Computer Science

We present incremental smoothing and mapping (iSAM), a novel approach to the simultaneous localization and mapping problem that is based on fast incremental matrix factorization. iSAM provides an efficient and exact solution by updating a QR factorization of the naturally sparse smoothing information matrix, therefore recalculating only the matrix entries that actually change. iSAM is efficient even for robot trajectories with many loops as it avoids unnecessary fill-in in the factor matrix by periodic variable reordering. Also, to enable data association in real-time, we provide efficient algorithms to access the estimation uncertainties of interest based on the factored information matrix. We systematically evaluate the different components of iSAM as well as the overall algorithm using various simulated and real-world datasets for both landmark and pose-only settings.

We propose a novel approach to the problem of simultaneous localization and mapping (SLAM) based on incremental smoothing, that is suitable for real-time applications in large-scale environments. The main advantages over filter-based algorithms are that we solve the full SLAM problem without the need for any approximations, and that we do not suffer from linearization errors. We achieve efficiency by updating the square-root information matrix, a factored version of the naturally sparse smoothing information matrix. We can efficiently recover the exact trajectory and map at any given time by back-substitution. Furthermore, our approach allows access to the exact covariances, as it does not suffer from under-estimation of uncertainties, which is another problem inherent to filters. We present simulation-based results for the linear case, showing constant time updates for exploration tasks. We further evaluate the behavior in the presence of loops, and discuss how our approach extends to the non-linear case. Finally, we evaluate the overall non-linear algorithm on the standard Victoria Park data set.

Solving the SLAM problem is one way to enable a robot to explore, map, and navigate in a previously unknown environment. We investigate smoothing approaches as a viable alternative to extended Kalman filter-based solutions to the problem. In particular, we look at approaches that factorize either the associated information matrix or the measurement Jacobian into square root form. Such techniques have several significant advantages over the EKF: they are faster yet exact, they can be used in either batch or incremental mode, are better equipped to deal with non-linear process and measurement models, and yield the entire robot trajectory, at lower cost for a large class of SLAM problems. In addition, in an indirect but dramatic way, column ordering heuristics automatically exploit the locality inherent in the geographic nature of the SLAM problem. In this paper we present the theory underlying these methods, along with an interpretation of factorization in terms of the graphical model associated with the SLAM problem. We present both simulation results and actual SLAM experiments in large-scale environments that underscore the potential of these methods as an alternative to EKF-based approaches.

Pose estimation of outdoor robots presents some distinct challenges due to the various uncertainties in the robot sensing and action. In particular, global positioning sensors of outdoor robots do not always work perfectly, causing large drift in the location estimate of the robot. To overcome this common problem, we propose a new approach for global localization using place recognition. First, we learn the location of some arbitrary key places using odometry measurements and GPS measurements only at the start and the end of the robot trajectory. In subsequent runs, when the robot perceives a key place, our fixed-lag smoother fuses odometry measurements with the relative location to the key place to improve its pose estimate. Outdoor mobile robot experiments show that place recognition measurements significantly improve the estimate of the smoother in the absence of GPS measurements.

We present an algorithm for pose estimation using fixed-lag smoothing. We show that fixed-lag smoothing enables inclusion of measurements from multiple asynchronous measurement sources in an optimal manner. Since robots usually have a plurality of uncoordinated sensors, our algorithm has an advantage over filtering-based estimation algorithms, which cannot incorporate delayed measurements optimally. We provide an implementation of the general fixed-lag smoothing algorithm using square root smoothing, a technique that has recently become prominent. Square root smoothing uses fast sparse matrix factorization and enables our fixed-lag pose estimation algorithm to run at upwards of 20 Hz. Our algorithm has been extensively tested over hundreds of hours of operation on a robot operating in outdoor environments. We present results based on these tests that verify our claims using wheel encoders, visual odometry, and GPS as sensors.

We introduce incremental smoothing and mapping (iSAM), a novel approach to the problem of simultaneous localization and mapping (SLAM) that addresses the data association problem and allows real-time application in large-scale environments. We employ smoothing to obtain the complete trajectory and map without the need for any approximations, exploiting the natural sparsity of the smoothing information matrix. A QR-factorization of this information matrix is at the heart of our approach. It provides efficient access to the exact covariances as well as to conservative estimates that are used for online data association. It also allows recovery of the exact trajectory and map at any given time by backsubstitution. Instead of refactoring in each step, we update the QR-factorization whenever a new measurement arrives. We analyze the effect of loops, and show how our approach extends to the non-linear case. Finally, we provide experimental validation of the overall non-linear algorithm based on the standard Victoria Park data set with unknown correspondences.

The problem of simultaneous localization and mapping has received much attention over the last years. Especially large scale environments, where the robot trajectory loops back on itself, are a challenge. In this paper we introduce a new solution to this problem of closing the loop. Our algorithm is EM-based, but differs from previous work. The key is a probability distribution over partitions of feature tracks that is determined in the E-step, based on the current estimate of the motion. This virtual structure is then used in the M-step to obtain a better estimate for the motion. We demonstrate the success of our algorithm in experiments on real laser data.

In some applications objects are known to have nonsmooth or "jagged" edges, which are not well approximated by smooth curves. We use subdivision curves as a simple but flexible curve representation, which allows tagging corners to model non-smooth features along otherwise smooth curves. A Markov chain Monte Carlo approach yields an approximate posterior distribution over tags, while Rao-Blackwellization allows us to integrate out the control point locations by an approximation. We apply this general methodology to multi-view reconstruction of piecewise smooth curves from multiple calibrated views in which the object has been segmented from the background. Results are shown for multiple images of two pot shards as would be encountered in archaeological applications.

This work focuses on real-time compression of laser data on board a mobile robot platform. Data is transmitted from the robot over low-bandwidth channels or incrementally in short bursts to a host, where it can be further processed for visualization. For compression purposes, the data is represented as a gray scale depth image. Considered are existing lossless image and file compression schemes (Unix compress, gzip, bzip2, PNG,, as well as wavelet transformations tailored to the specific nature of the data. Testing is done on several sets of indoor data acquired by a robot moving through rooms and hallways. The results show that Jpeg-LS compression performs best in this setting.

We describe our entry in the AAAI 2002 Urban Search and Rescue (USAR) competition, a marsupial team consisting of a larger wheeled robot and several small legged robots, carried around by the larger robot. This setup exploits complimentary strengths of each robot type in a challenging domain. We describe both the hardware and software architecture, and the on-board real-time mapping which forms the basis of accurate victim-localization crucial to the USAR domain. We also evaluate what challenges remain to be resolved in order to deploy search and rescue robots in realistic scenarios.

This paper presents an approach to learning an optimal behavioral parameterization in the framework of a Case-Based Reasoning methodology for autonomous navigation tasks. It is based on our previous work on a behavior-based robotic system that also employed spatio-temporal case-based reasoning in the selection of behavioral parameters but was not capable of learning new parameterizations. The present method extends the case-based reasoning module by making it capable of learning new and optimizing the existing cases where each case is a set of behavioral parameters. The learning process can either be a separate training process or be part of the mission execution. In either case, the robot learns an optimal parameterization of its behavior for different environments it encounters. The goal of this research is not only to automatically optimize the performance of the robot but also to avoid the manual configuration of behavioral parameters and the initial configuration of a case library, both of which require the user to possess good knowledge of robot behavior and the performance of numerous experiments. The presented method was integrated within a hybrid robot architecture and evaluated in extensive computer simulations, showing a significant increase in the performance over a nonadaptive system and a performance comparable to a non-learning CBR system that uses a hand-coded case library.

Camera-based simultaneous localization and mapping or visual SLAM has received much attention recently. Typically single cameras, multiple cameras in a stereo setup or omni-directional cameras are used. We propose a different approach, where multiple cameras can be mounted on a robot in an arbitrary configuration. Allowing the cameras to face in different directions yields better constraints than single cameras or stereo setups can provide, simplifying the reconstruction of large-scale environments. And in contrast to omni-directional sensors, the available resolution can be focused on areas of interest depending on the application. We describe a sparse SLAM approach that is suitable for real-time reconstruction from such multi-camera configurations. We have implemented the system and show experimental results in a large-scale environment, using a custom made eight-camera rig.

Abstract���This paper describes a system for performing multi-session visual mapping in large-scale environments. Multi-session mapping considers the problem of combining the results of multiple Simultaneous Localisation and Mapping (SLAM) missions performed repeatedly over time in the same environment. The goal is to robustly combine multiple maps in a common metrical coordinate system, with consistent estimates of uncertainty.

Abstract The problem of simultaneous localization and mapping has received much attention over the last years. Especially large scale environments, where the robot trajectory loops back on itself, are a challenge. In this paper we introduce a new solution to this problem of closing the loop. Our algorithm is EM-based, but differs from previous work. The key is a probability distribution over partitions of feature tracks that is determined in the E-step, based on the current estimate of the motion.

Abstract In this paper, we present incremental smoothing and mapping (iSAM), which is a novel approach to the simultaneous localization and mapping problem that is based on fast incremental matrix factorization. iSAM provides an efficient and exact solution by updating a QR factorization of the naturally sparse smoothing information matrix, thereby recalculating only those matrix entries that actually change.

Abstract In some applications objects are known to have non-smooth or" jagged" edges, which are not well approximated by smooth curves. We use subdivision curves as a simple but flexible curve representation, which allows tagging corners to model non-smooth features along otherwise smooth curves. A Markov chain Monte Carlo approach yields an approximate posterior distribution over tags, while Rao-Blackwellization allows us to integrate out the control point locations by an approximation.

In this paper we present a novel data structure, the Bayes tree, which exploits the connections between graphical model inference and sparse linear algebra. The proposed data structure provides a new perspective on an entire class of simultaneous localization and mapping (SLAM) algorithms. Similar to a junction tree, a Bayes tree encodes a factored probability density, but unlike the junction tree it is directed and maps more naturally to the square root information matrix of the SLAM problem.

Abstract We describe our entry in the AAAI 2002 Urban Search and Rescue (USAR) competition, a marsupial team consisting of a larger wheeled robot and several small legged robots, carried around by the larger robot. This setup exploits complimentary strengths of each robot type in a challenging domain. We describe both the hardware and software architecture, and the on-board real-time mapping which forms the basis of accurate victim-localization crucial to the USAR domain.

Abstract Many online inference problems in computer vision and robotics are characterized by probability distributions whose factor graph representations are sparse and whose factors are all Gaussian functions of error residuals. Under these conditions, maximum likelihood estimation corresponds to solving a sequence of sparse least-squares minimization problems in which additional summands are added to the objective function over time.