Garrett Warnell

Haresh Karnan, Anirudh Nair, Xuesu Xiao et al.

Social navigation is the capability of an autonomous agent, such as a robot, to navigate in a 'socially compliant' manner in the presence of other intelligent agents such as humans. With the emergence of autonomously navigating mobile robots in human populated environments (e.g., domestic service robots in homes and restaurants and food delivery robots on public sidewalks), incorporating socially compliant navigation behaviors on these robots becomes critical to ensuring safe and comfortable human robot coexistence. To address this challenge, imitation learning is a promising framework, since it is easier for humans to demonstrate the task of social navigation rather than to formulate reward functions that accurately capture the complex multi objective setting of social navigation. The use of imitation learning and inverse reinforcement learning to social navigation for mobile robots, however, is currently hindered by a lack of large scale datasets that capture socially compliant robot navigation demonstrations in the wild. To fill this gap, we introduce Socially CompliAnt Navigation Dataset (SCAND) a large scale, first person view dataset of socially compliant navigation demonstrations. Our dataset contains 8.7 hours, 138 trajectories, 25 miles of socially compliant, human teleoperated driving demonstrations that comprises multi modal data streams including 3D lidar, joystick commands, odometry, visual and inertial information, collected on two morphologically different mobile robots a Boston Dynamics Spot and a Clearpath Jackal by four different human demonstrators in both indoor and outdoor environments. We additionally perform preliminary analysis and validation through real world robot experiments and show that navigation policies learned by imitation learning on SCAND generate socially compliant behaviors

Haresh Karnan, Elvin Yang, Daniel Farkash et al.

Terrain awareness, i.e., the ability to identify and distinguish different types of terrain, is a critical ability that robots must have to succeed at autonomous off-road navigation. Current approaches that provide robots with this awareness either rely on labeled data which is expensive to collect, engineered features and cost functions that may not generalize, or expert human demonstrations which may not be available. Towards endowing robots with terrain awareness without these limitations, we introduce Self-supervised TErrain Representation LearnING (STERLING), a novel approach for learning terrain representations that relies solely on easy-to-collect, unconstrained (e.g., non-expert), and unlabelled robot experience, with no additional constraints on data collection. STERLING employs a novel multi-modal self-supervision objective through non-contrastive representation learning to learn relevant terrain representations for terrain-aware navigation. Through physical robot experiments in off-road environments, we evaluate STERLING features on the task of preference-aligned visual navigation and find that STERLING features perform on par with fully supervised approaches and outperform other state-of-the-art methods with respect to preference alignment. Additionally, we perform a large-scale experiment of autonomously hiking a 3-mile long trail which STERLING completes successfully with only two manual interventions, demonstrating its robustness to real-world off-road conditions.

Caroline Wang, Garrett Warnell, Peter Stone

While combining imitation learning (IL) and reinforcement learning (RL) is a promising way to address poor sample efficiency in autonomous behavior acquisition, methods that do so typically assume that the requisite behavior demonstrations are provided by an expert that behaves optimally with respect to a task reward. If, however, suboptimal demonstrations are provided, a fundamental challenge appears in that the demonstration-matching objective of IL conflicts with the return-maximization objective of RL. This paper introduces D-Shape, a new method for combining IL and RL that uses ideas from reward shaping and goal-conditioned RL to resolve the above conflict. D-Shape allows learning from suboptimal demonstrations while retaining the ability to find the optimal policy with respect to the task reward. We experimentally validate D-Shape in sparse-reward gridworld domains, showing that it both improves over RL in terms of sample efficiency and converges consistently to the optimal policy in the presence of suboptimal demonstrations.

Eddy Hudson, Ishan Durugkar, Garrett Warnell et al.

Given a dataset of expert agent interactions with an environment of interest, a viable method to extract an effective agent policy is to estimate the maximum likelihood policy indicated by this data. This approach is commonly referred to as behavioral cloning (BC). In this work, we describe a key disadvantage of BC that arises due to the maximum likelihood objective function; namely that BC is mean-seeking with respect to the state-conditional expert action distribution when the learner's policy is represented with a Gaussian. To address this issue, we introduce a modified version of BC, Adversarial Behavioral Cloning (ABC), that exhibits mode-seeking behavior by incorporating elements of GAN (generative adversarial network) training. We evaluate ABC on toy domains and a domain based on Hopper from the DeepMind Control suite, and show that it outperforms standard BC by being mode-seeking in nature.

Haresh Karnan, Elvin Yang, Garrett Warnell et al.

Autonomous mobility tasks such as lastmile delivery require reasoning about operator indicated preferences over terrains on which the robot should navigate to ensure both robot safety and mission success. However, coping with out of distribution data from novel terrains or appearance changes due to lighting variations remains a fundamental problem in visual terrain adaptive navigation. Existing solutions either require labor intensive manual data recollection and labeling or use handcoded reward functions that may not align with operator preferences. In this work, we posit that operator preferences for visually novel terrains, which the robot should adhere to, can often be extrapolated from established terrain references within the inertial, proprioceptive, and tactile domain. Leveraging this insight, we introduce Preference extrApolation for Terrain awarE Robot Navigation, PATERN, a novel framework for extrapolating operator terrain preferences for visual navigation. PATERN learns to map inertial, proprioceptive, tactile measurements from the robots observations to a representation space and performs nearest neighbor search in this space to estimate operator preferences over novel terrains. Through physical robot experiments in outdoor environments, we assess PATERNs capability to extrapolate preferences and generalize to novel terrains and challenging lighting conditions. Compared to baseline approaches, our findings indicate that PATERN robustly generalizes to diverse terrains and varied lighting conditions, while navigating in a preference aligned manner.

Dongmyeong Lee, Jesse Quattrociocchi, Christian Ellis et al.

We propose BEV-Patch-PF, a GPS-free sequential geo-localization system that integrates a particle filter with learned bird's-eye-view (BEV) and aerial feature maps. From onboard RGB and depth images, we construct a BEV feature map. For each 3-DoF particle pose hypothesis, we crop the corresponding patch from an aerial feature map computed from a local aerial image queried around the approximate location. BEV-Patch-PF computes a per-particle log-likelihood by matching the BEV feature to the aerial patch feature. On two real-world off-road datasets, our method achieves 7.5x lower absolute trajectory error (ATE) on seen routes and 7.0x lower ATE on unseen routes than a retrieval-based baseline, while maintaining accuracy under dense canopy and shadow. The system runs in real time at 10 Hz on an NVIDIA Tesla T4, enabling practical robot deployment.

Yunshu Du, Garrett Warnell, Assefaw Gebremedhin et al.

Experience replay (ER) improves the data efficiency of off-policy reinforcement learning (RL) algorithms by allowing an agent to store and reuse its past experiences in a replay buffer. While many techniques have been proposed to enhance ER by biasing how experiences are sampled from the buffer, thus far they have not considered strategies for refreshing experiences inside the buffer. In this work, we introduce Lucid Dreaming for Experience Replay (LiDER), a conceptually new framework that allows replay experiences to be refreshed by leveraging the agent's current policy. LiDER consists of three steps: First, LiDER moves an agent back to a past state. Second, from that state, LiDER then lets the agent execute a sequence of actions by following its current policy -- as if the agent were "dreaming" about the past and can try out different behaviors to encounter new experiences in the dream. Third, LiDER stores and reuses the new experience if it turned out better than what the agent previously experienced, i.e., to refresh its memories. LiDER is designed to be easily incorporated into off-policy, multi-worker RL algorithms that use ER; we present in this work a case study of applying LiDER to an actor-critic based algorithm. Results show LiDER consistently improves performance over the baseline in six Atari 2600 games. Our open-source implementation of LiDER and the data used to generate all plots in this work are available at github.com/duyunshu/lucid-dreaming-for-exp-replay.

Luisa Mao, Garrett Warnell, Peter Stone et al.

In autonomous robot navigation, terrain cost assignment is typically performed using a semantics-based paradigm in which terrain is first labeled using a pre-trained semantic classifier and costs are then assigned according to a user-defined mapping between label and cost. While this approach is rapidly adaptable to changing user preferences, only preferences over the types of terrain that are already known by the semantic classifier can be expressed. In this paper, we hypothesize that a machine-learning-based alternative to the semantics-based paradigm above will allow for rapid cost assignment adaptation to preferences expressed over new terrains at deployment time without the need for additional training. To investigate this hypothesis, we introduce and study PACER, a novel approach to costmap generation that accepts as input a single birds-eye view (BEV) image of the surrounding area along with a user-specified preference context and generates a corresponding BEV costmap that aligns with the preference context. Using both real and synthetic data along with a combination of proposed training tasks, we find that PACER is able to adapt quickly to new user preferences while also exhibiting better generalization to novel terrains compared to both semantics-based and representation-learning approaches.

Mohammad Nazeri, Anuj Pokhrel, Alexandyr Card et al.

Sophisticated learning architectures, e.g., Transformers, present a unique opportunity for robots to understand complex vehicle-terrain kinodynamic interactions for off-road mobility. While internet-scale data are available for Natural Language Processing (NLP) and Computer Vision (CV) tasks to train Transformers, real-world mobility data are difficult to acquire with physical robots navigating off-road terrain. Furthermore, training techniques specifically designed to process text and image data in NLP and CV may not apply to robot mobility. In this paper, we propose VertiFormer, a novel data-efficient multi-task Transformer model trained with only one hour of data to address such challenges of applying Transformer architectures for robot mobility on extremely rugged, vertically challenging, off-road terrain. Specifically, VertiFormer employs a new learnable masked modeling and next token prediction paradigm to predict the next pose, action, and terrain patch to enable a variety of off-road mobility tasks simultaneously, e.g., forward and inverse kinodynamics modeling. The non-autoregressive design mitigates computational bottlenecks and error propagation associated with autoregressive models. VertiFormer's unified modality representation also enhances learning of diverse temporal mappings and state representations, which, combined with multiple objective functions, further improves model generalization. Our experiments offer insights into effectively utilizing Transformers for off-road robot mobility with limited data and demonstrate our efficiently trained Transformer can facilitate multiple off-road mobility tasks onboard a physical mobile robot.

Michael J. Munje, Chen Tang, Shuijing Liu et al.

Robot navigation in dynamic, human-centered environments requires socially-compliant decisions grounded in robust scene understanding. Recent Vision-Language Models (VLMs) exhibit promising capabilities such as object recognition, common-sense reasoning, and contextual understanding-capabilities that align with the nuanced requirements of social robot navigation. However, it remains unclear whether VLMs can accurately understand complex social navigation scenes (e.g., inferring the spatial-temporal relations among agents and human intentions), which is essential for safe and socially compliant robot navigation. While some recent works have explored the use of VLMs in social robot navigation, no existing work systematically evaluates their ability to meet these necessary conditions. In this paper, we introduce the Social Navigation Scene Understanding Benchmark (SocialNav-SUB), a Visual Question Answering (VQA) dataset and benchmark designed to evaluate VLMs for scene understanding in real-world social robot navigation scenarios. SocialNav-SUB provides a unified framework for evaluating VLMs against human and rule-based baselines across VQA tasks requiring spatial, spatiotemporal, and social reasoning in social robot navigation. Through experiments with state-of-the-art VLMs, we find that while the best-performing VLM achieves an encouraging probability of agreeing with human answers, it still underperforms simpler rule-based approach and human consensus baselines, indicating critical gaps in social scene understanding of current VLMs. Our benchmark sets the stage for further research on foundation models for social robot navigation, offering a framework to explore how VLMs can be tailored to meet real-world social robot navigation needs. An overview of this paper along with the code and data can be found at https://larg.github.io/socialnav-sub .

Zichao Hu, Chen Tang, Michael J. Munje et al.

This paper considers the problem of enabling robots to navigate dynamic environments while following instructions. The challenge lies in the combinatorial nature of instruction specifications: each instruction can include multiple specifications, and the number of possible specification combinations grows exponentially as the robot's skill set expands. For example, "overtake the pedestrian while staying on the right side of the road" consists of two specifications: "overtake the pedestrian" and "walk on the right side of the road." To tackle this challenge, we propose ComposableNav, based on the intuition that following an instruction involves independently satisfying its constituent specifications, each corresponding to a distinct motion primitive. Using diffusion models, ComposableNav learns each primitive separately, then composes them in parallel at deployment time to satisfy novel combinations of specifications unseen in training. Additionally, to avoid the onerous need for demonstrations of individual motion primitives, we propose a two-stage training procedure: (1) supervised pre-training to learn a base diffusion model for dynamic navigation, and (2) reinforcement learning fine-tuning that molds the base model into different motion primitives. Through simulation and real-world experiments, we show that ComposableNav enables robots to follow instructions by generating trajectories that satisfy diverse and unseen combinations of specifications, significantly outperforming both non-compositional VLM-based policies and costmap composing baselines. Videos and additional materials can be found on the project page: https://amrl.cs.utexas.edu/ComposableNav/

Haresh Karnan, Kavan Singh Sikand, Pranav Atreya et al.

One of the key challenges in high speed off road navigation on ground vehicles is that the kinodynamics of the vehicle terrain interaction can differ dramatically depending on the terrain. Previous approaches to addressing this challenge have considered learning an inverse kinodynamics (IKD) model, conditioned on inertial information of the vehicle to sense the kinodynamic interactions. In this paper, we hypothesize that to enable accurate high-speed off-road navigation using a learned IKD model, in addition to inertial information from the past, one must also anticipate the kinodynamic interactions of the vehicle with the terrain in the future. To this end, we introduce Visual-Inertial Inverse Kinodynamics (VI-IKD), a novel learning based IKD model that is conditioned on visual information from a terrain patch ahead of the robot in addition to past inertial information, enabling it to anticipate kinodynamic interactions in the future. We validate the effectiveness of VI-IKD in accurate high-speed off-road navigation experimentally on a scale 1/5 UT-AlphaTruck off-road autonomous vehicle in both indoor and outdoor environments and show that compared to other state-of-the-art approaches, VI-IKD enables more accurate and robust off-road navigation on a variety of different terrains at speeds of up to 3.5 m/s.

Haresh Karnan, Garrett Warnell, Faraz Torabi et al.

The imitation learning research community has recently made significant progress towards the goal of enabling artificial agents to imitate behaviors from video demonstrations alone. However, current state-of-the-art approaches developed for this problem exhibit high sample complexity due, in part, to the high-dimensional nature of video observations. Towards addressing this issue, we introduce here a new algorithm called Visual Generative Adversarial Imitation from Observation using a State Observer VGAIfO-SO. At its core, VGAIfO-SO seeks to address sample inefficiency using a novel, self-supervised state observer, which provides estimates of lower-dimensional proprioceptive state representations from high-dimensional images. We show experimentally in several continuous control environments that VGAIfO-SO is more sample efficient than other IfO algorithms at learning from video-only demonstrations and can sometimes even achieve performance close to the Generative Adversarial Imitation from Observation (GAIfO) algorithm that has privileged access to the demonstrator's proprioceptive state information.

Kavan Singh Sikand, Sadegh Rabiee, Adam Uccello et al.

Autonomous mobile robots deployed in outdoor environments must reason about different types of terrain for both safety (e.g., prefer dirt over mud) and deployer preferences (e.g., prefer dirt path over flower beds). Most existing solutions to this preference-aware path planning problem use semantic segmentation to classify terrain types from camera images, and then ascribe costs to each type. Unfortunately, there are three key limitations of such approaches -- they 1) require pre-enumeration of the discrete terrain types, 2) are unable to handle hybrid terrain types (e.g., grassy dirt), and 3) require expensive labelled data to train visual semantic segmentation. We introduce Visual Representation Learning for Preference-Aware Path Planning (VRL-PAP), an alternative approach that overcomes all three limitations: VRL-PAP leverages unlabeled human demonstrations of navigation to autonomously generate triplets for learning visual representations of terrain that are viewpoint invariant and encode terrain types in a continuous representation space. The learned representations are then used along with the same unlabeled human navigation demonstrations to learn a mapping from the representation space to terrain costs. At run time, VRL-PAP maps from images to representations and then representations to costs to perform preference-aware path planning. We present empirical results from challenging outdoor settings that demonstrate VRL-PAP 1) is successfully able to pick paths that reflect demonstrated preferences, 2) is comparable in execution to geometric navigation with a highly detailed manually annotated map (without requiring such annotations), 3) is able to generalize to novel terrain types with minimal additional unlabeled demonstrations.

Zizhao Wang, Xuesu Xiao, Garrett Warnell et al.

Classical autonomous navigation systems can control robots in a collision-free manner, oftentimes with verifiable safety and explainability. When facing new environments, however, fine-tuning of the system parameters by an expert is typically required before the system can navigate as expected. To alleviate this requirement, the recently-proposed Adaptive Planner Parameter Learning paradigm allows robots to \emph{learn} how to dynamically adjust planner parameters using a teleoperated demonstration or corrective interventions from non-expert users. However, these interaction modalities require users to take full control of the moving robot, which requires the users to be familiar with robot teleoperation. As an alternative, we introduce \textsc{apple}, Adaptive Planner Parameter Learning from \emph{Evaluative Feedback} (real-time, scalar-valued assessments of behavior), which represents a less-demanding modality of interaction. Simulated and physical experiments show \textsc{apple} can achieve better performance compared to the planner with static default parameters and even yield improvement over learned parameters from richer interaction modalities.

Ruohan Zhang, Faraz Torabi, Garrett Warnell et al.

A longstanding goal of artificial intelligence is to create artificial agents capable of learning to perform tasks that require sequential decision making. Importantly, while it is the artificial agent that learns and acts, it is still up to humans to specify the particular task to be performed. Classical task-specification approaches typically involve humans providing stationary reward functions or explicit demonstrations of the desired tasks. However, there has recently been a great deal of research energy invested in exploring alternative ways in which humans may guide learning agents that may, e.g., be more suitable for certain tasks or require less human effort. This survey provides a high-level overview of five recent machine learning frameworks that primarily rely on human guidance apart from pre-specified reward functions or conventional, step-by-step action demonstrations. We review the motivation, assumptions, and implementation of each framework, and we discuss possible future research directions.

Haresh Karnan, Garrett Warnell, Xuesu Xiao et al.

While imitation learning for vision based autonomous mobile robot navigation has recently received a great deal of attention in the research community, existing approaches typically require state action demonstrations that were gathered using the deployment platform. However, what if one cannot easily outfit their platform to record these demonstration signals or worse yet the demonstrator does not have access to the platform at all? Is imitation learning for vision based autonomous navigation even possible in such scenarios? In this work, we hypothesize that the answer is yes and that recent ideas from the Imitation from Observation (IfO) literature can be brought to bear such that a robot can learn to navigate using only ego centric video collected by a demonstrator, even in the presence of viewpoint mismatch. To this end, we introduce a new algorithm, Visual Observation only Imitation Learning for Autonomous navigation (VOILA), that can successfully learn navigation policies from a single video demonstration collected from a physically different agent. We evaluate VOILA in the photorealistic AirSim simulator and show that VOILA not only successfully imitates the expert, but that it also learns navigation policies that can generalize to novel environments. Further, we demonstrate the effectiveness of VOILA in a real world setting by showing that it allows a wheeled Jackal robot to successfully imitate a human walking in an environment using a video recorded using a mobile phone camera.

Xuesu Xiao, Zizhao Wang, Zifan Xu et al.

While current autonomous navigation systems allow robots to successfully drive themselves from one point to another in specific environments, they typically require extensive manual parameter re-tuning by human robotics experts in order to function in new environments. Furthermore, even for just one complex environment, a single set of fine-tuned parameters may not work well in different regions of that environment. These problems prohibit reliable mobile robot deployment by non-expert users. As a remedy, we propose Adaptive Planner Parameter Learning (APPL), a machine learning framework that can leverage non-expert human interaction via several modalities -- including teleoperated demonstrations, corrective interventions, and evaluative feedback -- and also unsupervised reinforcement learning to learn a parameter policy that can dynamically adjust the parameters of classical navigation systems in response to changes in the environment. APPL inherits safety and explainability from classical navigation systems while also enjoying the benefits of machine learning, i.e., the ability to adapt and improve from experience. We present a suite of individual APPL methods and also a unifying cycle-of-learning scheme that combines all the proposed methods in a framework that can improve navigation performance through continual, iterative human interaction and simulation training.

Eddy Hudson, Garrett Warnell, Peter Stone

While Adversarial Imitation Learning (AIL) algorithms have recently led to state-of-the-art results on various imitation learning benchmarks, it is unclear as to what impact various design decisions have on performance. To this end, we present here an organizing, modular framework called Reinforcement-learning-based Adversarial Imitation Learning (RAIL) that encompasses and generalizes a popular subclass of existing AIL approaches. Using the view espoused by RAIL, we create two new IfO (Imitation from Observation) algorithms, which we term SAIfO: SAC-based Adversarial Imitation from Observation and SILEM (Skeletal Feature Compensation for Imitation Learning with Embodiment Mismatch). We go into greater depth about SILEM in a separate technical report. In this paper, we focus on SAIfO, evaluating it on a suite of locomotion tasks from OpenAI Gym, and showing that it outperforms contemporaneous RAIL algorithms that perform IfO.

Eddy Hudson, Garrett Warnell, Faraz Torabi et al.

Learning from demonstrations in the wild (e.g. YouTube videos) is a tantalizing goal in imitation learning. However, for this goal to be achieved, imitation learning algorithms must deal with the fact that the demonstrators and learners may have bodies that differ from one another. This condition -- "embodiment mismatch" -- is ignored by many recent imitation learning algorithms. Our proposed imitation learning technique, SILEM (\textbf{S}keletal feature compensation for \textbf{I}mitation \textbf{L}earning with \textbf{E}mbodiment \textbf{M}ismatch), addresses a particular type of embodiment mismatch by introducing a learned affine transform to compensate for differences in the skeletal features obtained from the learner and expert. We create toy domains based on PyBullet's HalfCheetah and Ant to assess SILEM's benefits for this type of embodiment mismatch. We also provide qualitative and quantitative results on more realistic problems -- teaching simulated humanoid agents, including Atlas from Boston Dynamics, to walk by observing human demonstrations.

Faraz Torabi, Garrett Warnell, Peter Stone

In imitation learning from observation IfO, a learning agent seeks to imitate a demonstrating agent using only observations of the demonstrated behavior without access to the control signals generated by the demonstrator. Recent methods based on adversarial imitation learning have led to state-of-the-art performance on IfO problems, but they typically suffer from high sample complexity due to a reliance on data-inefficient, model-free reinforcement learning algorithms. This issue makes them impractical to deploy in real-world settings, where gathering samples can incur high costs in terms of time, energy, and risk. In this work, we hypothesize that we can incorporate ideas from model-based reinforcement learning with adversarial methods for IfO in order to increase the data efficiency of these methods without sacrificing performance. Specifically, we consider time-varying linear Gaussian policies, and propose a method that integrates the linear-quadratic regulator with path integral policy improvement into an existing adversarial IfO framework. The result is a more data-efficient IfO algorithm with better performance, which we show empirically in four simulation domains: using far fewer interactions with the environment, the proposed method exhibits similar or better performance than the existing technique.

Xuesu Xiao, Bo Liu, Garrett Warnell et al.

Moving in complex environments is an essential capability of intelligent mobile robots. Decades of research and engineering have been dedicated to developing sophisticated navigation systems to move mobile robots from one point to another. Despite their overall success, a recently emerging research thrust is devoted to developing machine learning techniques to address the same problem, based in large part on the success of deep learning. However, to date, there has not been much direct comparison between the classical and emerging paradigms to this problem. In this article, we survey recent works that apply machine learning for motion planning and control in mobile robot navigation, within the context of classical navigation systems. The surveyed works are classified into different categories, which delineate the relationship of the learning approaches to classical methods. Based on this classification, we identify common challenges and promising future directions.

Zizhao Wang, Xuesu Xiao, Bo Liu et al.

While classical autonomous navigation systems can typically move robots from one point to another safely and in a collision-free manner, these systems may fail or produce suboptimal behavior in certain scenarios. The current practice in such scenarios is to manually re-tune the system's parameters, e.g. max speed, sampling rate, inflation radius, to optimize performance. This practice requires expert knowledge and may jeopardize performance in the originally good scenarios. Meanwhile, it is relatively easy for a human to identify those failure or suboptimal cases and provide a teleoperated intervention to correct the failure or suboptimal behavior. In this work, we seek to learn from those human interventions to improve navigation performance. In particular, we propose Adaptive Planner Parameter Learning from Interventions (APPLI), in which multiple sets of navigation parameters are learned during training and applied based on a confidence measure to the underlying navigation system during deployment. In our physical experiments, the robot achieves better performance compared to the planner with static default parameters, and even dynamic parameters learned from a full human demonstration. We also show APPLI's generalizability in another unseen physical test course, and a suite of 300 simulated navigation environments.

Zifan Xu, Gauraang Dhamankar, Anirudh Nair et al.

Classical navigation systems typically operate using a fixed set of hand-picked parameters (e.g. maximum speed, sampling rate, inflation radius, etc.) and require heavy expert re-tuning in order to work in new environments. To mitigate this requirement, it has been proposed to learn parameters for different contexts in a new environment using human demonstrations collected via teleoperation. However, learning from human demonstration limits deployment to the training environment, and limits overall performance to that of a potentially-suboptimal demonstrator. In this paper, we introduce APPLR, Adaptive Planner Parameter Learning from Reinforcement, which allows existing navigation systems to adapt to new scenarios by using a parameter selection scheme discovered via reinforcement learning (RL) in a wide variety of simulation environments. We evaluate APPLR on a robot in both simulated and physical experiments, and show that it can outperform both a fixed set of hand-tuned parameters and also a dynamic parameter tuning scheme learned from human demonstration.

Siddharth Desai, Ishan Durugkar, Haresh Karnan et al.

We examine the problem of transferring a policy learned in a source environment to a target environment with different dynamics, particularly in the case where it is critical to reduce the amount of interaction with the target environment during learning. This problem is particularly important in sim-to-real transfer because simulators inevitably model real-world dynamics imperfectly. In this paper, we show that one existing solution to this transfer problem - 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 grounded transfer learning.To validate our hypothesis we derive a new algorithm - generative adversarial reinforced action transformation (GARAT) - based on adversarial imitation from observation techniques. We run experiments in several domains with mismatched dynamics, and find that agents trained with GARAT achieve higher returns in the target environment compared to existing black-box transfer methods

Siddharth Desai, Haresh Karnan, Josiah P. Hanna et al.

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

Haresh Karnan, Siddharth Desai, Josiah P. Hanna et al.

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.

Xuesu Xiao, Bo Liu, Garrett Warnell et al.

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 and generalizes well to unseen environments, including those based on both classical and machine learning techniques.

Xuesu Xiao, Bo Liu, Garrett Warnell et al.

Existing autonomous robot navigation systems allow robots to move from one point to another in a collision-free manner. However, when facing new environments, these systems generally require re-tuning by expert roboticists with a good understanding of the inner workings of the navigation system. In contrast, even users who are unversed in the details of robot navigation algorithms can generate desirable navigation behavior in new environments via teleoperation. In this paper, we introduce APPLD, Adaptive Planner Parameter Learning from Demonstration, that allows existing navigation systems to be successfully applied to new complex environments, given only a human teleoperated demonstration of desirable navigation. APPLD is verified on two robots running different navigation systems in different environments. Experimental results show that APPLD can outperform navigation systems with the default and expert-tuned parameters, and even the human demonstrator themselves.

Nicholas Waytowich, Sean L. Barton, Vernon Lawhern et al.

While deep reinforcement learning techniques have led to agents that are successfully able to learn to perform a number of tasks that had been previously unlearnable, these techniques are still susceptible to the longstanding problem of reward sparsity. This is especially true for tasks such as training an agent to play StarCraft II, a real-time strategy game where reward is only given at the end of a game which is usually very long. While this problem can be addressed through reward shaping, such approaches typically require a human expert with specialized knowledge. Inspired by the vision of enabling reward shaping through the more-accessible paradigm of natural-language narration, we develop a technique that can provide the benefits of reward shaping using natural language commands. Our narration-guided RL agent projects sequences of natural-language commands into the same high-dimensional representation space as corresponding goal states. We show that we can get improved performance with our method compared to traditional reward-shaping approaches. Additionally, we demonstrate the ability of our method to generalize to unseen natural-language commands.

Faraz Torabi, Sean Geiger, Garrett Warnell et al.

Imitation from observation is the framework of learning tasks by observing demonstrated state-only trajectories. Recently, adversarial approaches have achieved significant performance improvements over other methods for imitating complex behaviors. However, these adversarial imitation algorithms often require many demonstration examples and learning iterations to produce a policy that is successful at imitating a demonstrator's behavior. This high sample complexity often prohibits these algorithms from being deployed on physical robots. In this paper, we propose an algorithm that addresses the sample inefficiency problem by utilizing ideas from trajectory centric reinforcement learning algorithms. We test our algorithm and conduct experiments using an imitation task on a physical robot arm and its simulated version in Gazebo and will show the improvement in learning rate and efficiency.

Brahma S. Pavse, Faraz Torabi, Josiah P. Hanna et al.

Augmenting reinforcement learning with imitation learning is often hailed as a method by which to improve upon learning from scratch. However, most existing methods for integrating these two techniques are subject to several strong assumptions---chief among them that information about demonstrator actions is available. In this paper, we investigate the extent to which this assumption is necessary by introducing and evaluating reinforced inverse dynamics modeling (RIDM), a novel paradigm for combining imitation from observation (IfO) and reinforcement learning with no dependence on demonstrator action information. Moreover, RIDM requires only a single demonstration trajectory and is able to operate directly on raw (unaugmented) state features. We find experimentally that RIDM performs favorably compared to a baseline approach for several tasks in simulation as well as for tasks on a real UR5 robot arm. Experiment videos can be found at https://sites.google.com/view/ridm-reinforced-inverse-dynami.

Faraz Torabi, Garrett Warnell, Peter Stone

Imitation learning is the process by which one agent tries to learn how to perform a certain task using information generated by another, often more-expert agent performing that same task. Conventionally, the imitator has access to both state and action information generated by an expert performing the task (e.g., the expert may provide a kinesthetic demonstration of object placement using a robotic arm). However, requiring the action information prevents imitation learning from a large number of existing valuable learning resources such as online videos of humans performing tasks. To overcome this issue, the specific problem of imitation from observation (IfO) has recently garnered a great deal of attention, in which the imitator only has access to the state information (e.g., video frames) generated by the expert. In this paper, we provide a literature review of methods developed for IfO, and then point out some open research problems and potential future work.

Faraz Torabi, Garrett Warnell, Peter Stone

Classically, imitation learning algorithms have been developed for idealized situations, e.g., the demonstrations are often required to be collected in the exact same environment and usually include the demonstrator's actions. Recently, however, the research community has begun to address some of these shortcomings by offering algorithmic solutions that enable imitation learning from observation (IfO), e.g., learning to perform a task from visual demonstrations that may be in a different environment and do not include actions. Motivated by the fact that agents often also have access to their own internal states (i.e., proprioception), we propose and study an IfO algorithm that leverages this information in the policy learning process. The proposed architecture learns policies over proprioceptive state representations and compares the resulting trajectories visually to the demonstration data. We experimentally test the proposed technique on several MuJoCo domains and show that it outperforms other imitation from observation algorithms by a large margin.

Nicholas Waytowich, Sean L. Barton, Vernon Lawhern et al.

While deep reinforcement learning techniques have led to agents that are successfully able to learn to perform a number of tasks that had been previously unlearnable, these techniques are still susceptible to the longstanding problem of {\em reward sparsity}. This is especially true for tasks such as training an agent to play StarCraft II, a real-time strategy game where reward is only given at the end of a game which is usually very long. While this problem can be addressed through reward shaping, such approaches typically require a human expert with specialized knowledge. Inspired by the vision of enabling reward shaping through the more-accessible paradigm of natural-language narration, we investigate to what extent we can contextualize these narrations by grounding them to the goal-specific states. We present a mutual-embedding model using a multi-input deep-neural network that projects a sequence of natural language commands into the same high-dimensional representation space as corresponding goal states. We show that using this model we can learn an embedding space with separable and distinct clusters that accurately maps natural-language commands to corresponding game states . We also discuss how this model can allow for the use of narrations as a robust form of reward shaping to improve RL performance and efficiency.

Prabhat Nagarajan, Garrett Warnell, Peter Stone

While deep reinforcement learning (DRL) has led to numerous successes in recent years, reproducing these successes can be extremely challenging. One reproducibility challenge particularly relevant to DRL is nondeterminism in the training process, which can substantially affect the results. Motivated by this challenge, we study the positive impacts of deterministic implementations in eliminating nondeterminism in training. To do so, we consider the particular case of the deep Q-learning algorithm, for which we produce a deterministic implementation by identifying and controlling all sources of nondeterminism in the training process. One by one, we then allow individual sources of nondeterminism to affect our otherwise deterministic implementation, and measure the impact of each source on the variance in performance. We find that individual sources of nondeterminism can substantially impact the performance of agent, illustrating the benefits of deterministic implementations. In addition, we also discuss the important role of deterministic implementations in achieving exact replicability of results.

Faraz Torabi, Garrett Warnell, Peter Stone

Imitation from observation (IfO) is the problem of learning directly from state-only demonstrations without having access to the demonstrator's actions. The lack of action information both distinguishes IfO from most of the literature in imitation learning, and also sets it apart as a method that may enable agents to learn from a large set of previously inapplicable resources such as internet videos. In this paper, we propose both a general framework for IfO approaches and also a new IfO approach based on generative adversarial networks called generative adversarial imitation from observation (GAIfO). We conduct experiments in two different settings: (1) when demonstrations consist of low-dimensional, manually-defined state features, and (2) when demonstrations consist of high-dimensional, raw visual data. We demonstrate that our approach performs comparably to classical imitation learning approaches (which have access to the demonstrator's actions) and significantly outperforms existing imitation from observation methods in high-dimensional simulation environments.

Faraz Torabi, Garrett Warnell, Peter Stone

Humans often learn how to perform tasks via imitation: they observe others perform a task, and then very quickly infer the appropriate actions to take based on their observations. While extending this paradigm to autonomous agents is a well-studied problem in general, there are two particular aspects that have largely been overlooked: (1) that the learning is done from observation only (i.e., without explicit action information), and (2) that the learning is typically done very quickly. In this work, we propose a two-phase, autonomous imitation learning technique called behavioral cloning from observation (BCO), that aims to provide improved performance with respect to both of these aspects. First, we allow the agent to acquire experience in a self-supervised fashion. This experience is used to develop a model which is then utilized to learn a particular task by observing an expert perform that task without the knowledge of the specific actions taken. We experimentally compare BCO to imitation learning methods, including the state-of-the-art, generative adversarial imitation learning (GAIL) technique, and we show comparable task performance in several different simulation domains while exhibiting increased learning speed after expert trajectories become available.

Garrett Warnell, Nicholas Waytowich, Vernon Lawhern et al.

While recent advances in deep reinforcement learning have allowed autonomous learning agents to succeed at a variety of complex tasks, existing algorithms generally require a lot of training data. One way to increase the speed at which agents are able to learn to perform tasks is by leveraging the input of human trainers. Although such input can take many forms, real-time, scalar-valued feedback is especially useful in situations where it proves difficult or impossible for humans to provide expert demonstrations. Previous approaches have shown the usefulness of human input provided in this fashion (e.g., the TAMER framework), but they have thus far not considered high-dimensional state spaces or employed the use of deep learning. In this paper, we do both: we propose Deep TAMER, an extension of the TAMER framework that leverages the representational power of deep neural networks in order to learn complex tasks in just a short amount of time with a human trainer. We demonstrate Deep TAMER's success by using it and just 15 minutes of human-provided feedback to train an agent that performs better than humans on the Atari game of Bowling - a task that has proven difficult for even state-of-the-art reinforcement learning methods.

Alec Koppel, Garrett Warnell, Ethan Stump et al.

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 tradeoff of objective function evaluation, classification performance, and complexity of the nonparametric regressor extracted the proposed method.

Gordon Christie, Garrett Warnell, Kevin Kochersberger

Localization in a global map is critical to success in many autonomous robot missions. This is particularly challenging for multi-robot operations in unknown and adverse environments. Here, we are concerned with providing a small unmanned ground vehicle (UGV) the ability to localize itself within a 2.5D aerial map generated from imagery captured by a low-flying unmanned aerial vehicle (UAV). We consider the scenario where GPS is unavailable and appearance-based scene changes may have occurred between the UAV's flight and the start of the UGV's mission. We present a GPS-free solution to this localization problem that is robust to appearance shifts by exploiting high-level, semantic representations of image and depth data. Using data gathered at an urban test site, we empirically demonstrate that our technique yields results within five meters of a GPS-based approach.

Alec Koppel, Garrett Warnell, Ethan Stump et al.

We consider discriminative dictionary learning in a distributed online setting, where a network of agents aims to learn a common set of dictionary elements of a feature space and model parameters while sequentially receiving observations. 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. Using Lagrange multipliers to penalize the discrepancy between them, only neighboring nodes exchange model information. We show that decisions made with this saddle point algorithm asymptotically achieve a first-order stationarity condition on average.

Garrett Warnell, Sourabh Bhattacharya, Rama Chellappa et al.

We provide two novel adaptive-rate compressive sensing (CS) strategies for sparse, time-varying signals using side information. Our first method utilizes extra cross-validation measurements, and the second one exploits extra low-resolution measurements. Unlike the majority of current CS techniques, we do not assume that we know an upper bound on the number of significant coefficients that comprise the images in the video sequence. Instead, we use the side information to predict the number of significant coefficients in the signal at the next time instant. For each image in the video sequence, our techniques specify a fixed number of spatially-multiplexed CS measurements to acquire, and adjust this quantity from image to image. Our strategies are developed in the specific context of background subtraction for surveillance video, and we experimentally validate the proposed methods on real video sequences.