A Comparative Study of Registration Methods for RGB-D Video of Static Scenes

The use of RGB-D sensors for mapping and recognition tasks in robotics or, in general, for virtual reconstruction has increased in recent years. The key aspect of these kinds of sensors is that they provide both depth and color information using the same device. In this paper, we present a comparative analysis of the most important methods used in the literature for the registration of subsequent RGB-D video frames in static scenarios. The analysis begins by explaining the characteristics of the registration problem, dividing it into two representative applications: scene modeling and object reconstruction. Then, a detailed experimentation is carried out to determine the behavior of the different methods depending on the application. For both applications, we used standard datasets and a new one built for object reconstruction.

Temporo-Parietal Junction encodes the Embodied Self: Neuro-Robotics, fMRI, and Lesion mapping.

Introduction: Neuroimaging of the self focused on high-level mechanisms such as language, memory or imagery of the self. Recent evidence suggests that low-level mechanisms of multisensory and sensorimotor integration may play a fundamental role in encoding self-location and the first-person perspective (Blanke and Metzinger, 2009). Neurological patients with out-of body experiences (OBE) suffer from abnormal self-location and the first-person perspective due to a damage in the temporo-parietal junction (Blanke et al., 2004). Although self-location and the first-person perspective can be studied experimentally (Lenggenhager et al., 2009), the neural underpinnings of self-location have yet to be investigated. To investigate the brain network involved in self-location and first-person perspective we used visuo-tactile multisensory conflict, magnetic resonance (MR)-compatible robotics, and fMRI in study 1, and lesion analysis in a sample of 9 patients with OBE due to focal brain damage in study 2. Methods: Twenty-two participants saw a video showing either a person's back or an empty room being stroked (visual stimuli) while the MR-compatible robotic device stroked their back (tactile stimulation). Direction and speed of the seen stroking could either correspond (synchronous) or not (asynchronous) to those of the seen stroking. Each run comprised the four conditions according to a 2x2 factorial design with Object (Body, No-Body) and Synchrony (Synchronous, Asynchronous) as main factors. Self-location was estimated using the mental ball dropping (MBD; Lenggenhager et al., 2009). After the fMRI session participants completed a 6-item adapted from the original questionnaire created by Botvinick and Cohen (1998) and based on questions and data obtained by Lenggenhager et al. (2007, 2009). They were also asked to complete a questionnaire to disclose the perspective they adopted during the illusion. Response times (RTs) for the MBD and fMRI data were analyzed with a 3-way mixed model ANOVA with the in-between factor Perspective (up, down) and the two with-in factors Object (body, no-body) and Stroking (synchronous, asynchronous). Quantitative lesion analysis was performed using MRIcron (Rorden et al., 2007). We compared the distributions of brain lesions confirmed by multimodality imaging (Knowlton, 2004) in patients with OBE with those showing complex visual hallucinations involving people or faces, but without any disturbance of self-location and first person perspective. Nine patients with OBE were investigated. The control group comprised 8 patients. Structural imaging data were available for normalization and co-registration in all the patients. Normalization of each patient's lesion into the common MNI (Montreal Neurological Institute) reference space permitted simple, voxel-wise, algebraic comparisons to be made. Results: Even if in the scanner all participants were lying on their back and were facing upwards, analysis of perspective showed that half of the participants had the impression to be looking down at the virtual human body below them, despite any cues about their body position (Down-group). The other participants had the impression to be looking up at the virtual body above them (Up-group). Analysis of Q3 ("How strong was the feeling that the body you saw was you?") indicated stronger self-identification with the virtual body during the synchronous stroking. RTs in the MBD task confirmed these subjective data (significant 3-way interaction between perspective, object and stroking). fMRI results showed eight cortical regions where the BOLD signal was significantly different during at least one of the conditions resulting from the combination of Object and Stroking, relative to baseline: right and left temporo-parietal junction, right EBA, left middle occipito-temporal gyrus, left postcentral gyrus, right medial parietal lobe, bilateral medial occipital lobe (Fig 1). The activation patterns in right and left temporo-parietal junction and right EBA reflected changes in self-location and perspective as revealed by statistical analysis that was performed on the percentage of BOLD change with respect to the baseline. Statistical lesion overlap comparison (using nonparametric voxel based lesion symptom mapping) with respect to the control group revealed the right temporo-parietal junction, centered at the angular gyrus (Talairach coordinates x = 54, y =-52, z = 26; p>0.05, FDR corrected). Conclusions: The present questionnaire and behavioural results show that - despite the noisy and constraining MR environment) our participants had predictable changes in self-location, self-identification, and first-person perspective when robotic tactile stroking was applied synchronously with the robotic visual stroking. fMRI data in healthy participants and lesion data in patients with abnormal self-location and first-person perspective jointly revealed that the temporo-parietal cortex especially in the right hemisphere encodes these conscious experiences. We argue that temporo-parietal activity reflects the experience of the conscious "I" as embodied and localized within bodily space.

Application of Visual-Inertial SLAM for 3D Mapping of Underground Environments

The underground scenarios are one of the most challenging environments for accurate and precise 3d mapping where hostile conditions like absence of Global Positioning Systems, extreme lighting variations and geometrically smooth surfaces may be expected. So far, the state-of-the-art methods in underground modelling remain restricted to environments in which pronounced geometric features are abundant. This limitation is a consequence of the scan matching algorithms used to solve the localization and registration problems. This paper contributes to the expansion of the modelling capabilities to structures characterized by uniform geometry and smooth surfaces, as is the case of road and train tunnels. To achieve that, we combine some state of the art techniques from mobile robotics, and propose a method for 6DOF platform positioning in such scenarios, that is latter used for the environment modelling. A visual monocular Simultaneous Localization and Mapping (MonoSLAM) approach based on the Extended Kalman Filter (EKF), complemented by the introduction of inertial measurements in the prediction step, allows our system to localize himself over long distances, using exclusively sensors carried on board a mobile platform. By feeding the Extended Kalman Filter with inertial data we were able to overcome the major problem related with MonoSLAM implementations, known as scale factor ambiguity. Despite extreme lighting variations, reliable visual features were extracted through the SIFT algorithm, and inserted directly in the EKF mechanism according to the Inverse Depth Parametrization. Through the 1-Point RANSAC (Random Sample Consensus) wrong frame-to-frame feature matches were rejected. The developed method was tested based on a dataset acquired inside a road tunnel and the navigation results compared with a ground truth obtained by post-processing a high grade Inertial Navigation System and L1/L2 RTK-GPS measurements acquired outside the tunnel. Results from the localization strategy are presented and analyzed.

Intelligent task-level grasp mapping for robot control

In the future, robots will enter our everyday lives to help us with various tasks.For a complete integration and cooperation with humans, these robots needto be able to acquire new skills. Sensor capabilities for navigation in real humanenvironments and intelligent interaction with humans are some of the keychallenges.Learning by demonstration systems focus on the problem of human robotinteraction, and let the human teach the robot by demonstrating the task usinghis own hands. In this thesis, we present a solution to a subproblem within thelearning by demonstration field, namely human-robot grasp mapping. Robotgrasping of objects in a home or office environment is challenging problem.Programming by demonstration systems, can give important skills for aidingthe robot in the grasping task.The thesis presents two techniques for human-robot grasp mapping, directrobot imitation from human demonstrator and intelligent grasp imitation. Inintelligent grasp mapping, the robot takes the size and shape of the object intoconsideration, while for direct mapping, only the pose of the human hand isavailable.These are evaluated in a simulated environment on several robot platforms.The results show that knowing the object shape and size for a grasping taskimproves the robot precision and performance

Multi-sensor Simultaneous Localization and Mapping

Simultaneous localization and mapping(SLAM) is a very important problem in mobile robotics. Many solutions have been proposed by different scientists during the last two decades, nevertheless few studies have considered the use of multiple sensors simultane¬ously. The solution is on combining several data sources with the aid of an Extended Kalman Filter (EKF). Two approaches are proposed. The first one is to use the ordinary EKF SLAM algorithm for each data source separately in parallel and then at the end of each step, fuse the results into one solution. Another proposed approach is the use of multiple data sources simultaneously in a single filter. The comparison of the computational com¬plexity of the two methods is also presented. The first method is almost four times faster than the second one.

Efficient topology estimation for large scale optical mapping

Large scale image mosaicing methods are in great demand among scientists who study different aspects of the seabed, and have been fostered by impressive advances in the capabilities of underwater robots in gathering optical data from the seafloor. Cost and weight constraints mean that lowcost Remotely operated vehicles (ROVs) usually have a very limited number of sensors. When a low-cost robot carries out a seafloor survey using a down-looking camera, it usually follows a predetermined trajectory that provides several non time-consecutive overlapping image pairs. Finding these pairs (a process known as topology estimation) is indispensable to obtaining globally consistent mosaics and accurate trajectory estimates, which are necessary for a global view of the surveyed area, especially when optical sensors are the only data source. This thesis presents a set of consistent methods aimed at creating large area image mosaics from optical data obtained during surveys with low-cost underwater vehicles. First, a global alignment method developed within a Feature-based image mosaicing (FIM) framework, where nonlinear minimisation is substituted by two linear steps, is discussed. Then, a simple four-point mosaic rectifying method is proposed to reduce distortions that might occur due to lens distortions, error accumulation and the difficulties of optical imaging in an underwater medium. The topology estimation problem is addressed by means of an augmented state and extended Kalman filter combined framework, aimed at minimising the total number of matching attempts and simultaneously obtaining the best possible trajectory. Potential image pairs are predicted by taking into account the uncertainty in the trajectory. The contribution of matching an image pair is investigated using information theory principles. Lastly, a different solution to the topology estimation problem is proposed in a bundle adjustment framework. Innovative aspects include the use of fast image similarity criterion combined with a Minimum spanning tree (MST) solution, to obtain a tentative topology. This topology is improved by attempting image matching with the pairs for which there is the most overlap evidence. Unlike previous approaches for large-area mosaicing, our framework is able to deal naturally with cases where time-consecutive images cannot be matched successfully, such as completely unordered sets. Finally, the efficiency of the proposed methods is discussed and a comparison made with other state-of-the-art approaches, using a series of challenging datasets in underwater scenarios

Semantic mapping using mobile robots

Robotic mapping is the process of automatically constructing an environment representation using mobile robots. We address the problem of semantic mapping, which consists of using mobile robots to create maps that represent not only metric occupancy but also other properties of the environment. Specifically, we develop techniques to build maps that represent activity and navigability of the environment. Our approach to semantic mapping is to combine machine learning techniques with standard mapping algorithms. Supervised learning methods are used to automatically associate properties of space to the desired classification patterns. We present two methods, the first based on hidden Markov models and the second on support vector machines. Both approaches have been tested and experimentally validated in two problem domains: terrain mapping and activity-based mapping.

Cross-situational learning of object-word mapping using Neural Modeling Fields

The issue of how children learn the meaning of words is fundamental to developmental psychology. The recent attempts to develop or evolve efficient communication protocols among interacting robots or Virtual agents have brought that issue to a central place in more applied research fields, such as computational linguistics and neural networks, as well. An attractive approach to learning an object-word mapping is the so-called cross-situational learning. This learning scenario is based on the intuitive notion that a learner can determine the meaning of a word by finding something in common across all observed uses of that word. Here we show how the deterministic Neural Modeling Fields (NMF) categorization mechanism can be used by the learner as an efficient algorithm to infer the correct object-word mapping. To achieve that we first reduce the original on-line learning problem to a batch learning problem where the inputs to the NMF mechanism are all possible object-word associations that Could be inferred from the cross-situational learning scenario. Since many of those associations are incorrect, they are considered as clutter or noise and discarded automatically by a clutter detector model included in our NMF implementation. With these two key ingredients - batch learning and clutter detection - the NMF mechanism was capable to infer perfectly the correct object-word mapping. (C) 2009 Elsevier Ltd. All rights reserved.

Environment mapping for mobile robots navigation using hierarchical neural network and omnivision

Autonomous robots must be able to learn and maintain models of their environments. In this context, the present work considers techniques for the classification and extraction of features from images in joined with artificial neural networks in order to use them in the system of mapping and localization of the mobile robot of Laboratory of Automation and Evolutive Computer (LACE). To do this, the robot uses a sensorial system composed for ultrasound sensors and a catadioptric vision system formed by a camera and a conical mirror. The mapping system is composed by three modules. Two of them will be presented in this paper: the classifier and the characterizer module. The first module uses a hierarchical neural network to do the classification; the second uses techiniques of extraction of attributes of images and recognition of invariant patterns extracted from the places images set. The neural network of the classifier module is structured in two layers, reason and intuition, and is trained to classify each place explored for the robot amongst four predefine classes. The final result of the exploration is the construction of a topological map of the explored environment. Results gotten through the simulation of the both modules of the mapping system will be presented in this paper. © 2008 IEEE.

Underwater mapping using sonar

[ES]El proyecto contiene módulos de simulación, procesado de datos, mapeo y localización, desarrollados en C++ utilizando ROS (Robot Operating System) y PCL (Point Cloud Library). Ha sido desarrollado bajo el proyecto de robótica submarina AVORA.Se han caracterizado el vehículo y el sensor, y se han analizado diferentes tecnologías de sensores y mapeo. Los datos pasan por tres etapas: Conversión a nube de puntos, filtrado por umbral, eliminación de puntos espureos y, opcionalmente, detección de formas. Estos datos son utilizados para construir un mapa de superficie multinivel. La otra herramienta desarrollada es un algoritmo de Punto más Cercano Iterativo (ICP) modificado, que tiene en cuenta el modo de funcionamiento del sonar de imagen utilizado.

RatSLAM: A hippocampal model for simultaneous localisation and mapping

The work presents a new approach to the problem of simultaneous localization and mapping - SLAM - inspired by computational models of the hippocampus of rodents. The rodent hippocampus has been extensively studied with respect to navigation tasks, and displays many of the properties of a desirable SLAM solution. RatSLAM is an implementation of a hippocampal model that can perform SLAM in real time on a real robot. It uses a competitive attractor network to integrate odometric information with landmark sensing to form a consistent representation of the environment. Experimental results show that RatSLAM can operate with ambiguous landmark information and recover from both minor and major path integration errors.

Biologically inspired visual landmark processing for simultaneous localization and mapping

This paper illustrates a method for finding useful visual landmarks for performing simultaneous localization and mapping (SLAM). The method is based loosely on biological principles, using layers of filtering and pooling to create learned templates that correspond to different views of the environment. Rather than using a set of landmarks and reporting range and bearing to the landmark, this system maps views to poses. The challenge is to produce a system that produces the same view for small changes in robot pose, but provides different views for larger changes in pose. The method has been developed to interface with the RatSLAM system, a biologically inspired method of SLAM. The paper describes the method of learning and recalling visual landmarks in detail, and shows the performance of the visual system in real robot tests.