We demonstrate a system built using probabilistic techniques that allows for remarkably accurate localization across our entire office building using nothing more than the built-in signal intensity meter supplied by standard 802.11 cards. While prior systems have required significant investments of human labor to build a detailed signal map, we can train our system by spending less than one minute per office or region, walking around with a laptop and recording the observed signal intensities of our building’s unmodified base stations. We actually collected over two minutes of data per office or region, about 28 man-hours of effort. Using less than half of this data to train the localizer, we can localize a user to the precise, correct location in over 95 % of our attempts, across the entire building. Even in the most pathological cases, we almost never localize a user any more distant than to the neighboring office. A user can obtain this level of accuracy with only two or three signal intensity measurements, allowing for a high frame rate of localization results. Furthermore, with a brief calibration period, our system can be adapted to work with previously unknown user hardware. We present results demonstrating the robustness of our system against a variety of untrained time-varying phenomena, including the presence or absence of people in the building across the day. Our system is sufficiently robust to enable a variety of locationaware applications without requiring special-purpose hardware or complicated training and calibration procedures.

Topological and metrical methods for representing spatial knowledge have complementary strengths. We present a hybrid extension to the Spatial Semantic Hierarchy that combines their strengths and avoids their weaknesses. Metrical SLAM methods are used to build local maps of small-scale space within the sensory horizon of the agent, while topological methods are used to represent the structure of large-scale space. We describe how a local perceptual map is analyzed to identify a local topology description and is abstracted to a topological place. The mapbuilding method creates a set of topological map hypotheses that are consistent with travel experience. The set of maps is guaranteed under reasonable assumptions to include the correct map. We demonstrate the method on a real environment with multiple nested large-scale loops.

Autonomous place detection has long been a major hurdle to topological map-building techniques. Theoretical work on topological mapping has assumed that places can be reliably detected by a robot, resulting in deterministic actions. Whether or not deterministic place detection is always achievable is controversial; however, even topological mapping algorithms that assume non-determinism benefit from highly reliable place detection. Unfortunately, topological map-building implementations often have handcoded place detection algorithms that are brittle and domain dependent.

Most simultaneous localization and mapping (SLAM) approaches focus on purely metrical approaches to map-building. We present a method for computing the global metrical map that builds on the structure provided by a topological map. This allows us to factor the uncertainty in the map into local metrical uncertainty (which is handled well by existing SLAM methods), global topological uncertainty (which is handled well by recently developed topological maplearning methods), and global metrical uncertainty (which can be handled effectively once the other types of uncertainty are factored out). We believe that this method for building the global metrical map will be scalable to very large environments.

The need for spatial representations and spatial reasoning is ubiquitous in AI – from robot planning and navigation, to interpreting visual inputs, to understanding natural language – in all these cases the need to represent and reason about spatial aspects of the world is of key importance. Related fields of research, such as geographic information science

To be autonomous, intelligent robots must learn the foundations of commonsense knowledge from their own sensorimotor experience in the world. We describe four recent research results that contribute to a theory of how a robot learning agent can bootstrap from the ‘blooming buzzing confusion ’ of the pixel level to a higher level ontology including distinctive states, places, objects, and actions. This is not a single learning problem, but a lattice of related learning tasks, each providing prerequisites for tasks to come later. Starting with completely uninterpreted sense and motor vectors, as well as an unknown environment, we show how a learning agent can separate the sense vector into modalities, learn the structure of individual modalities, learn natural primitives for the motor system, identify reliable relations between primitive actions and created sensory features, and can define useful control laws for homing and path-following. Building on this framework, we show how an agent can use self-organizing maps to identify useful sensory features in the environment, and can learn effective hill-climbing control laws to define distinctive states in terms of those features, and trajectory-following control laws to move from one distinctive state to another. Moving on to place recognition, we show how an agent can combine unsupervised learning, map-learning, and supervised learning to achieve high-performance recognition of places from rich sensory input. Finally, we take the first steps toward learning an ontology of objects, showing that a bootstrap learning robot can learn to individuate objects through motion, separating them from the static environment and from each other, and can learn properties useful for classification. These are four key steps in a larger research enterprise on the foundations of human and robot commonsense knowledge.

Loop-closing has long been recognized as a critical issue when building maps of large-scale environments from local observations. Topological mapping methods abstract the problem of determining the topological structure of the environment (i.e., how loops are closed) from the problem of determining the metrical layout of places in the map and dealing with noisy sensors. A recently developed incremental topological mapping algorithm [1], [2] generates all possible topological maps consistent with the experienced sequence of actions and observations and the topological axioms. These are then ordered by a preference criterion such as minimality or probability, to determine the single best map for continued planning and exploration. This paper presents the planarity constraint and analyzes its impact on the search-tree of all topological maps consistent with (non-metrical) exploration experience. Experimental studies demonstrate excellent results even in artificial environments where loop-closing is particularly difficult due to large amounts of perceptual aliasing and structural symmetry.

Abstract — This paper considers what can be accomplished using a mobile robot that has limited sensing. For navigation and mapping, the robot has only one sensor, which tracks the directions of depth discontinuities. There are no coordinates, and the robot is given a motion primitive that allows it to move toward discontinuities. The robot is incapable of performing localization or measuring any distances or angles. Nevertheless, when dropped into an unknown planar environment, the robot builds a data structure, called the Gap Navigation Tree, which enables it to navigate optimally in terms of Euclidean distance traveled. In a sense, the robot is able to learn the critical information contained in the classical shortest-path roadmap, although surprisingly it is unable to extract metric information. We prove these results for the case of a point robot placed into a simply connected, piecewise-analytic planar environment. The case of multiply connected environments is also addressed, in which it is shown that further sensing assumptions are needed. Due to the limited sensor given to the robot, globally optimal navigation is impossible; however, our approach achieves locally optimal (within a homotopy class) navigation, which is the best that is theoretically possible under this robot model. Index Terms — Visibility, navigation, optimality, map building, minimal sensing, shortest paths, information spaces, sensor-based

Abstract—While probabilistic techniques have previously been investigated extensively for performing inference over the space of metric maps, no corresponding general-purpose methods exist for topological maps. We present the concept of probabilistic topological maps (PTMs), a sample-based representation that approximates the posterior distribution over topologies, given available sensor measurements. We show that the space of topologies is equivalent to the intractably large space of set partitions on the set of available measurements. The combinatorial nature of the problem is overcome by computing an approximate, sample-based representation of the posterior. The PTM is obtained by performing Bayesian inference over the space of all possible topologies, and provides a systematic solution to the problem of perceptual aliasing in the domain of topological mapping. In this paper, we describe a general framework for modeling measurements, and the use of a Markov-chain Monte Carlo algorithm that uses specific instances of these models for odometry and appearance measurements to estimate the posterior distribution. We present experimental results that validate our technique and generate good maps when using odometry and appearance, derived from panoramic images, as sensor measurements. Index Terms—Bayesian inference, Markov-chain Monte Carlo (MCMC), mobile robots, perceptual aliasing, probability distributions, sample-based representations, topological maps. I.

A robotic chauffeur should reason about spatial information with a variety of scales, dimensions, and ontologies. Rich representations of both the quantitative and qualitative characteristics of space not only enable robust navigation behavior, but also permit natural communication with a human passenger. We apply a hierarchical framework of spatial knowledge inspired by human cognitive abilities, the Hybrid Spatial Semantic Hierarchy, to common navigation tasks: safe motion, localization, map-building, and route planning. We also discuss the straightforward mapping between the variety of ways in which people communicate with a chauffeur and the framework’s heterogeneous concepts of spatial knowledge. We present pilot experiments with a virtual chauffeur.