce global uncertainty relative poses. It shown that, addition EKF, also selection measurement can be distributed among robots, scalable...

We present a novel calibration method for off-axis catadioptric cameras, i.e. standard perspective cameras placed in generic position w.r.t. an axial-symmetric mirror of unknown shape. The proposed estimates the intrinsic parameters natural camera, 3D shape and its pose camera. peculiarity our approach is that, unlike several other methods, we do not require any cross section to be visible image. Instead, that image contains at least one space line. then derive some constraints combined with...

This paper presents the main results obtained in Agencies for PErception environmental monitoring (APE) Project, funded by Italian Ministry of University and Research, aimed at studying employment advanced distributed artificial intelligence techniques developing cooperative multiagent systems oriented to support applications. In particular, this describes issues about communication between perceptive agents two implemented electromagnetic fields pollutant evaluation.

The reconstruction of 3D scenes constituted by straight lines can find many applications both in computer vision and mobile robotics. Most the approaches to this problem involve either stereo-vision or analysis a sequence images taken from different viewpoints: cases, solution correspondence is required. This paper studies localization space single ID images, acquired catadioptric camera. In general, using noncentral camera, viewing rays starting points line constitute non-planar surface: if...

The availability of efficient mapping systems to produce accurate representations initially unknown environments is undoubtedly one the main requirements for autonomous mobile robots. This paper presents a system that has been implemented on robot equipped with laser range scanner. builds geometrical maps environment employing an exploration strategy takes into account both distance travelled and information gathered determining observation positions. based stronger mathematical foundations...

We consider the problem of localizing a moving ball from single calibrated perspective image; after showing that ordinary algorithms fail in analyzing motion blurred scenes, we describe theoretically-sound model for image ball. Then, present an algorithm capable recovering both 3D position and its velocity. The is experimentally validated on real synthetic images.

The self-localization of a mobile robot within known environment, by means an orientable range finder, is considered. problem the determination sensor orientation which minimizes position uncertainty addressed. An efficient technique proposed to determine optimal exploration, given current estimate and its uncertainty. Once tentative exploration given, avoids take any worst into account, allowing efficiently one. Both location accuracy efficiency have been analyzed in paper. time needed plan...

A criterion is presented for the automatic selection of a sensor measurement aimed at observing state system which described both by discrete variables and continuous ones. The based on expected value entropy variation associated to observation. This then applied object recognition localization tasks, in observed characterized class, represented variable, pose, i.e., position orientation, vector parameters. proposed also accounts information obtained case missed measurement.

We consider the problem of localizing a moving ball from single calibrated perspective image; after showing that ordinary algorithms fail in analyzing motion blurred scenes, we describe theoretically-sound model for image ball. Then, present an algorithm capable recovering both 3D position and its velocity. The is experimentally validated on real synthetic images.

In-door environments often contain several straight line segments. The 3D reconstruction of such can thus reduce to the localization lines in space. Multi-view requires solution correspondence problem. use a single image localize space is attractive, since problem be avoided. However, using perspective camera (or central one), not localized, its viewing surface planar, and hence it infinite other than correct one. In this paper we study number planar surfaces for general class catadioptric...

Photographic images of moving objects are often characterized by motion blur; analyzing blurred is problematic since the object boundaries appear fuzzy and seamlessly blend with background.In extreme cases, when fast in relation to exposure time, image becomes an elongated, semitransparent smear.We consider a motion-blurred color over still background: we introduce meaningful entities, "alpha map" "color map", which bear information about during exposure, its texture; draw connections...

Line localization from a single image of central camera is an ill-posed problem unless other constraints or apriori knowledge are exploited. Recently, it has been proved that noncentral catadioptric cameras allow space lines to be localized image. In this paper we propose two novel algorithms. The first method exploits pair coplanar viewing rays localize the line. second follows constrained non-linear minimization procedure using suitable parametrization represent lines. We compare accuracy...

Sensor networks are distributed systems that detect phenomena to produce detailed environmental assessments. The use of multirobot for carrying sensors around the environment represents a solution has recently received considerable attention. In this paper we present an architecture system performs tasks related perception; in particular, it is devoted monitoring electro-magnetic fields.

In this paper we propose a novel algorithm to estimate motion parameters from single blurred image, exploiting geometrical relations between image intensities at pixels of region that contains corner. Corners are significant both for scene and understanding since they permit univocal interpretation parameters. Motion estimated locally in regions, without assuming uniform blur on so the works also with produced by camera rotation and, more general, space variant blur.