Metric ground navigation addresses the problem of autonomously moving a robot from one point to another in an obstacle-occupied planar environment in a collision-free manner. It is one of the most fundamental capabilities of intelligent mobile robots. This paper presents a standardized testbed with a set of environments and metrics to benchmark difficulty of different scenarios and performance of different systems of metric ground navigation. Current benchmarks focus on individual components of mobile robot navigation, such as perception and state estimation, but the navigation performance as a whole is rarely measured in a systematic and standardized fashion. As a result, navigation systems are usually tested and compared in an ad hoc manner, such as in one or two manually chosen environments. The introduced benchmark provides a general testbed for ground robot navigation in a metric world. The dataset includes 300 navigation environments, which are ordered by a set of difficulty metrics. Navigation performance can be tested and compared in those environments in a systematic and objective fashion. This benchmark can be used to predict navigation difficulty of a new environment, compare navigation systems, and potentially serve as a cost function and a curriculum for planning-based and learning-based navigation systems. We have published our dataset and the source code to generate datasets for different robot footprints at www.cs.utexas.edu/~attruong/metrics_dataset.html.
Temporal difference (TD) learning is one of the main foundations of modern reinforcement learning. This paper studies the use of TD(0), a canonical TD algorithm, to estimate the value function of a given policy from a batch of data. In this batch setting, we show that TD(0) may converge to an inaccurate value function because the update following an action is weighted according to the number of times that action occurred in the batch -- not the true probability of the action under the given policy. To address this limitation, we introduce \textit{policy sampling error corrected}-TD(0) (PSEC-TD(0)). PSEC-TD(0) first estimates the empirical distribution of actions in each state in the batch and then uses importance sampling to correct for the mismatch between the empirical weighting and the correct weighting for updates following each action. We refine the concept of a certainty-equivalence estimate and argue that PSEC-TD(0) is a more data efficient estimator than TD(0) for a fixed batch of data. Finally, we conduct an empirical evaluation of PSEC-TD(0) on three batch value function learning tasks, with a hyperparameter sensitivity analysis, and show that PSEC-TD(0) produces value function estimates with lower mean squared error than TD(0).
The sim to real transfer problem deals with leveraging large amounts of inexpensive simulation experience to help artificial agents learn behaviors intended for the real world more efficiently. One approach to sim-to-real transfer is using interactions with the real world to make the simulator more realistic, called grounded sim to-real transfer. In this paper, we show that a particular grounded sim-to-real approach, grounded action transformation, is closely related to the problem of imitation from observation IfO, learning behaviors that mimic the observations of behavior demonstrations. After establishing this relationship, we hypothesize that recent state-of-the-art approaches from the IfO literature can be effectively repurposed for such grounded sim-to-real transfer. To validate our hypothesis we derive a new sim-to-real transfer algorithm - generative adversarial reinforced action transformation (GARAT) - based on adversarial imitation from observation techniques. We run experiments in several simulation domains with mismatched dynamics, and find that agents trained with GARAT achieve higher returns in the real world compared to existing black-box sim-to-real methods
Robot control policies learned in simulation do not often transfer well to the real world. Many existing solutions to this sim-to-real problem, such as the Grounded Action Transformation (GAT) algorithm, seek to correct for or ground these differences by matching the simulator to the real world. However, the efficacy of these approaches is limited if they do not explicitly account for stochasticity in the target environment. In this work, we analyze the problems associated with grounding a deterministic simulator in a stochastic real world environment, and we present examples where GAT fails to transfer a good policy due to stochastic transitions in the target domain. In response, we introduce the Stochastic Grounded Action Transformation(SGAT) algorithm,which models this stochasticity when grounding the simulator. We find experimentally for both simulated and physical target domains that SGAT can find policies that are robust to stochasticity in the target domain
Robots can learn to do complex tasks in simulation, but often, learned behaviors fail to transfer well to the real world due to simulator imperfections (the reality gap). Some existing solutions to this sim-to-real problem, such as Grounded Action Transformation (GAT), use a small amount of real-world experience to minimize the reality gap by grounding the simulator. While very effective in certain scenarios, GAT is not robust on problems that use complex function approximation techniques to model a policy. In this paper, we introduce Reinforced Grounded Action Transformation(RGAT), a new sim-to-real technique that uses Reinforcement Learning (RL) not only to update the target policy in simulation, but also to perform the grounding step itself. This novel formulation allows for end-to-end training during the grounding step, which, compared to GAT, produces a better grounded simulator. Moreover, we show experimentally in several MuJoCo domains that our approach leads to successful transfer for policies modeled using neural networks.
While classical approaches to autonomous robot navigation currently enable operation in certain environments, they break down in tightly constrained spaces, e.g., where the robot needs to engage in agile maneuvers to squeeze between obstacles. Recent machine learning techniques have the potential to address this shortcoming, but existing approaches require vast amounts of navigation experience for training, during which the robot must operate in close proximity to obstacles and risk collision. In this paper, we propose to side-step this requirement by introducing a new machine learning paradigm for autonomous navigation called learning from hallucination (LfH), which can use training data collected in completely safe environments to compute navigation controllers that result in fast, smooth, and safe navigation in highly constrained environments. Our experimental results show that the proposed LfH system outperforms three autonomous navigation baselines on a real robot, including those based on both classical and machine learning techniques (anonymized video: https://tinyurl.com/corl20lfh).
This paper presents a continually self-improving lifelong learning framework for a mobile robot navigating in different environments. Classical static navigation methods require environment-specific in-situ system adjustment, e.g. from human experts, or may repeat their mistakes regardless of how many times they have navigated in the same environment. Having the potential to improve with experience, learning-based navigation is highly dependent on access to training resources, e.g. sufficient memory and fast computation, and is prone to forgetting previously learned capability, especially when facing different environments. In this work, we propose a lifelong learning framework for mobile robot navigation which (1) improves from its own experience without expert demonstrations, and (2) retains capability to navigate in previous environments after learning in new ones. This framework is implemented and tested entirely onboard a physical robot with a limited memory and computation budget.
In continuing tasks, average-reward reinforcement learning may be a more appropriate problem formulation than the more common discounted reward formulation. As usual, learning an optimal policy in this setting typically requires a large amount of training experiences. Reward shaping is a common approach for incorporating domain knowledge into reinforcement learning in order to speed up convergence to an optimal policy. However, to the best of our knowledge, the theoretical properties of reward shaping have thus far only been established in the discounted setting. This paper presents the first reward shaping framework for average-reward learning and proves that, under standard assumptions, the optimal policy under the original reward function can be recovered. In order to avoid the need for manual construction of the shaping function, we introduce a method for utilizing domain knowledge expressed as a temporal logic formula. The formula is automatically translated to a shaping function that provides additional reward throughout the learning process. We evaluate the proposed method on three continuing tasks. In all cases, shaping speeds up the average-reward learning rate without any reduction in the performance of the learned policy compared to relevant baselines.