Matthew R. Walter

Anytime Motion Planning using the RRT

The Rapidly-exploring Random Tree (RRT) algorithm, based on incremental sampling, efficiently computes motion plans. Although the RRT algorithm quickly produces candidate feasible solutions, it tends to converge to a solution that is far from optimal. Practical applications favor “anytime” algorithms that quickly identify an initial feasible plan, then, given more computation time available during plan execution, improve the plan toward an optimal solution. This paper describes an anytime algorithm based on the RRT* which (like the RRT) finds an initial feasible solution quickly, but (unlike the RRT) almost surely converges to an optimal solution. We present two key extensions to the RRT*, committed trajectories and branch-and-bound tree adaptation, that together enable the algorithm to make more efficient use of computation time online, resulting in an anytime algorithm for real-time implementation. We evaluate the method using a series of Monte Carlo runs in a high-fidelity simulation environment, and compare the operation of the RRT and RRT* methods. We also demonstrate experimental results for an outdoor wheeled

Exactly Sparse Extended Information Filters for Feature-based SLAM

Recent research concerning the Gaussian canonical form for Simultaneous Localization and Mapping (SLAM) has given rise to a handful of algorithms that attempt to solve the SLAM scalability problem for arbitrarily large environments. One such estimator that has received due attention is the Sparse Extended Information Filter (SEIF) proposed by Thrun et al., which is reported to be nearly constant time, irrespective of the size of the map. The key to the SEIFs scalability is to prune weak links in what is a dense information (inverse covariance) matrix to achieve a sparse approximation that allows for efficient, scalable SLAM. We demonstrate that the SEIF sparsification strategy yields error estimates that are overconfident when expressed in the global reference frame, while empirical results show that relative map consistency is maintained. In this paper, we propose an alternative scalable estimator based on an information form that maintains sparsity while preserving consistency. The paper describes a method for controlling the population of the information matrix, whereby we track a modified version of the SLAM posterior, essentially by ignoring a small fraction of temporal measurements. In this manner, the Exactly Sparse Extended Information Filter (ESEIF) performs inference over a model that is conservative relative to the standard Gaussian distribution. We compare our algorithm to the SEIF and standard EKF both in simulation as well as on two nonlinear datasets. The results convincingly show that our method yields conservative estimates for the robot pose and map that are nearly identical to those of the EKF.

Visually Navigating the RMS Titanic with SLAM Information Filters

This paper describes a vision-based large-area simultaneous localization and mapping (SLAM) algorithm that respects the constraints of low-overlap imagery typical of underwater vehicles while exploiting the information associated with the inertial sensors that are routinely available on such platforms. We present a novel strategy for efficiently accessing and maintaining consistent covariance bounds within a SLAM information filter, greatly increasing the reliability of data association. The technique is based upon solving a sparse system of linear equations coupled with the application of constant-time Kalman updates. The method is shown to produce consistent covariance estimates suitable for robot planning and data association. Realworld results are presented for a vision-based 6 DOF SLAM implementation using data from a recent ROV survey of the wreck of the RMS Titanic.

Visually Mapping the RMS Titanic: Conservative Covariance Estimates for SLAM Information Filters

This paper describes a vision-based, large-area, simultaneous localization and mapping (SLAM) algorithm that respects the low-overlap imagery constraints typical of underwater vehicles while exploiting the inertial sensor information that is routinely available on such platforms. We present a novel strategy for efficiently accessing and maintaining consistent covariance bounds within a SLAM information filter, thereby greatly increasing the reliability of data association. The technique is based upon solving a sparse system of linear equations coupled with the application of constant-time Kalman updates. The method is shown to produce consistent covariance estimates suitable for robot planning and data association. Real-world results are reported for a vision-based, six degree of freedom SLAM implementation using data from a recent survey of the wreck of the RMS Titanic.

Sparse extended information filters: insights into sparsification

Recently, there have been a number of variant simultaneous localization and mapping (SLAM) algorithms that have made substantial progress towards large-area scalability by parameterizing the SLAM posterior within the information (canonical/inverse covariance) form. Of these, probably the most well known and popular approach is the sparse extended information filter (SEIF) by Thrun et al. While SEIFs have been successfully implemented with a variety of challenging real world datasets and have led to new insights into scalable SLAM, open research questions remain regarding the approximate sparsification procedure and its effect on map error consistency. In this paper, we examine the constant time SEIF sparsification procedure in depth and offer new insight into issues of consistency. In particular, we show that exaggerated map inconsistency occurs within the global reference frame where estimation is performed, but that empirical testing shows that relative local map relationships are preserved. We then present a slightly modified version of their sparsification procedure, which is shown to preserve sparsity while also generating both local and global map estimates comparable to those obtained by the nonsparsified SLAM filter. While this modified approximation is no longer constant time, it does serve as a theoretical benchmark against which to compare SEIFs constant time results. We demonstrate our findings by benchmark comparison of the modified and original SEIF sparsification rule using simulation in the linear Gaussian SLAM case and real world experiments for a nonlinear dataset.

Approaching the Symbol Grounding Problem with Probabilistic Graphical Models

n order for robots to engage in dialog with human teammates, they must have the ability to map between words in the language and aspects of the external world. A solution to this symbol grounding problem (Harnad, 1990) would enable a robot to interpret commands such as “Drive over to receiving and pick up the tire pallet.” In this article we describe several of our results that use probabilistic inference to address the symbol grounding problem. Our specific approach is to develop models that factor according to the linguistic structure of a command. We first describe an early result, a generative model that factors according to the sequential structure of language, and then discuss our new framework, generalized grounding graphs (G3). The G3 framework dynamically instantiates a probabilistic graphical model for a natural language input, enabling a mapping between words in language and concrete objects, places, paths and events in the external world. We report on corpus-based experiments where the robot is able to learn and use word meanings in three real-world tasks: indoor navigation, spatial language video retrieval, and mobile manipulation.

A voice-commandable robotic forklift working alongside humans in minimally-prepared outdoor environments

One long-standing challenge in robotics is the realization of mobile autonomous robots able to operate safely in existing human workplaces in a way that their presence is accepted by the human occupants. We describe the development of a multi-ton robotic forklift intended to operate alongside human personnel, handling palletized materials within existing, busy, semi-structured outdoor storage facilities.

SLAM for ship hull inspection using exactly sparse extended information filters

Many important missions for autonomous underwater vehicles (AUVs), such as undersea inspection of ship hulls, require integrated navigation, control, and motion planning in complex, 3D environments. This paper describes a SLAM implementation using forward-looking sonar (FLS) data from a highly maneuverable, hovering AUV performing a ship hull inspection mission. The exactly sparse extended information filter (ESEIF) algorithm is applied to perform SLAM based upon features manually selected within FLS images. The results demonstrate the ability to effectively map a ship hull in a challenging marine environment. This provides a foundation for future work in which real-time SLAM will be integrated with motion planning and control to achieve autonomous coverage of a complete ship hull.

A provably consistent method for imposing sparsity in feature-based SLAM information filters

An open problem in Simultaneous Localization and Mapping (SLAM) is the development of algorithms which scale with the size of the environment. A few promising methods exploit the key insight that representing the posterior in the canonical form parameterized by a sparse information matrix provides significant advantages regarding computational efficiency and storage requirements. Because the information matrix is naturally dense in the case of feature-based SLAM, additional steps are necessary to achieve sparsity. The delicate issue then becomes one of performing this sparsification in a manner which is consistent with the original distribution.

What to talk about and how? Selective Generation using LSTMs with Coarse-to-Fine Alignment

We propose an end-to-end, domain-independent neural encoder-aligner-decoder model for selective generation, i.e., the joint task of content selection and surface realization. Our model first encodes a full set of over-determined database event records via an LSTM-based recurrent neural network, then utilizes a novel coarse-to-fine aligner to identify the small subset of salient records to talk about, and finally employs a decoder to generate free-form descriptions of the aligned, selected records. Our model achieves the best selection and generation results reported to-date (with 59% relative improvement in generation) on the benchmark WeatherGov dataset, despite using no specialized features or linguistic resources. Using an improved k-nearest neighbor beam filter helps further. We also perform a series of ablations and visualizations to elucidate the contributions of our key model components. Lastly, we evaluate the generalizability of our model on the RoboCup dataset, and get results that are competitive with or better than the state-of-the-art, despite being severely data-starved.