Stefanie Tellex

Ziyi Yang, Shreyas S. Raman, Ankit Shah et al. · cmu

Recent advancements in large language models (LLMs) have enabled a new research domain, LLM agents, for solving robotics and planning tasks by leveraging the world knowledge and general reasoning abilities of LLMs obtained during pretraining. However, while considerable effort has been made to teach the robot the "dos," the "don'ts" received relatively less attention. We argue that, for any practical usage, it is as crucial to teach the robot the "don'ts": conveying explicit instructions about prohibited actions, assessing the robot's comprehension of these restrictions, and, most importantly, ensuring compliance. Moreover, verifiable safe operation is essential for deployments that satisfy worldwide standards such as ISO 61508, which defines standards for safely deploying robots in industrial factory environments worldwide. Aiming at deploying the LLM agents in a collaborative environment, we propose a queryable safety constraint module based on linear temporal logic (LTL) that simultaneously enables natural language (NL) to temporal constraints encoding, safety violation reasoning and explaining, and unsafe action pruning. To demonstrate the effectiveness of our system, we conducted experiments in VirtualHome environment and on a real robot. The experimental results show that our system strictly adheres to the safety constraints and scales well with complex safety constraints, highlighting its potential for practical utility.

Jason Xinyu Liu, Ziyi Yang, Ifrah Idrees et al. · cmu

Grounding navigational commands to linear temporal logic (LTL) leverages its unambiguous semantics for reasoning about long-horizon tasks and verifying the satisfaction of temporal constraints. Existing approaches require training data from the specific environment and landmarks that will be used in natural language to understand commands in those environments. We propose Lang2LTL, a modular system and a software package that leverages large language models (LLMs) to ground temporal navigational commands to LTL specifications in environments without prior language data. We comprehensively evaluate Lang2LTL for five well-defined generalization behaviors. Lang2LTL demonstrates the state-of-the-art ability of a single model to ground navigational commands to diverse temporal specifications in 21 city-scaled environments. Finally, we demonstrate a physical robot using Lang2LTL can follow 52 semantically diverse navigational commands in two indoor environments.

Shreyas Sundara Raman, Vanya Cohen, Ifrah Idrees et al.

Extracting commonsense knowledge from a large language model (LLM) offers a path to designing intelligent robots. Existing approaches that leverage LLMs for planning are unable to recover when an action fails and often resort to retrying failed actions, without resolving the error's underlying cause. We propose a novel approach (CAPE) that attempts to propose corrective actions to resolve precondition errors during planning. CAPE improves the quality of generated plans by leveraging few-shot reasoning from action preconditions. Our approach enables embodied agents to execute more tasks than baseline methods while ensuring semantic correctness and minimizing re-prompting. In VirtualHome, CAPE generates executable plans while improving a human-annotated plan correctness metric from 28.89% to 49.63% over SayCan. Our improvements transfer to a Boston Dynamics Spot robot initialized with a set of skills (specified in language) and associated preconditions, where CAPE improves the correctness metric of the executed task plans by 76.49% compared to SayCan. Our approach enables the robot to follow natural language commands and robustly recover from failures, which baseline approaches largely cannot resolve or address inefficiently.

Kaiyu Zheng, Anirudha Paul, Stefanie Tellex

Searching for objects is a fundamental skill for robots. As such, we expect object search to eventually become an off-the-shelf capability for robots, similar to e.g., object detection and SLAM. In contrast, however, no system for 3D object search exists that generalizes across real robots and environments. In this paper, building upon a recent theoretical framework that exploited the octree structure for representing belief in 3D, we present GenMOS (Generalized Multi-Object Search), the first general-purpose system for multi-object search (MOS) in a 3D region that is robot-independent and environment-agnostic. GenMOS takes as input point cloud observations of the local region, object detection results, and localization of the robot's view pose, and outputs a 6D viewpoint to move to through online planning. In particular, GenMOS uses point cloud observations in three ways: (1) to simulate occlusion; (2) to inform occupancy and initialize octree belief; and (3) to sample a belief-dependent graph of view positions that avoid obstacles. We evaluate our system both in simulation and on two real robot platforms. Our system enables, for example, a Boston Dynamics Spot robot to find a toy cat hidden underneath a couch in under one minute. We further integrate 3D local search with 2D global search to handle larger areas, demonstrating the resulting system in a 25m$^2$ lobby area.

Ziyi Yang, Benned Hedegaard, Ahmed Jaafar et al.

Generalizing from individual skill executions to solving long-horizon tasks remains a core challenge in building autonomous agents. A promising direction is learning high-level, symbolic abstractions of the low-level skills of the agents, enabling reasoning and planning independent of the low-level state space. Among possible high-level representations, object-centric skill abstraction with symbolic predicates has been proven to be efficient because of its compatibility with domain-independent planners. Recent advances in foundation models have made it possible to generate symbolic predicates that operate on raw sensory inputs, a process we call generative predicate invention, to facilitate downstream abstraction learning. However, it remains unclear which formal properties the learned representations must satisfy, and how they can be learned to guarantee these properties. In this paper, we address both questions by presenting a formal theory of generative predicate invention for skill abstraction, resulting in symbolic operators that can be used for provably sound and complete planning. Within this framework, we propose SkillWrapper, a method that leverages foundation models to actively collect robot data and learn human-interpretable, plannable representations of black-box skills, using only RGB image observations. Our extensive empirical evaluation in simulation and on real robots shows that SkillWrapper learns abstract representations that enable solving unseen, long-horizon tasks in the real world with black-box skills.

Rafael Rodriguez-Sanchez, Benjamin A. Spiegel, Jennifer Wang et al.

We introduce RLang, a domain-specific language (DSL) for communicating domain knowledge to an RL agent. Unlike existing RL DSLs that ground to \textit{single} elements of a decision-making formalism (e.g., the reward function or policy), RLang can specify information about every element of a Markov decision process. We define precise syntax and grounding semantics for RLang, and provide a parser that grounds RLang programs to an algorithm-agnostic \textit{partial} world model and policy that can be exploited by an RL agent. We provide a series of example RLang programs demonstrating how different RL methods can exploit the resulting knowledge, encompassing model-free and model-based tabular algorithms, policy gradient and value-based methods, hierarchical approaches, and deep methods.

Thao Nguyen, Vladislav Hrosinkov, Eric Rosen et al.

Object search is a challenging task because when given complex language descriptions (e.g., "find the white cup on the table"), the robot must move its camera through the environment and recognize the described object. Previous works map language descriptions to a set of fixed object detectors with predetermined noise models, but these approaches are challenging to scale because new detectors need to be made for each object. In this work, we bridge the gap in realistic object search by posing the search problem as a partially observable Markov decision process (POMDP) where the object detector and visual sensor noise in the observation model is determined by a single Deep Neural Network conditioned on complex language descriptions. We incorporate the neural network's outputs into our language-conditioned observation model (LCOM) to represent dynamically changing sensor noise. With an LCOM, any language description of an object can be used to generate an appropriate object detector and noise model, and training an LCOM only requires readily available supervised image-caption datasets. We empirically evaluate our method by comparing against a state-of-the-art object search algorithm in simulation, and demonstrate that planning with our observation model yields a significantly higher average task completion rate (from 0.46 to 0.66) and more efficient and quicker object search than with a fixed-noise model. We demonstrate our method on a Boston Dynamics Spot robot, enabling it to handle complex natural language object descriptions and efficiently find objects in a room-scale environment.

Ifrah Idrees, Tian Yun, Naveen Sharma et al.

Conversational assistive robots can aid people, especially those with cognitive impairments, to accomplish various tasks such as cooking meals, performing exercises, or operating machines. However, to interact with people effectively, robots must recognize human plans and goals from noisy observations of human actions, even when the user acts sub-optimally. Previous works on Plan and Goal Recognition (PGR) as planning have used hierarchical task networks (HTN) to model the actor/human. However, these techniques are insufficient as they do not have user engagement via natural modes of interaction such as language. Moreover, they have no mechanisms to let users, especially those with cognitive impairments, know of a deviation from their original plan or about any sub-optimal actions taken towards their goal. We propose a novel framework for plan and goal recognition in partially observable domains -- Dialogue for Goal Recognition (D4GR) enabling a robot to rectify its belief in human progress by asking clarification questions about noisy sensor data and sub-optimal human actions. We evaluate the performance of D4GR over two simulated domains -- kitchen and blocks domain. With language feedback and the world state information in a hierarchical task model, we show that D4GR framework for the highest sensor noise performs 1% better than HTN in goal accuracy in both domains. For plan accuracy, D4GR outperforms by 4% in the kitchen domain and 2% in the blocks domain in comparison to HTN. The ALWAYS-ASK oracle outperforms our policy by 3% in goal recognition and 7%in plan recognition. D4GR does so by asking 68% fewer questions than an oracle baseline. We also demonstrate a real-world robot scenario in the kitchen domain, validating the improved plan and goal recognition of D4GR in a realistic setting.

Benedict Quartey, Sebastian Castro, Eric Rosen et al.

Learning from Demonstration (LfD) enables robots to learn complex behaviors from expert examples, yet existing approaches often fail to generalize to new compositions of known skills without retraining. Modern generative policies model distributions over action trajectories alone, thus are unable to reason about the symbolic outcomes required for robust composition. We propose that skills should jointly model action trajectories and the symbolic outcomes they induce. To address this gap, we introduce Predicate Action Skills (PACTS), a class of closed-loop visuomotor policies that model skills as a joint generative process over action and predicate belief trajectories, producing coherent action-outcome rollouts within a single model. Jointly generating actions and predicates enables PACTS to learn internal representations that improve both action generation and predicate classification. Furthermore, we demonstrate zero-shot composition of learned skills via planning by leveraging online predicate predictions from PACTS as a symbolic interface for sequencing and monitoring execution. Project website: https://planpacts.github.io/

Lance Ying, Xinyi Li, Shivam Aarya et al.

Spoken language instructions are ubiquitous in agent collaboration. However, in real-world human-robot collaboration, following human spoken instructions can be challenging due to various speaker and environmental factors, such as background noise or mispronunciation. When faced with noisy auditory inputs, humans can leverage the collaborative context in the embodied environment to interpret noisy spoken instructions and take pragmatic assistive actions. In this paper, we present a cognitively inspired neurosymbolic model, Spoken Instruction Following through Theory of Mind (SIFToM), which leverages a Vision-Language Model with model-based mental inference to enable robots to pragmatically follow human instructions under diverse speech conditions. We test SIFToM in both simulated environments (VirtualHome) and real-world human-robot collaborative settings with human evaluations. Results show that SIFToM can significantly improve the performance of a lightweight base VLM (Gemini 2.5 Flash), outperforming state-of-the-art VLMs (Gemini 2.5 Pro) and approaching human-level accuracy on challenging spoken instruction following tasks.

Lakshita Dodeja, Ondrej Biza, Shivam Vats et al.

Behavior Cloning (BC) has emerged as a highly effective paradigm for robot learning. However, BC lacks a self-guided mechanism for online improvement after demonstrations have been collected. Existing offline-to-online learning methods often cause policies to replace previously learned good actions due to a distribution mismatch between offline data and online learning. In this work, we propose Q2RL, Q-Estimation and Q-Gating from BC for Reinforcement Learning, an algorithm for efficient offline-to-online learning. Our method consists of two parts: (1) Q-Estimation extracts a Q-function from a BC policy using a few interaction steps with the environment, followed by online RL with (2) Q-Gating, which switches between BC and RL policy actions based on their respective Q-values to collect samples for RL policy training. Across manipulation tasks from D4RL and robomimic benchmarks, Q2RL outperforms SOTA offline-to-online learning baselines on success rate and time to convergence. Q2RL is efficient enough to be applied in an on-robot RL setting, learning robust policies for contact-rich and high precision manipulation tasks such as pipe assembly and kitting, in 1-2 hours of online interaction, achieving success rates of up to 100% and up to 3.75x improvement against the original BC policy. Code and video are available at https://pages.rai-inst.com/q2rl_website/

Vanya Cohen, Jason Xinyu Liu, Raymond Mooney et al.

With large language models, robots can understand language more flexibly and more capable than ever before. This survey reviews and situates recent literature into a spectrum with two poles: 1) mapping between language and some manually defined formal representation of meaning, and 2) mapping between language and high-dimensional vector spaces that translate directly to low-level robot policy. Using a formal representation allows the meaning of the language to be precisely represented, limits the size of the learning problem, and leads to a framework for interpretability and formal safety guarantees. Methods that embed language and perceptual data into high-dimensional spaces avoid this manually specified symbolic structure and thus have the potential to be more general when fed enough data but require more data and computing to train. We discuss the benefits and tradeoffs of each approach and finish by providing directions for future work that achieves the best of both worlds.

Benedict Quartey, Eric Rosen, Stefanie Tellex et al.

When instructing robots, users want to flexibly express constraints, refer to arbitrary landmarks, and verify robot behavior, while robots must disambiguate instructions into specifications and ground instruction referents in the real world. To address this problem, we propose Language Instruction grounding for Motion Planning (LIMP), an approach that enables robots to verifiably follow complex, open-ended instructions in real-world environments without prebuilt semantic maps. LIMP constructs a symbolic instruction representation that reveals the robot's alignment with an instructor's intended motives and affords the synthesis of correct-by-construction robot behaviors. We conduct a large-scale evaluation of LIMP on 150 instructions across five real-world environments, demonstrating its versatility and ease of deployment in diverse, unstructured domains. LIMP performs comparably to state-of-the-art baselines on standard open-vocabulary tasks and additionally achieves a 79\% success rate on complex spatiotemporal instructions, significantly outperforming baselines that only reach 38\%. See supplementary materials and demo videos at https://robotlimp.github.io

Ahmed Jaafar, Shreyas Sundara Raman, Sudarshan Harithas et al.

Learning to execute long-horizon mobile manipulation tasks is crucial for advancing robotics in household and workplace settings. However, current approaches are typically data-inefficient, underscoring the need for improved models that require realistically sized benchmarks to evaluate their efficiency. To address this, we introduce the LAMBDA (λ) benchmark-Long-horizon Actions for Mobile-manipulation Benchmarking of Directed Activities-which evaluates the data efficiency of models on language-conditioned, long-horizon, multi-room, multi-floor, pick-and-place tasks using a dataset of manageable size, more feasible for collection. Our benchmark includes 571 human-collected demonstrations that provide realism and diversity in simulated and real-world settings. Unlike planner-generated data, these trajectories offer natural variability and replay-verifiability, ensuring robust learning and evaluation. We leverage λ to benchmark current end-to-end learning methods and a modular neuro-symbolic approach that combines foundation models with task and motion planning. We find that learning methods, even when pretrained, yield lower success rates, while a neuro-symbolic method performs significantly better and requires less data.

Benned Hedegaard, Ziyi Yang, Yichen Wei et al.

Task and motion planning is a well-established approach for solving long-horizon robot planning problems. However, traditional methods assume that each task-level robot action, or skill, can be reduced to kinematic motion planning. In this work, we address the challenge of planning with both kinematic skills and closed-loop motor controllers that go beyond kinematic considerations. We propose a novel method that integrates these controllers into motion planning using Composable Interaction Primitives (CIPs), enabling the use of diverse, non-composable pre-learned skills in hierarchical robot planning. Toward validating our Task and Skill Planning (TASP) approach, we describe ongoing robot experiments in real-world scenarios designed to demonstrate how CIPs can allow a mobile manipulator robot to effectively combine motion planning with general-purpose skills to accomplish complex tasks.

Rachel Ma, Lyndon Lam, Benjamin A. Spiegel et al.

It is imperative that robots can understand natural language commands issued by humans. Such commands typically contain verbs that signify what action should be performed on a given object and that are applicable to many objects. We propose a method for generalizing manipulation skills to novel objects using verbs. Our method learns a probabilistic classifier that determines whether a given object trajectory can be described by a specific verb. We show that this classifier accurately generalizes to novel object categories with an average accuracy of 76.69% across 13 object categories and 14 verbs. We then perform policy search over the object kinematics to find an object trajectory that maximizes classifier prediction for a given verb. Our method allows a robot to generate a trajectory for a novel object based on a verb, which can then be used as input to a motion planner. We show that our model can generate trajectories that are usable for executing five verb commands applied to novel instances of two different object categories on a real robot.

Lakshita Dodeja, Karl Schmeckpeper, Shivam Vats et al.

Residual Reinforcement Learning (RL) is a popular approach for adapting pretrained policies by learning a lightweight residual policy that provides corrective actions. While Residual RL is more sample-efficient than finetuning the entire base policy, existing methods struggle with sparse rewards and are designed for deterministic base policies. We propose two improvements to Residual RL that further enhance its sample efficiency and make it suitable for stochastic base policies. First, we leverage uncertainty estimates of the base policy to focus exploration on regions in which the base policy is not confident. Second, we propose a simple modification to off-policy residual learning that allows it to observe base actions and better handle stochastic base policies. We evaluate our method with both Gaussian-based and Diffusion-based stochastic base policies on tasks from Robosuite and D4RL, and compare against state-of-the-art finetuning methods, demo-augmented RL methods, and other residual RL methods. Our algorithm significantly outperforms existing baselines in a variety of simulation benchmark environments. We also deploy our learned polices in the real world to demonstrate their robustness with zero-shot sim-to-real transfer.

Casey Kennington, Malihe Alikhani, Heather Pon-Barry et al. · cmu

The ability to interact with machines using natural human language is becoming not just commonplace, but expected. The next step is not just text interfaces, but speech interfaces and not just with computers, but with all machines including robots. In this paper, we chronicle the recent history of this growing field of spoken dialogue with robots and offer the community three proposals, the first focused on education, the second on benchmarks, and the third on the modeling of language when it comes to spoken interaction with robots. The three proposals should act as white papers for any researcher to take and build upon.

Cassie Huang, Stuti Mohan, Ziyi Yang et al.

LLMs have been widely used in planning, either as planners to generate action sequences end-to-end, or as formalizers to represent the planning domain and problem in a formal language that can derive plans deterministically. However, both lines of work rely on standard benchmarks that only include generic and simplistic environmental specifications, leading to potential overestimation of the planning ability of LLMs and safety concerns in downstream tasks. We bridge this gap by augmenting widely used planning benchmarks with manually annotated, fine-grained, and rich natural language constraints spanning four formally defined categories. Over 4 state-of-the-art reasoning LLMs, 3 formal languages, 5 methods, and 4 datasets, we show that the introduction of constraints not only consistently halves performance, but also significantly challenges robustness to problem complexity and lexical shift.

Ifrah Idrees, Zahid Hasan, Steven P. Reiss et al.

Robots equipped with situational awareness can help humans efficiently find their lost objects by leveraging spatial and temporal structure. Existing approaches to video and image retrieval do not take into account the unique constraints imposed by a moving camera with a partial view of the environment. We present a Detection-based 3-level hierarchical Association approach, D3A, to create an efficient query-able spatial-temporal representation of unique object instances in an environment. D3A performs online incremental and hierarchical learning to identify keyframes that best represent the unique objects in the environment. These keyframes are learned based on both spatial and temporal features and once identified their corresponding spatial-temporal information is organized in a key-value database. D3A allows for a variety of query patterns such as querying for objects with/without the following: 1) specific attributes, 2) spatial relationships with other objects, and 3) time slices. For a given set of 150 queries, D3A returns a small set of candidate keyframes (which occupy only 0.17% of the total sensory data) with 81.98\% mean accuracy in 11.7 ms. This is 47x faster and 33% more accurate than a baseline that naively stores the object matches (detections) in the database without associating spatial-temporal information.

Kaiyu Zheng, Rohan Chitnis, Yoonchang Sung et al.

In realistic applications of object search, robots will need to locate target objects in complex environments while coping with unreliable sensors, especially for small or hard-to-detect objects. In such settings, correlational information can be valuable for planning efficiently. Previous approaches that consider correlational information typically resort to ad-hoc, greedy search strategies. We introduce the Correlational Object Search POMDP (COS-POMDP), which models correlations while preserving optimal solutions with a reduced state space. We propose a hierarchical planning algorithm to scale up COS-POMDPs for practical domains. Our evaluation, conducted with the AI2-THOR household simulator and the YOLOv5 object detector, shows that our method finds objects more successfully and efficiently compared to baselines,particularly for hard-to-detect objects such as srub brush and remote control.

Eric Hsiung, Hiloni Mehta, Junchi Chu et al.

Recent work on using natural language to specify commands to robots has grounded that language to LTL. However, mapping natural language task specifications to LTL task specifications using language models require probability distributions over finite vocabulary. Existing state-of-the-art methods have extended this finite vocabulary to include unseen terms from the input sequence to improve output generalization. However, novel out-of-vocabulary atomic propositions cannot be generated using these methods. To overcome this, we introduce an intermediate contextual query representation which can be learned from single positive task specification examples, associating a contextual query with an LTL template. We demonstrate that this intermediate representation allows for generalization over unseen object references, assuming accurate groundings are available. We compare our method of mapping natural language task specifications to intermediate contextual queries against state-of-the-art CopyNet models capable of translating natural language to LTL, by evaluating whether correct LTL for manipulation and navigation task specifications can be output, and show that our method outperforms the CopyNet model on unseen object references. We demonstrate that the grounded LTL our method outputs can be used for planning in a simulated OO-MDP environment. Finally, we discuss some common failure modes encountered when translating natural language task specifications to grounded LTL.

Sreehari Rammohan, Shangqun Yu, Bowen He et al.

Learning continuous control in high-dimensional sparse reward settings, such as robotic manipulation, is a challenging problem due to the number of samples often required to obtain accurate optimal value and policy estimates. While many deep reinforcement learning methods have aimed at improving sample efficiency through replay or improved exploration techniques, state of the art actor-critic and policy gradient methods still suffer from the hard exploration problem in sparse reward settings. Motivated by recent successes of value-based methods for approximating state-action values, like RBF-DQN, we explore the potential of value-based reinforcement learning for learning continuous robotic manipulation tasks in multi-task sparse reward settings. On robotic manipulation tasks, we empirically show RBF-DQN converges faster than current state of the art algorithms such as TD3, SAC, and PPO. We also perform ablation studies with RBF-DQN and have shown that some enhancement techniques for vanilla Deep Q learning such as Hindsight Experience Replay (HER) and Prioritized Experience Replay (PER) can also be applied to RBF-DQN. Our experimental analysis suggests that value-based approaches may be more sensitive to data augmentation and replay buffer sample techniques than policy-gradient methods, and that the benefits of these methods for robot manipulation are heavily dependent on the transition dynamics of generated subgoal states.

Monica Roy, Kaiyu Zheng, Jason Liu et al.

We envision robots that can collaborate and communicate seamlessly with humans. It is necessary for such robots to decide both what to say and how to act, while interacting with humans. To this end, we introduce a new task, dialogue object search: A robot is tasked to search for a target object (e.g. fork) in a human environment (e.g., kitchen), while engaging in a "video call" with a remote human who has additional but inexact knowledge about the target's location. That is, the robot conducts speech-based dialogue with the human, while sharing the image from its mounted camera. This task is challenging at multiple levels, from data collection, algorithm and system development,to evaluation. Despite these challenges, we believe such a task blocks the path towards more intelligent and collaborative robots. In this extended abstract, we motivate and introduce the dialogue object search task and analyze examples collected from a pilot study. We then discuss our next steps and conclude with several challenges on which we hope to receive feedback.

Matt Corsaro, Stefanie Tellex, George Konidaris

We propose an approach to multi-modal grasp detection that jointly predicts the probabilities that several types of grasps succeed at a given grasp pose. Given a partial point cloud of a scene, the algorithm proposes a set of feasible grasp candidates, then estimates the probabilities that a grasp of each type would succeed at each candidate pose. Predicting grasp success probabilities directly from point clouds makes our approach agnostic to the number and placement of depth sensors at execution time. We evaluate our system both in simulation and on a real robot with a Robotiq 3-Finger Adaptive Gripper and compare our network against several baselines that perform fewer types of grasps. Our experiments show that a system that explicitly models grasp type achieves an object retrieval rate 8.5% higher in a complex cluttered environment than our highest-performing baseline.

Ben Abbatematteo, Eric Rosen, Stefanie Tellex et al.

Learning a robot motor skill from scratch is impractically slow; so much so that in practice, learning must be bootstrapped using a good skill policy obtained from human demonstration. However, relying on human demonstration necessarily degrades the autonomy of robots that must learn a wide variety of skills over their operational lifetimes. We propose using kinematic motion planning as a completely autonomous, sample efficient way to bootstrap motor skill learning for object manipulation. We demonstrate the use of motion planners to bootstrap motor skills in two complex object manipulation scenarios with different policy representations: opening a drawer with a dynamic movement primitive representation, and closing a microwave door with a deep neural network policy. We also show how our method can bootstrap a motor skill for the challenging dynamic task of learning to hit a ball off a tee, where a kinematic plan based on treating the scene as static is insufficient to solve the task, but sufficient to bootstrap a more dynamic policy. In all three cases, our method is competitive with human-demonstrated initialization, and significantly outperforms starting with a random policy. This approach enables robots to to efficiently and autonomously learn motor policies for dynamic tasks without human demonstration.

Kaiyu Zheng, Deniz Bayazit, Rebecca Mathew et al.

Humans use spatial language to naturally describe object locations and their relations. Interpreting spatial language not only adds a perceptual modality for robots, but also reduces the barrier of interfacing with humans. Previous work primarily considers spatial language as goal specification for instruction following tasks in fully observable domains, often paired with reference paths for reward-based learning. However, spatial language is inherently subjective and potentially ambiguous or misleading. Hence, in this paper, we consider spatial language as a form of stochastic observation. We propose SLOOP (Spatial Language Object-Oriented POMDP), a new framework for partially observable decision making with a probabilistic observation model for spatial language. We apply SLOOP to object search in city-scale environments. To interpret ambiguous, context-dependent prepositions (e.g. front), we design a simple convolutional neural network that predicts the language provider's latent frame of reference (FoR) given the environment context. Search strategies are computed via an online POMDP planner based on Monte Carlo Tree Search. Evaluation based on crowdsourced language data, collected over areas of five cities in OpenStreetMap, shows that our approach achieves faster search and higher success rate compared to baselines, with a wider margin as the spatial language becomes more complex. Finally, we demonstrate the proposed method in AirSim, a realistic simulator where a drone is tasked to find cars in a neighborhood environment.

Michael Fishman, Nishanth Kumar, Cameron Allen et al.

A general-purpose planning agent requires an open-scope world model: one rich enough to tackle any of the wide range of tasks it may be asked to solve over its operational lifetime. This stands in contrast with typical planning approaches, where the scope of a model is limited to a specific family of tasks that share significant structure. Unfortunately, planning to solve any specific task using an open-scope model is computationally intractable - even for state-of-the-art methods - due to the many states and actions that are necessarily present in the model but irrelevant to that problem. We propose task scoping: a method that exploits knowledge of the initial state, goal conditions, and transition system to automatically and efficiently remove provably irrelevant variables and actions from a planning problem. Our approach leverages causal link analysis and backwards reachability over state variables (rather than states) along with operator merging (when effects on relevant variables are identical). Using task scoping as a pre-planning step can shrink the search space by orders of magnitude and dramatically decrease planning time. We empirically demonstrate that these improvements occur across a variety of open-scope domains, including Minecraft, where our approach leads to a 75x reduction in search time with a state-of-the-art numeric planner, even after including the time required for task scoping itself.

Atsunobu Kotani, Stefanie Tellex, James Tompkin

Representing a space of handwriting stroke styles includes the challenge of representing both the style of each character and the overall style of the human writer. Existing VRNN approaches to representing handwriting often do not distinguish between these different style components, which can reduce model capability. Instead, we introduce the Decoupled Style Descriptor (DSD) model for handwriting, which factors both character- and writer-level styles and allows our model to represent an overall greater space of styles. This approach also increases flexibility: given a few examples, we can generate handwriting in new writer styles, and also now generate handwriting of new characters across writer styles. In experiments, our generated results were preferred over a state of the art baseline method 88% of the time, and in a writer identification task on 20 held-out writers, our DSDs achieved 89.38% accuracy from a single sample word. Overall, DSDs allows us to improve both the quality and flexibility over existing handwriting stroke generation approaches.

Eric Rosen, Stefanie Tellex, Geroge Konidaris

There are unwritten guidelines for how to make robot videos that researchers learn from their advisors and pass onto their students. We believe that it is important for the community to collaboratively discuss and develop a standard set of best practices when making robot. We suggest a starting set of maxims for best robot video practices, and highlight positive examples from the community and negative examples only from videos made by the authors of this article. In addition, we offer a checklist that we hope can act as an document that can be given to robotic researchers to inform them of how to make robot videos that truthfully characterize what a robot can and can not do. We consider this a first draft, and are looking for feedback from the community as we refine and grow our maxims and checklist.

Thao Nguyen, Nakul Gopalan, Roma Patel et al.

Natural language object retrieval is a highly useful yet challenging task for robots in human-centric environments. Previous work has primarily focused on commands specifying the desired object's type such as "scissors" and/or visual attributes such as "red," thus limiting the robot to only known object classes. We develop a model to retrieve objects based on descriptions of their usage. The model takes in a language command containing a verb, for example "Hand me something to cut," and RGB images of candidate objects and selects the object that best satisfies the task specified by the verb. Our model directly predicts an object's appearance from the object's use specified by a verb phrase. We do not need to explicitly specify an object's class label. Our approach allows us to predict high level concepts like an object's utility based on the language query. Based on contextual information present in the language commands, our model can generalize to unseen object classes and unknown nouns in the commands. Our model correctly selects objects out of sets of five candidates to fulfill natural language commands, and achieves an average accuracy of 62.3% on a held-out test set of unseen ImageNet object classes and 53.0% on unseen object classes and unknown nouns. Our model also achieves an average accuracy of 54.7% on unseen YCB object classes, which have a different image distribution from ImageNet objects. We demonstrate our model on a KUKA LBR iiwa robot arm, enabling the robot to retrieve objects based on natural language descriptions of their usage. We also present a new dataset of 655 verb-object pairs denoting object usage over 50 verbs and 216 object classes.

Kaiyu Zheng, Yoonchang Sung, George Konidaris et al.

Robots operating in households must find objects on shelves, under tables, and in cupboards. In such environments, it is crucial to search efficiently at 3D scale while coping with limited field of view and the complexity of searching for multiple objects. Principled approaches to object search frequently use Partially Observable Markov Decision Process (POMDP) as the underlying framework for computing search strategies, but constrain the search space in 2D. In this paper, we present a POMDP formulation for multi-object search in a 3D region with a frustum-shaped field-of-view. To efficiently solve this POMDP, we propose a multi-resolution planning algorithm based on online Monte-Carlo tree search. In this approach, we design a novel octree-based belief representation to capture uncertainty of the target objects at different resolution levels, then derive abstract POMDPs at lower resolutions with dramatically smaller state and observation spaces. Evaluation in a simulated 3D domain shows that our approach finds objects more efficiently and successfully compared to a set of baselines without resolution hierarchy in larger instances under the same computational requirement. We demonstrate our approach on a mobile robot to find objects placed at different heights in two 10m$^2 \times 2$m regions by moving its base and actuating its torso.

Kaiyu Zheng, Stefanie Tellex

In this paper, we present pomdp_py, a general purpose Partially Observable Markov Decision Process (POMDP) library written in Python and Cython. Existing POMDP libraries often hinder accessibility and efficient prototyping due to the underlying programming language or interfaces, and require extra complexity in software toolchain to integrate with robotics systems. pomdp_py features simple and comprehensive interfaces capable of describing large discrete or continuous (PO)MDP problems. Here, we summarize the design principles and describe in detail the programming model and interfaces in pomdp_py. We also describe intuitive integration of this library with ROS (Robot Operating System), which enabled our torso-actuated robot to perform object search in 3D. Finally, we note directions to improve and extend this library for POMDP planning and beyond.

Ifrah Idrees, Steven P. Reiss, Stefanie Tellex

Robots have the potential to improve health monitoring outcomes for the elderly by providing doctors, and caregivers with information about the person's behavior, health activities and their surrounding environment. Over the years, less work has been done to enable robots to preserve information for longer periods of time, on the order of months and years of data, and use this contextual information to answer queries. Time complexity to process this massive sensor data in a timely fashion, inability to anticipate the future queries in advance and imprecision involved in the results have been the main impediments in making progress in this area. We make a contribution by introducing RoboMem, a query answering system for health-care assistance of elderly over long term; continuous data feeds that intends to overcome the challenges of giving long term memory to robots. The design for our framework preprocesses the sensor data and stores this preprocessed data into the database. This data is updated in the database by going through successive refinements, improving its accuracy for responding to queries. If data in the database is not enough to answer a query, a small set of relevant frames (also obtained from the database) will be reprocessed to obtain the answer. [Our initial prototype of RoboMem stores 3.5MB of data in the database as compared to 535.8MB of actual video frames and with minimal data in the database it is able to fetch information fundamental to respond to queries in 0.0002 seconds on average].

Jonathan Chang, Nishanth Kumar, Sean Hastings et al.

Robots need to learn skills that can not only generalize across similar problems but also be directed to a specific goal. Previous methods either train a new skill for every different goal or do not infer the specific target in the presence of multiple goals from visual data. We introduce an end-to-end method that represents targetable visuomotor skills as a goal-parameterized neural network policy. By training on an informative subset of available goals with the associated target parameters, we are able to learn a policy that can zero-shot generalize to previously unseen goals. We evaluate our method in a representative 2D simulation of a button-grid and on both button-pressing and peg-insertion tasks on two different physical arms. We demonstrate that our model trained on 33% of the possible goals is able to generalize to more than 90% of the targets in the scene for both simulation and robot experiments. We also successfully learn a mapping from target pixel coordinates to a robot policy to complete a specified goal.

Luke Eller, Theo Guerin, Baichuan Huang et al.

PiDrone is a quadrotor platform created to accompany an introductory robotics course. Students build an autonomous flying robot from scratch and learn to program it through assignments and projects. Existing educational robots do not have significant autonomous capabilities, such as high-level planning and mapping. We present a hardware and software framework for an autonomous aerial robot, in which all software for autonomy can run onboard the drone, implemented in Python. We present an Unscented Kalman Filter (UKF) for accurate state estimation. Next, we present an implementation of Monte Carlo (MC) Localization and FastSLAM for Simultaneous Localization and Mapping (SLAM). The performance of UKF, localization, and SLAM is tested and compared to ground truth, provided by a motion-capture system. Our evaluation demonstrates that our autonomous educational framework runs quickly and accurately on a Raspberry Pi in Python, making it ideal for use in educational settings.

Vanya Cohen, Benjamin Burchfiel, Thao Nguyen et al.

We develop a system to disambiguate object instances within the same class based on simple physical descriptions. The system takes as input a natural language phrase and a depth image containing a segmented object and predicts how similar the observed object is to the object described by the phrase. Our system is designed to learn from only a small amount of human-labeled language data and generalize to viewpoints not represented in the language-annotated depth image training set. By decoupling 3D shape representation from language representation, this method is able to ground language to novel objects using a small amount of language-annotated depth-data and a larger corpus of unlabeled 3D object meshes, even when these objects are partially observed from unusual viewpoints. Our system is able to disambiguate between novel objects, observed via depth images, based on natural language descriptions. Our method also enables view-point transfer; trained on human-annotated data on a small set of depth images captured from frontal viewpoints, our system successfully predicted object attributes from rear views despite having no such depth images in its training set. Finally, we demonstrate our approach on a Baxter robot, enabling it to pick specific objects based on human-provided natural language descriptions.

Yoonseon Oh, Roma Patel, Thao Nguyen et al.

Often times, we specify tasks for a robot using temporal language that can also span different levels of abstraction. The example command ``go to the kitchen before going to the second floor'' contains spatial abstraction, given that ``floor'' consists of individual rooms that can also be referred to in isolation ("kitchen", for example). There is also a temporal ordering of events, defined by the word "before". Previous works have used Linear Temporal Logic (LTL) to interpret temporal language (such as "before"), and Abstract Markov Decision Processes (AMDPs) to interpret hierarchical abstractions (such as "kitchen" and "second floor"), separately. To handle both types of commands at once, we introduce the Abstract Product Markov Decision Process (AP-MDP), a novel approach capable of representing non-Markovian reward functions at different levels of abstractions. The AP-MDP framework translates LTL into its corresponding automata, creates a product Markov Decision Process (MDP) of the LTL specification and the environment MDP, and decomposes the problem into subproblems to enable efficient planning with abstractions. AP-MDP performs faster than a non-hierarchical method of solving LTL problems in over 95% of tasks, and this number only increases as the size of the environment domain increases. We also present a neural sequence-to-sequence model trained to translate language commands into LTL expression, and a new corpus of non-Markovian language commands spanning different levels of abstraction. We test our framework with the collected language commands on a drone, demonstrating that our approach enables a robot to efficiently solve temporal commands at different levels of abstraction.

Stefanie Tellex, Andy Brown, Sergei Lupashin

This document describes standard approaches for filtering and estimation for quadrotors, created for the Udacity Flying Cars course. We assume previous knowledge of probability and some knowledge of linear algebra. We do not assume previous knowledge of Kalman filters or Bayes filters. This document derives an EKF for various models of drones in 1D, 2D, and 3D. We use the EKF and notation as defined in Thrun et al. [13]. We also give pseudocode for the Bayes filter, the EKF, and the Unscented Kalman filter [14]. The motivation behind this document is the lack of a step-by-step EKF tutorial that provides the derivations for a quadrotor helicopter. The goal of estimation is to infer the drone's state (pose, velocity, acceleration, and biases) from its sensor values and control inputs. This problem is challenging because sensors are noisy. Additionally, because of weight and cost issues, many drones have limited on-board computation so we want to estimate these values as quickly as possible. The standard method for performing this method is the Extended Kalman filter, a nonlinear extension of the Kalman filter which linearizes a nonlinear transition and measurement model around the current state. However the Unscented Kalman filter is better in almost every respect: simpler to implement, more accurate to estimate, and comparable runtimes.

Nicholas DeMarinis, Stefanie Tellex, Vasileios Kemerlis et al.

Because robots can directly perceive and affect the physical world, security issues take on particular importance. In this paper, we describe the results of our work on scanning the entire IPv4 address space of the Internet for instances of the Robot Operating System (ROS), a widely used robotics platform for research. Our results identified that a number of hosts supporting ROS are exposed to the public Internet, thereby allowing anyone to access robotic sensors and actuators. As a proof of concept, and with consent, we were able to read image sensor information and move the robot of a research group in a US university. This paper gives an overview of our findings, including the geographic distribution of publicly-accessible platforms, the sorts of sensor and actuator data that is available, as well as the different kinds of robots and sensors that our scan uncovered. Additionally, we offer recommendations on best practices to mitigate these security issues in the future.

Thomas Kollar, Stefanie Tellex, Matthew Walter et al.

Many task domains require robots to interpret and act upon natural language commands which are given by people and which refer to the robot's physical surroundings. Such interpretation is known variously as the symbol grounding problem, grounded semantics and grounded language acquisition. This problem is challenging because people employ diverse vocabulary and grammar, and because robots have substantial uncertainty about the nature and contents of their surroundings, making it difficult to associate the constitutive language elements (principally noun phrases and spatial relations) of the command text to elements of those surroundings. Symbolic models capture linguistic structure but have not scaled successfully to handle the diverse language produced by untrained users. Existing statistical approaches can better handle diversity, but have not to date modeled complex linguistic structure, limiting achievable accuracy. Recent hybrid approaches have addressed limitations in scaling and complexity, but have not effectively associated linguistic and perceptual features. Our framework, called Generalized Grounding Graphs (G^3), addresses these issues by defining a probabilistic graphical model dynamically according to the linguistic parse structure of a natural language command. This approach scales effectively, handles linguistic diversity, and enables the system to associate parts of a command with the specific objects, places, and events in the external world to which they refer. We show that robots can learn word meanings and use those learned meanings to robustly follow natural language commands produced by untrained users. We demonstrate our approach for both mobility commands and mobile manipulation commands involving a variety of semi-autonomous robotic platforms, including a wheelchair, a micro-air vehicle, a forklift, and the Willow Garage PR2.

Melrose Roderick, James MacGlashan, Stefanie Tellex

The Deep Q-Network proposed by Mnih et al. [2015] has become a benchmark and building point for much deep reinforcement learning research. However, replicating results for complex systems is often challenging since original scientific publications are not always able to describe in detail every important parameter setting and software engineering solution. In this paper, we present results from our work reproducing the results of the DQN paper. We highlight key areas in the implementation that were not covered in great detail in the original paper to make it easier for researchers to replicate these results, including termination conditions and gradient descent algorithms. Finally, we discuss methods for improving the computational performance and provide our own implementation that is designed to work with a range of domains, and not just the original Arcade Learning Environment [Bellemare et al., 2013].

Melrose Roderick, Christopher Grimm, Stefanie Tellex

We examine the problem of learning and planning on high-dimensional domains with long horizons and sparse rewards. Recent approaches have shown great successes in many Atari 2600 domains. However, domains with long horizons and sparse rewards, such as Montezuma's Revenge and Venture, remain challenging for existing methods. Methods using abstraction (Dietterich 2000; Sutton, Precup, and Singh 1999) have shown to be useful in tackling long-horizon problems. We combine recent techniques of deep reinforcement learning with existing model-based approaches using an expert-provided state abstraction. We construct toy domains that elucidate the problem of long horizons, sparse rewards and high-dimensional inputs, and show that our algorithm significantly outperforms previous methods on these domains. Our abstraction-based approach outperforms Deep Q-Networks (Mnih et al. 2015) on Montezuma's Revenge and Venture, and exhibits backtracking behavior that is absent from previous methods.

Eric Rosen, David Whitney, Elizabeth Phillips et al.

Efficient motion intent communication is necessary for safe and collaborative work environments with collocated humans and robots. Humans efficiently communicate their motion intent to other humans through gestures, gaze, and social cues. However, robots often have difficulty efficiently communicating their motion intent to humans via these methods. Many existing methods for robot motion intent communication rely on 2D displays, which require the human to continually pause their work and check a visualization. We propose a mixed reality head-mounted display visualization of the proposed robot motion over the wearer's real-world view of the robot and its environment. To evaluate the effectiveness of this system against a 2D display visualization and against no visualization, we asked 32 participants to labeled different robot arm motions as either colliding or non-colliding with blocks on a table. We found a 16% increase in accuracy with a 62% decrease in the time it took to complete the task compared to the next best system. This demonstrates that a mixed-reality HMD allows a human to more quickly and accurately tell where the robot is going to move than the compared baselines.

Lucas Lehnert, Stefanie Tellex, Michael L. Littman

One question central to Reinforcement Learning is how to learn a feature representation that supports algorithm scaling and re-use of learned information from different tasks. Successor Features approach this problem by learning a feature representation that satisfies a temporal constraint. We present an implementation of an approach that decouples the feature representation from the reward function, making it suitable for transferring knowledge between domains. We then assess the advantages and limitations of using Successor Features for transfer.

Siddharth Karamcheti, Edward C. Williams, Dilip Arumugam et al.

Robots operating alongside humans in diverse, stochastic environments must be able to accurately interpret natural language commands. These instructions often fall into one of two categories: those that specify a goal condition or target state, and those that specify explicit actions, or how to perform a given task. Recent approaches have used reward functions as a semantic representation of goal-based commands, which allows for the use of a state-of-the-art planner to find a policy for the given task. However, these reward functions cannot be directly used to represent action-oriented commands. We introduce a new hybrid approach, the Deep Recurrent Action-Goal Grounding Network (DRAGGN), for task grounding and execution that handles natural language from either category as input, and generalizes to unseen environments. Our robot-simulation results demonstrate that a system successfully interpreting both goal-oriented and action-oriented task specifications brings us closer to robust natural language understanding for human-robot interaction.

Dilip Arumugam, Siddharth Karamcheti, Nakul Gopalan et al.

Humans can ground natural language commands to tasks at both abstract and fine-grained levels of specificity. For instance, a human forklift operator can be instructed to perform a high-level action, like "grab a pallet" or a low-level action like "tilt back a little bit." While robots are also capable of grounding language commands to tasks, previous methods implicitly assume that all commands and tasks reside at a single, fixed level of abstraction. Additionally, methods that do not use multiple levels of abstraction encounter inefficient planning and execution times as they solve tasks at a single level of abstraction with large, intractable state-action spaces closely resembling real world complexity. In this work, by grounding commands to all the tasks or subtasks available in a hierarchical planning framework, we arrive at a model capable of interpreting language at multiple levels of specificity ranging from coarse to more granular. We show that the accuracy of the grounding procedure is improved when simultaneously inferring the degree of abstraction in language used to communicate the task. Leveraging hierarchy also improves efficiency: our proposed approach enables a robot to respond to a command within one second on 90% of our tasks, while baselines take over twenty seconds on half the tasks. Finally, we demonstrate that a real, physical robot can ground commands at multiple levels of abstraction allowing it to efficiently plan different subtasks within the same planning hierarchy.