This paper presents an approach and introduces new open-source tools that can be used to evaluate robotic mapping algorithms. Also described is an extensive subterranean mine rescue dataset based upon the DARPA Subterranean (SubT) challenge including professionally surveyed ground truth. Finally, some commonly available approaches are evaluated using this metric.

Despite their attractiveness, popular perception is that techniques for nonparametric function approximation do not scale to streaming data due to an intractable growth in the amount of storage they require. To solve this problem in a memory-affordable way, we propose an online technique based on functional stochastic gradient descent in tandem with supervised sparsification based on greedy function subspace projections. The method, called parsimonious online learning with kernels (POLK), provides a controllable tradeoff between its solution accuracy and the amount of memory it requires. We derive conditions under which the generated function sequence converges almost surely to the optimal function, and we establish that the memory requirement remains finite. We evaluate POLK for kernel multi-class logistic regression and kernel hinge-loss classification on three canonical data sets: a synthetic Gaussian mixture model, the MNIST hand-written digits, and the Brodatz texture database. On all three tasks, we observe a favorable trade-off of objective function evaluation, classification performance, and complexity of the nonparametric regressor extracted by the proposed method.

We present a reinforcement learning algorithm for learning sparse non-parametric controllers in a Reproducing Kernel Hilbert Space. We improve the sample complexity of this approach by imposing a structure of the state-action function through a normalized advantage function (NAF). This representation of the policy enables efficiently composing multiple learned models without additional training samples or interaction with the environment. We demonstrate the performance of this algorithm on learning obstacle-avoidance policies in multiple simulations of a robot equipped with a laser scanner while navigating in a 2D environment. We apply the composition operation to various policy combinations and test them to show that the composed policies retain the performance of their components. We also transfer the composed policy directly to a physical platform operating in an arena with obstacles in order to demonstrate a degree of generalization.

We consider stochastic nonparametric regression problems in a reproducing kernel Hilbert space (RKHS), an extension of expected risk minimization to nonlinear function estimation. Popular perception is that kernel methods are inapplicable to online settings, since the generalization of stochastic methods to kernelized function spaces require memory storage that is cubic in the iteration index (“the curse of kernelization”). We alleviate this intractability in two ways: (1) we consider the use of functional stochastic gradient method (FSGD) which operates on a subset of training examples at each step; and (2), we extract parsimonious approximations of the resulting stochastic sequence via a greedy sparse subspace projection scheme based on kernel orthogonal matching pursuit (KOMP). We establish that this method converges almost surely in both diminishing and constant algorithm step-size regimes for a specific selection of sparse approximation budget. The method is evaluated on a kernel multi-class support vector machine problem, where data samples are generated from class-dependent Gaussian mixture models.

The output encodings of neural nets determine the structure of the space in which inference occurs. Yet, they are generally given very little thought. It is common practice for neural nets to use 1-Hot encoding when training to discriminate among many classes. The primary exceptions to this are error correcting output codes, and semantic output encodings. Output encodings based upon semantic descriptors cause a net to learn responses for classes to which it has not been exposed, provided those classes may be characterized by the same semantic descriptors. This raises a number of questions, such as “Can a net implicitly learn encodings for unobserved classes in the absence of a semantic encoding, or any encoding that requires some form of prior knowledge and hand crafting?”. Also, are some output encodings better than others for learning these implicit encodings? In this paper, we will compare how effectively different non-semantic encodings are at causing a neural net to implicitly learn encodings for unobserved classes. Also, while evaluating the efficacy of these implicit encodings, we will look for evidence of a phenomenon akin to over-training. Specifically, as training on the observed classes occurs, we initially see improvement in how well the implicitly learned encodings can be used to differentiate among the classes which are unobserved during the net's training. However, as training continues to improve discrimination among the observed classes, the efficacy of the implicit codes either remains steady, or undergoes a degradation. This degradation is akin to the overtraining that one generally tries to guard against when training a neural net.

In pursuit of increasing the operational tempo of a ground robotics platform in unknown domains, we consider the problem of predicting the distribution of structural state-estimation error due to poorly-modeled platform dynamics as well as environmental effects. Such predictions are a critical component of any modern control approach that utilizes uncertainty information to provide robustness in control design. We use an online learning algorithm based on matrix factorization techniques to fit a statistical model of error that provides enough expressive power to enable prediction directly from motion control signals and low-level visual features. Moreover, we empirically demonstrate that this technique compares favorably to predictors that do not incorporate this information.

Historically, neural nets have learned new things at the cost of forgetting what they already know. This problem is known as 'catastrophic forgetting'. Here, we examine how training a neural net in accordance with latently learned [1] output encodings drastically reduces catastrophic forgetting. Previous approaches to dealing with catastrophic forgetting have tended either to add extra samples to new training sets, modify the training of hidden nodes or model the interaction between short term and long term memory. Our approach is unique in that it both uses transfer learning to mitigate catastrophic forgetting and focuses upon the output nodes of a neural network. This results in a technique that makes it easier rather than harder to learn new tasks while retaining existing knowledge; is architecture independent and trivial to implement on any existing net. Additionally, we examine the use of ternary output codes. Binary codes assign a value to each output bit that may be thought of as either affirmative or negative. Ternary codes allow for the possibility that not every output bit has a meaningful response to every given input. By not forcing each output bit to train for a specific response for each new class, we hope to lessen catastrophic forgetting.

We consider discriminative dictionary learning in a distributed online setting, where a team of networked robots aims to jointly learn both a common basis of the feature space and a classifier over this basis from sequentially observed signals. We formulate this problem as a distributed stochastic program with a non-convex objective and present a block variant of the Arrow-Hurwicz saddle point algorithm to solve it. Only neighboring nodes in the communications network need to exchange information, and we penalize the discrepency between the individual feature basis and classifiers using Lagrange multipliers. The application we consider is for a team of robots to collaboratively recognize objects of interest in dynamic environments. As a preliminary performance benchmark, we consider the problem of learning a texture classifier across a network of robots moving around an urban setting where separate training examples are sequentially observed at each robot. Results are shown for both a standard texture dataset and a new dataset from an urban training facility, and we compare the performance of the standard centralized construction to the new distributed algorithm for the case when distinct samples from all classes are seen by the robots. These experiments yield comparable performance between the decentralized and the centralized cases, demonstrating the proposed method's practical utility.

We begin to bridge the gap between high-level motion planning and execution by adopting models to abstract the complicated skid-steer vehicle dynamics and evaluating their suitability as motion predictors for a feed-forward control framework. We consider three kinematic motion models and a drivetrain model in experiments on two surface types with a small tracked vehicle. We perform statistical analysis of the predictive accuracy of these models when used to create optimal open-loop plans for a set of canonical maneuvers and discuss the applicability of these models for a closed-loop control framework.

We consider the problem of deploying robots in formations that ensure network connectivity between a fixed base station and a set of independent agents wandering in the environment. We adopt a communications model that requires line-of-sight and then solve for robot placements by finding mutually-visible configurations in a polygonal decomposition of the environment map. Both the static deployment case and the case of finding deployments that minimize total robot movement are considered. We provide algorithms for the moving agent case, consider their performance on various discretizations for a range of problem sizes, and discuss our experimental implementation of the presented ideas.

In this paper, we focus on the detailing of a system architecture capable of addressing the problem of persistent surveillance with a team of autonomous micro-aerial vehicles (MAVs). We detail the problem of interest, discuss system requirements, and provide an overview of our approach. The remainder of the paper is dedicated to the system design and evaluation on a team of quadrotors in simulation and experiments.

Tele-operation of robotic platforms has long suffered from the inability of the operator to effectively perform ancillary tasks while controlling the robot. Because of the focus required to perform tele-operation, the operator is limited in their ability to use the robot to improve their situational awareness, with tele-operation often becoming a detriment to their task rather than an enhancement. We present experimental results on the use of a tele-operated robotic system modified to include layered obstacle detection and avoidance routines and an open space planner. These algorithms help the operator shift their focus towards high-level tasks instead of low-level navigation. While experimental data shows that the increase in autonomy did not lead to a reduction in task completion time, obstacle collisions were reduced and has led to further investigation of cognitive load reduction during operation.

We develop a framework for controlling a team of robots to maintain and improve a communication bridge between a stationary robot and an independently exploring robot in a walled environment. We make use of two metrics for characterizing the communication: the Fiedler value of the weighted Laplacian describing the communication interactions of all the robots in the system, and the k-connectivity matrix that expresses which robots can interact through k or less intermediary robots. At each step, we move in such a way as to improve the Fiedler value as much as possible while keeping the number of intermediary robots between the two robots of interest below a desired value. We demonstrate the use of this framework in a scenario where the hop-count constraint cannot be satisfied, but show that communication quality is maintained anyways.