Jonathan Ko

Distributed Multirobot Exploration and Mapping

Efficient exploration of unknown environments is a fundamental problem in mobile robotics. We present an approach to distributed multirobot mapping and exploration. Our system enables teams of robots to efficiently explore environments from different, unknown locations. In order to ensure consistency when combining their data into shared maps, the robots actively seek to verify their relative locations. Using shared maps, they coordinate their exploration strategies to maximize the efficiency of exploration. This system was evaluated under extremely realistic real-world conditions. An outside evaluation team found the system to be highly efficient and robust. The maps generated by our approach are consistently more accurate than those generated by manually measuring the locations and extensions of rooms and objects

GP-BayesFilters: Bayesian filtering using Gaussian process prediction and observation models

Bayesian filtering is a general framework for recursively estimating the state of a dynamical system. Key components of each Bayes filter are probabilistic prediction and observation models. This paper shows how non-parametric Gaussian process (GP) regression can be used for learning such models from training data. We also show how Gaussian process models can be integrated into different versions of Bayes filters, namely particle filters and extended and unscented Kalman filters. The resulting GP-BayesFilters can have several advantages over standard (parametric) filters. Most importantly, GP-BayesFilters do not require an accurate, parametric model of the system. Given enough training data, they enable improved tracking accuracy compared to parametric models, and they degrade gracefully with increased model uncertainty. These advantages stem from the fact that GPs consider both the noise in the system and the uncertainty in the model. If an approximate parametric model is available, it can be incorporated into the GP, resulting in further performance improvements. In experiments, we show different properties of GP-BayesFilters using data collected with an autonomous micro-blimp as well as synthetic data.

A practical, decision-theoretic approach to multi-robot mapping and exploration

An important assumption underlying virtually all approaches to multi-robot exploration is prior knowledge about their relative locations. This is due to the fact that robots need to merge their maps so as to coordinate their exploration strategies. The key step in map merging is to estimate the relative locations of the individual robots. This paper presents a novel approach to multi-robot map merging under global uncertainty about the robots relative locations. Our approach uses an adapted version of particle filters to estimate the position of one robot in the other robots partial map. The risk of false-positive map matches is avoided by verifying match hypotheses using a rendezvous approach. We show how to seamlessly integrate this approach into a decision-theoretic multi-robot coordination strategy. The experiments show that our sample-based technique can reliably find good hypotheses for map matches. Furthermore, we present results obtained with two robots successfully merging their maps using the decision-theoretic rendezvous strategy.

Gaussian Processes and Reinforcement Learning for Identification and Control of an Autonomous Blimp

Blimps are a promising platform for aerial robotics and have been studied extensively for this purpose. Unlike other aerial vehicles, blimps are relatively safe and also possess the ability to loiter for long periods. These advantages, however, have been difficult to exploit because blimp dynamics are complex and inherently non-linear. The classical approach to system modeling represents the system as an ordinary differential equation (ODE) based on Newtonian principles. A more recent modeling approach is based on representing state transitions as a Gaussian process (GP). In this paper, we present a general technique for system identification that combines these two modeling approaches into a single formulation. This is done by training a Gaussian process on the residual between the non-linear model and ground truth training data. The result is a GP-enhanced model that provides an estimate of uncertainty in addition to giving better state predictions than either ODE or GP alone. We show how the GP-enhanced model can be used in conjunction with reinforcement learning to generate a blimp controller that is superior to those learned with ODE or GP models alone.

Map merging for distributed robot navigation

A set of robots mapping an area can potentially combine their information to produce a distributed map more efficiently than a single robot alone. We describe a general framework for distributed map building in the presence of uncertain communication. Within this framework, we then present a technical solution to the key decision problem of determining relative location within partial maps.

Centibots: very large scale distributed robotic teams

We describe the development of Centibots, a framework for very large teams of robots that are able to perceive, explore, plan and collaborate in unknown environments.

GP-UKF: Unscented kalman filters with Gaussian process prediction and observation models

This paper considers the use of non-parametric system models for sequential state estimation. In particular, motion and observation models are learned from training examples using Gaussian process (GP) regression. The state estimator is an unscented Kalman filter (UKF). The resulting GP-UKF algorithm has a number of advantages over standard (parametric) UKFs. These include the ability to estimate the state of arbitrary nonlinear systems, improved tracking quality compared to a parametric UKF, and graceful degradation with increased model uncertainty. These advantages stem from the fact that GPs consider both the noise in the system and the uncertainty in the model. If an approximate parametric model is available, it can be incorporated into the GP; resulting in further performance improvements. In experiments, we show how the GP-UKF algorithm can be applied to the problem of tracking an autonomous micro-blimp.

Mechanisms of the Anatomically Correct Testbed Hand

We have built an anatomically correct testbed (ACT) hand with the purpose of understanding the intrinsic biomechanical and control features in human hands that are critical for achieving robust, versatile, and dexterous movements, as well as rich object and world exploration. By mimicking the underlying mechanics and controls of the human hand in a hardware platform, our goal is to achieve previously unmatched grasping and manipulation skills. In this paper, the novel constituting mechanisms, unique muscle to joint relationships, and movement demonstrations of the thumb, index finger, middle finger, and wrist of the ACT Hand are presented. The grasping and manipulation abilities of the ACT Hand are also illustrated. The fully functional ACT Hand platform allows for the possibility to design and experiment with novel control algorithms leading to a deeper understanding of human dexterity.

Learning GP-BayesFilters via Gaussian process latent variable models

GP-BayesFilters are a general framework for integrating Gaussian process prediction and observation models into Bayesian filtering techniques, including particle filters and extended and unscented Kalman filters. GP-BayesFilters have been shown to be extremely well suited for systems for which accurate parametric models are difficult to obtain. GP-BayesFilters learn non-parametric models from training data containing sequences of control inputs, observations, and ground truth states. The need for ground truth states limits the applicability of GP-BayesFilters to systems for which the ground truth can be estimated without significant overhead. In this paper we introduce GPBF-Learn, a framework for training GP-BayesFilters without ground truth states. Our approach extends Gaussian Process Latent Variable Models to the setting of dynamical robotics systems. We show how weak labels for the ground truth states can be incorporated into the GPBF-Learn framework. The approach is evaluated using a difficult tracking task, namely tracking a slotcar based on inertial measurement unit (IMU) observations only. We also show some special features enabled by this framework, including time alignment, and control replay for both the slotcar, and a robotic arm.

Distributed multirobot exploration, mapping, and task allocation

We present an integrated approach to multirobot exploration, mapping and searching suitable for large teams of robots operating in unknown areas lacking an existing supporting communications infrastructure. We present a set of algorithms that have been both implemented and experimentally verified on teams—of what we refer to as Centibots—consisting of as many as 100 robots. The results that we present involve search tasks that can be divided into a mapping stage in which robots must jointly explore a large unknown area with the goal of generating a consistent map from the fragment, a search stage in which robots are deployed within the environment in order to systematically search for an object of interest, and a protection phase in which robots are distributed to track any intruders in the search area. During the first stage, the robots actively seek to verify their relative locations in order to ensure consistency when combining data into shared maps; they must also coordinate their exploration strategies so as to maximize the efficiency of exploration. In the second and third stages, robots allocate search tasks among themselves; since tasks are not defined a priori, the robots first produce a topological graph of the area of interest and then generate a set of tasks that reflect spatial and communication constraints. Our system was evaluated under extremely realistic real-world conditions. An outside evaluation team found the system to be highly efficient and robust.