Abstract: Accurate 3d perception from video sequences is a core subject in computer vision and robotics, since it forms the basis of subsequent scene analysis. In practice however, online requirements often severely limit the utilizable camera resolution and hence also reconstruction accuracy. Furthermore, real-time systems often rely on heavy parallelism which can prevent applications in mobile devices or driver assistance systems, especially in cases where FPGAs cannot be employed. This paper proposes a novel approach to build 3d maps from high-resolution stereo sequences in real-time. Inspired by recent progress in stereo matching, we propose a sparse feature matcher in conjunction with an efficient and robust visual odometry algorithm. Our reconstruction pipeline combines both techniques with efficient stereo matching and a multi-view linking scheme for generating consistent 3d point clouds. In our experiments we show that the proposed odometry method achieves state-of-the-art accuracy. Including feature matching, the visual odometry part of our algorithm runs at 25 frames per second, while - at the same time - we obtain new depth maps at 3-4 fps, sufficient for online 3d reconstructions.

Abstract: This paper presents a fixed-lag smoothing algorithm for tracking the motion of a mobile robot in real time. The algorithm processes measurements from proprioceptive (e.g., odometry, inertial measurement unit) and exteroceptive (e.g., camera, laser scanner) sensors, in order to estimate the trajectory of the vehicle. Smoothing is carried out in the information-filtering framework, and utilizes iterative minimization, which renders the method well-suited for applications where the effects of the measurements' nonlinearity are significant. The algorithm attains bounded computational complexity by marginalizing out older states. The key contribution of this work is a detailed analysis of the effects of the marginalization process on the consistency properties of the estimator. Based on this analysis, a linearization scheme that results in substantially improved accuracy, compared to the standard linearization approach, is proposed. Both simulation and real-world experimental results are presented, which demonstrate that the proposed method attains localization accuracy superior to that of competing approaches.

Abstract: Inertial navigation systems (INS) are composed of inertial sensors, such as accelerometers and gyroscopes. An INS updates its orientation and position automatically; it has an acceptable stability over the short term, however this stability deteriorates over time. Odometry, used to estimate the position of a mobile robot, employs encoders attached to the robot’s wheels. However, errors occur caused by the integrative nature of the rotating speed and the slippage between the wheel and the ground. In this paper, we discuss mobile robot position estimation without using external signals in indoor environments. In order to achieve optimal solutions, a Kalman filter that estimates the orientation and velocity of mobile robots has been designed. The proposed system combines INS and odometry and delivers more accurate position information than standalone odometry.

Abstract: In RGB-D sensor based visual odometry the goal is to estimate a sequence of camera movements using image and/or range measurements Direct methods solve the problem by minimizing intensity error In this work a depth map obtained from a RGB-D sensor is considered as a new measurement which is combined with a direct photometric cost function The minimization of the bi-objective cost function produces 3D camera motion parameters which registers two 3D surfaces within a same coordinate system The given formulation does not require any predetermined temporal correspondencies nor feature extraction when having a sufficient frame rate It is shown how incorporating the depth measurement robustifies the cost function in case of insufficient texture information and non-Lambertian surfaces Finally the method is demonstrated in the Planetary Robotics Vision Ground Processing (PRoVisG) competition where visual odometry and 3D reconstruction results are solved for a stereo image sequence captured using a Mars rover

Abstract: Dead reckoning techniques such as inertial navigation and odometry are integrated with GPS to avoid interruption of navigation solutions due to lack of visible satellites. A common method to achieve a low-cost navigation solution for land vehicles is to use a MEMS-based inertial measurement unit (IMU) for integration with GPS. This integration is traditionally accomplished by means of a Kalman filter (KF). Due to the significant inherent errors of MEMS inertial sensors and their time-varying changes, which are difficult to model, severe position error growth happens during GPS outages. The positional accuracy provided by the KF is limited by its linearized models. A Particle filter (PF), being a nonlinear technique, can accommodate for arbitrary inertial sensor characteristics and motion dynamics. An enhanced version of the PF, called Mixture PF, is employed in this paper. It samples from both the prior importance density and the observation likelihood, leading to an improved performance. Furthermore, in order to enhance the performance of MEMS-based IMU/GPS integration during GPS outages, the use of pitch and roll calculated from the longitudinal and transversal accelerometers together with the odometer data as a measurement update is proposed in this paper. These updates aid the IMU and limit the positional error growth caused by two horizontal gyroscopes, which are a major source of error during GPS outages. The performance of the proposed method is examined on road trajectories, and results are compared to the three different KF-based solutions. The proposed Mixture PF with velocity, pitch, and roll updates outperformed all the other solutions and exhibited an average improvement of approximately 64% over KF with the same updates, about 85% over KF with velocity updates only, and around 95% over KF without any updates during GPS outages.

Abstract: This paper considers the problem of state estimation in autonomous navigation from a theoretical perspective. In particular, the investigation concerns problems where the information provided by the sensor data is not sufficient to carry out the state estimation (i.e., the state is not observable). For these systems, the concept of continuous symmetry is introduced. Detection of the continuous symmetries of a given system has a very practical importance. It allows the detection of an observable state whose components are nonlinear functions of the original nonobservable state. So far, this theoretical and very general concept has been applied to deal with two distinct fundamental estimation problems in the framework of mobile robotics. The former is in the framework of self-calibration, and the latter is in the framework of the fusion of the data provided by inertial sensors and vision sensors. For reasons of length, only the former is discussed. In particular, the theoretical machinery is used to address a specific calibration problem. The solution constrains the robot to move along specific trajectories in order to be able to apply the calibration algorithm. This paper provides two distinct contributions. The first is the introduction of this concept of continuous symmetry. The second is the introduction of a simple and efficient strategy to extrinsically calibrate a bearing sensor (e.g., a vision sensor) mounted on a vehicle and, simultaneously, estimate the parameters describing the systematic error of its odometry system. Many accurate simulations and real experiments show the robustness, the efficiency, and the accuracy of the proposed strategy.

Abstract: We present a novel algorithm for topological mapping, which is the problem of finding the graph structure of an environment from a sequence of measurements. Our algorithm, called Online Probabilistic Topological Mapping (OPTM), systematically addresses the problem by constructing the posterior on the space of all possible topologies given measurements. With each successive measurement, the posterior is updated incrementally using a Rao-Blackwellized particle filter. We present efficient sampling mechanisms using data-driven proposals and prior distributions on topologies that further enable OPTM's operation in an online manner. OPTM can incorporate various sensors seamlessly, as is demonstrated by our use of appearance, laser, and odometry measurements. OPTM is the first topological mapping algorithm that is theoretically accurate, systematic, sensor independent, and online, and thus advances the state of the art significantly. We evaluate the algorithm on a robot in diverse environments.

Abstract: Entropy-Based Active Vision for a Humanoid Soccer Robot- A Realistic Simulation Tool for Testing Face Recognition Systems under Real-World Conditions- Thermal Face Recognition Using Local Interest Points and Descriptors for HRI Applications- On Progress in RoboCup: The Simulation League Showcase- Odometry Correction for Humanoid Robots Using Optical Sensors- SSL-Humanoid: RoboCup Soccer Using Humanoid Robots under the Global Vision- Localization with Non-unique Landmark Observations- MR-Simulator: A Simulator for the Mixed Reality Competition of RoboCup- Learning Footstep Prediction from Motion Capture Kicking a Ball - Modeling Complex Dynamic Motions for Humanoid Robots- Towards Semantic Scene Analysis with Time-of-Flight Cameras- Providing Ground-Truth Data for the Nao Robot- Optimizing Particle Filter Parameters for Self-localization- Improving People Awareness of Service Robots by Semantic Scene Knowledge- An Evaluation of Open Source SURF Implementations- A Semantic World Model for Urban Search and Rescue Based on Heterogeneous Sensors- Improving Biped Walk Stability Using Real-Time Corrective Human Feedback- A Review of Shape Memory Alloy Actuators in Robotics- Biologically Inspired Mobile Robot Control Robust to Hardware Failures and Sensor Noise- TopLeague & Bundesliga Manager New Generation Online Soccer Games- Human vs Robotic Soccer: How Far Are They? A Statistical Comparison- Learning Powerful Kicks on the Aibo ERS-7: The Quest for a Striker- Real-Time Active Vision by Entropy Minimization Applied to Localization- Multi-agent Behavior Composition through Adaptable Software Architectures and Tangible Interfaces- A Novel Real-Time Local Visual Feature for Omnidirectional Vision Based on FAST and LBP- Mixed 2D/3D Perception for Autonomous Robots in Unstructured Environments- Hierarchical Multi-robot Coordination- Biped Walking Using Coronal and Sagittal Movements Based on Truncated Fourier Series- Realistic Simulation of Laser Range Finder Behavior in a Smoky Environment- Cooperative Localization Based on Visually Shared Objects- Parameter Optimization of a Signal-Based Omni-Directional Biped Locomotion Using Evolutionary Strategies- Designing Effective Humanoid Soccer Goalies- A Supporter Behavior for Soccer Playing Humanoid Robots- Utilizing the Structure of Field Lines for Efficient Soccer Robot Localization- Robot Detection with a Cascade of Boosted Classifiers Based on Haar-Like Features- LearnPNP: A Tool for Learning Agent Behaviors

Abstract: The paper summarises the main features concerning the definition of an efficient odometry algorithm to be used in modern automatic train protection and control (ATP/ATC) systems. The availability of a reliable speed and travelled distance estimation is essential for the efficiency and the safety of the whole system. The first essential step in odometric subsystem design is the choice of the sensors, whose output signals will be used for velocity estimation. Then a suitable procedure fusing sensor signals has to be defined as a function of number and type of sensors and accuracy and safety targets. In the paper, the main features of an innovative solution will be summarised and its performance will be presented, in terms of precision in speed and travelled distance estimation.

Abstract: It is an important ability for any mobile robot to be able to estimate its posture and to gauge the distance it travelled. The information can be obtained from various sources. In this work, we have addressed this problem in a dynamic quadruped robot. We have designed and implemented a navigation algorithm for full body state (position, velocity, and attitude) estimation that does not use any external reference (such as GPS, or visual landmarks). Extended Kalman Filter was used to provide error estimation and data fusion from two independent sources of information: Inertial Navigation System mechanization algorithm processing raw inertial data, and legged odometry, which provided velocity aiding. We present a novel data-driven architecture for legged odometry that relies on a combination of joint sensor signals and pressure sensors. Our navigation system ensures precise tracking of a running robot's posture (roll and pitch), and satisfactory tracking of its position over medium time intervals. We have shown our method to work for two different dynamic turning gaits and on two terrains with significantly different friction. We have also successfully demonstrated how our method generalizes to different velocities.

Abstract: The prevention of vehicle accidents is targeted. A road texture model is created based on a vehicle camera image. An initial vehicle location estimate is determined, and map imagery is obtained based on this location estimate. A refined vehicle location is determined using visual egomotion. In particular, 3D features of the vehicle image and the retrieved map imagery are identified and aligned. A map image is selected based on this alignment, and the location associated with the map image is modified by a displacement between the selected map image and the vehicle image to produce a refined vehicle location. A road boundary model is created based on the road texture model and the refined vehicle location, and a road departure model is created based on the road boundary model and vehicle odometry information. The operator of the vehicle is warned of a road departure based on the road departure model.

Abstract: This paper addresses the problem of using visual information to estimate vehicle motion (a.k.a. visual odometry) from a machine learning perspective. The vast majority of current visual odometry algorithms are heavily based on geometry, using a calibrated camera model to recover relative translation (up to scale) and rotation by tracking image features over time. Our method eliminates the need for a parametric model by jointly learning how image structure and vehicle dynamics affect camera motion. This is achieved with a Gaussian Process extension, called Coupled GP, which is trained in a supervised manner to infer the underlying function mapping optical flow to relative translation and rotation. Matched image features parameters are used as inputs and linear and angular velocities are the outputs in our non-linear multi-task regression problem. We show here that it is possible, using a single uncalibrated camera and establishing a first-order temporal dependency between frames, to jointly estimate not only a full 6 DoF motion (along with a full covariance matrix) but also relative scale, a non-trivial problem in monocular configurations. Experiments were performed with imagery collected with an unmanned aerial vehicle (UAV) flying over a deserted area at speeds of 100–120 km/h and altitudes of 80-100 m, a scenario that constitutes a challenge for traditional visual odometry estimators.

Abstract: The FeetSLAM technique builds on iterative processing of multiple sets of pedestrian odometry data, based on Foot- SLAM The objective is to obtain maps of large areas based on many data sets The central idea is that maps originating from other data sets are used as a so-called prior map for a given data set We show that this follows from the optimal FeetSLAM derivation but is more suited to practical computation limitations such as limited memory It also yields maps which are not overly dominated by one data set but rather balances the characteristics of each with the effect of averaging out errors Over iterations, FootSLAM maps are gradually combined to yield a high-accuracy global map - the iteration speed is controlled by employing concepts from simulated annealing We validate our approach using two data sets from two locations, consisting of four and five walks respectively

Abstract: This paper describes a relative localization system used to achieve the navigation of a convoy of robotic units in indoor environments. This positioning system is carried out fusing two sensorial sources: (a) an odometric system and (b) a laser scanner together with artificial landmarks located on top of the units. The laser source allows one to compensate the cumulative error inherent to dead-reckoning; whereas the odometry source provides less pose uncertainty in short trajectories. A discrete Extended Kalman Filter, customized for this application, is used in order to accomplish this aim under real time constraints. Different experimental results with a convoy of Pioneer P3-DX units tracking non-linear trajectories are shown. The paper shows that a simple setup based on low cost laser range systems and robot built-in odometry sensors is able to give a high degree of robustness and accuracy to the relative localization problem of convoy units for indoor applications.

Abstract: This paper presents a perception-based GPS-free approach for localizing a mobile robot in an orchard environment. An extended Kalman filter (EKF) algorithm is presented that uses a wheel odometry prediction step and laser rangefinder update steps. There are two update steps, one that uses measurements to reflective point features and one that uses measurements to linear features formed by tree rows. The features are associated to landmarks in previously surveyed maps. The practical issues of dealing with uncertainty both from the environment and the on-board sensors are discussed and accounted for. The resulting algorithm is demonstrated in over 20km of online operation in a variety of real orchard environments.

Abstract: This work develops a particle filter algorithm to localise a vehicle in the direction of travel without the use of GPS. The inputs to the algorithm include a terrain map of road grade, pitch measurements from an in-vehicle pitch sensor, and wheel odometry. Simulations and experiments at The Thomas D. Larson Transportation Institute test track are used to demonstrate the algorithm, observe the speed of convergence, and to determine key parameters for practical implementation. The results indicate that the method can quickly localise a vehicle with 1 m accuracy or better. Experiments over 5 km along Highway 322 in State College, Pennsylvania, were also used to demonstrate the algorithm.

Abstract: The MagneBike inspection robot is a climbing robot equipped with magnetic wheels. The robot is designed to drive on three-dimensional (3D) complexly shaped pipe structures; therefore it is necessary to provide 3D visualization tools for the user, who remotely controls the robot out of sight. The localization system is required to provide a 3D map of the unknown environment and the 3D location of the robot in the environment's map. The localization strategy proposed in this paper consists of combining 3D odometry with 3D scan registration. The odometry model is based on wheel encoders and a three-axis accelerometer. Odometry enables the tracking of the robot trajectory between consecutive 3D scans and is used as a prior for the scan matching algorithm. The 3D scan registration facilitates the construction of a 3D map of the environment and refines the robot position computed with odometry. This paper describes in detail the implementation of the localization concept. It presents the lightweight, small-sized 3D range finder that has been developed for the MagneBike. It also proposes an innovative 3D odometry model that estimates the local surface curvature to compensate for the absence of angular velocity inputs. The different tools are characterized in detail based on laboratory and field experiments. They show that the localization concepts reliably track the robot moving in the specific application environment. We also describe various techniques to optimize the 3D scanning process, which is time consuming, and to compensate for the identified limitations. These techniques are useful inputs for the future automatization of the robot's control and optimization of its localization process. © 2010 Wiley Periodicals, Inc. © 2011 Wiley Periodicals, Inc.

Abstract: We address the problem of simultaneous localization and mapping by combining visual loop-closure detection with metrical information given by the robot odometry. The proposed algorithm builds in real-time topo-metric maps of an unknown environment, with a monocular or omnidirectional camera and odometry gathered by motors encoders. A dedicated improved version of our previous work on purely appearance-based loop-closure detection [1] is used to extract potential loop-closure locations. Potential locations are then verified and classified using a new validation stage. The main contributions we bring are the generalization of the validation method for the use of monocular and omnidirectional camera with the removal of the camera calibration stage, the inclusion of an odometry-based evolution model in the Bayesian filter which improves accuracy and responsiveness, and the addition of a consistent metric position estimation. This new SLAM method does not require any calibration or learning stage (i.e. no a priori information about environment). It is therefore fully incremental and generates maps usable for global localization and planned navigation. This algorithm is moreover well suited for remote processing and can be used on toy robots with very small computational power.

Abstract: This paper describes the use of derivative free filters for mobile robot localization and navigation in an orchard. The localization algorithm fuses odometry and gyro measurements with line features representing the surrounding fruit trees of the orchard. The line features are created on basis of 2D laser scanner data by a least square algorithm. The three derivative free filters are compared to an EKF based localization method on a typical run covering four rows in the orchard. The Matlab® toolbox Kalmtool is used for easy switching between different filter implementations without the need for changing the base structure of the system.

Abstract: Recently, visual odometry has gained more attention due to its application to autonomous vehicles and robots In this paper, we propose visual odometry using a single camera This configuration is economically more attractive than other popular configurations using stereo cameras To estimate frame-to-frame motion more accurately, we thoroughly utilize prior knowledge on our operating space and motion Our visual odometry assumes that its operating space is locally flat so that its movement undergoes planar motion From this assumption, we can adopt the novel normalized 1-point algorithm to estimate relative pose between a pair of images Moreover, under this assumption, we can resolve scale ambiguity, the most serious problem in monocular visual odometry To estimate scale of motion, we introduce the other 1-point algorithm derived from planar homography of the ground plane Two 1-point algorithms are the key to accelerate our visual odometry for its further application to real-time systems Our experiments on synthetic data verified that the normalized 1-point algorithm was more accurate than other known method Moreover, our experiments on real data presented effectiveness of our proposed visual odometry

Abstract: In urban search and rescue scenarios, typical applications of robots include autonomous exploration of possibly dangerous sites, and the recognition of victims and other objects of interest. In complex scenarios, relying on only one type of sensor is often misleading, while using complementary sensors frequently helps improving the performance. To that end, we propose a probabilistic world model that leverages information from heterogeneous sensors and integrates semantic attributes. This method of reasoning about complementary information is shown to be advantageous, yielding increased reliability compared to considering all sensors separately. We report results from several experiments with a wheeled USAR robot in a complex indoor scenario. The robot is able to learn an accurate map, and to detect real persons and signs of hazardous materials based on inertial sensing, odometry, a laser range finder, visual detection, and thermal imaging. The results show that combining heterogeneous sensor information increases the detection performance, and that semantic attributes can be successfully integrated into the world model.

Abstract: A walking robot and a simultaneous localization and mapping method thereof in which odometry data acquired during movement of the walking robot are applied to image-based SLAM technology so as to improve accuracy and convergence of localization of the walking robot. The simultaneous localization and mapping method includes acquiring image data of a space about which the walking robot walks and rotational angle data of rotary joints relating to walking of the walking robot, calculating odometry data using kinematic data of respective links constituting the walking robot and the rotational angle data, and localizing the walking robot and mapping the space about which the walking robot walks using the image data and the odometry data.

Abstract: Vision-based mobile robot's simultaneous localization and mapping (SLAM) and navigation has been the source of countless research contributions because of rich sensory output and cost effectiveness of vision sensors. However, existing methods of vision-based SLAM and navigation are not effective for robots to be used in crowded environments such as train stations and shopping malls, because when we extract feature points from an image in crowded environments, many feature points are extracted from not only static objects but also dynamic objects such as humans. By recognizing all such feature points as landmarks, the algorithm collapses and errors occur in map building and self-localization. In this paper, we propose a SLAM and navigation method that is effective even in crowded environments by extracting robust 3D feature points from sequential vision images and odometry. By using the proposed method, we can eliminate unstable feature points extracted from dynamic objects and perform SLAM and navigation stably. We present experiments showing the utility of our approach in crowded environments, including map building and navigation.

Abstract: In this paper, we present a novel approach to localization for planetary rovers, in which sun sensor and inclinometer measurements are incorporated directly into a stereo visual odometry pipeline. Utilizing the absolute orientation information provided by the sun sensor significantly reduces the error growth of the visual odometry path estimate. The measurements have minimal computation, power, and mass requirements, providing a localization improvement at nearly negligible cost. We describe the mathematical formulation of error terms for the stereo camera, sun sensor, and inclinometer measurements, as well as the bundle adjustment framework for determining the maximum likelihood vehicle transformation. Improved localization accuracy is demonstrated through extensive experimental results from a 10 kilometre traversal of a Mars analogue site on Devon Island in the Canadian High Arctic.

Abstract: This paper presents a simultaneous localization and mapping (SLAM) algorithm that allows a recursive state estimation process to be both distributed and decentralized in a sparse robot network that is never guaranteed to be fully connected (communication-wise). In such a sparse network, a robot may not always have the latest odometry and measurements from other robots. Our approach allows robots to obtain a temporary (localization and map) estimate at the current timestep using information available locally, but we also ensure that the centralized-equivalent estimate can always be recovered by all robots at a later time; we do not require a robot to keep track of what other robots know when it applies the Markov property to discard past information. Our method is validated through a hardware SLAM experiment where we distribute data association hypotheses amongst a team of robots. Estimate errors are shown to validate the performance of our approach. We also discuss the trade-offs and show comparisons between our distributed approach versus a non-distributed one.

Abstract: We propose a particle filter-based algorithm for monocular vision-aided odometry for mobile robot localization. The algorithm fuses information from odometry with observations of naturally occurring static point features in the environment. A key contribution of this work is a novel approach for computing the particle weights, which does not require including the feature positions in the state vector. As a result, the computational and sample complexities of the algorithm remain low even in feature-dense environments. We validate the effectiveness of the approach extensively with both simulations as well as real-world data, and compare its performance against that of the extended Kalman filter (EKF) and FastSLAM. Results from the simulation tests show that the particle filter approach is better than these competing approaches in terms of the RMS error. Moreover, the experiments demonstrate that the approach is capable of achieving good localization accuracy in complex environments.

Abstract: This paper presents a method for computing the visual odometry of ground vehicles. By mounting a downward-looking camera on the vehicle, it can recover the motion of the vehicle only by using the video input. Different from other feature based algorithms, the key technique deployed in this system is an appearance-based approach — template matching. A rotated template matching approach is used for motion estimation. For each two consecutive images captured from the camera, an image patch is first selected from the first image, and then the image patch is rotated by every angle within a predefined range and matched against the second image. The best result from the matching indicates both the position translation and heading changes. For performance evaluation, this algorithm is tested on the indoor granite surface. In the end, we give a comparison of the result between wheel odometry and visual odometry.