Michael Hagenow

Seongheon Park, Wendi Li, Changdae Oh et al.

Vision-Language-Action (VLA) models enable robots to follow natural language instructions and generalize across diverse tasks, but they remain vulnerable to execution failures that compromise reliability in real-world deployment. Detecting such failures during execution is therefore critical for the robust deployment of embodied systems. Existing failure detection methods either rely on expensive action resampling or external models, while alternatives propagate trajectory-level labels uniformly across every timestep, obscuring localized failure signals. In this paper, we propose \textbf{Hide-and-Seek}, a framework that formulates VLA failure detection as a coarsely supervised learning problem. By combining inter-trajectory and intra-trajectory contrastive objectives, Hide-and-Seek localizes failure-indicative actions and induces temporally structured failure signals from trajectory-level supervision alone, without any step-level annotation. We evaluate Hide-and-Seek on LIBERO, VLABench, and a real-world robotic platform across three representative VLA policies: OpenVLA, $π_0$, and $π_{0.5}$.Our method achieves state-of-the-art multi-task failure detection performance with a practical accuracy--timeliness trade-off under conformal prediction, and generalizes well to both seen and unseen tasks.

Alex Cuellar, Michael Hagenow, Julie Shah

Effective human-robot teaming is crucial for the practical deployment of robots in human workspaces. However, optimizing joint human-robot plans remains a challenge due to the difficulty of modeling individualized human capabilities and preferences. While prior research has leveraged the multi-cycle structure of domains like manufacturing to learn an individual's tendencies and adapt plans over repeated interactions, these techniques typically consider task-level and motion-level adaptation in isolation. Task-level methods optimize allocation and scheduling but often ignore spatial interference in close-proximity scenarios; conversely, motion-level methods focus on collision avoidance while ignoring the broader task context. This paper introduces RAPIDDS, a framework that unifies these approaches by modeling an individual's spatial behavior (motion paths) and temporal behavior (time required to complete tasks) over multiple cycles. RAPIDDS then jointly adapts task schedules and steers diffusion models of robot motions to maximize efficiency and minimize proximity accounting for these individualized models. We demonstrate the importance of this dual adaptation through an ablation study in simulation and a physical robot scenario using a 7-DOF robot arm. Finally, we present a user study (n=32) showing significant plan improvement compared to non-adaptive systems across both objective metrics, such as efficiency and proximity, and subjective measures, including fluency and user preference. See this paper's companion video at: https://youtu.be/55Q3lq1fINs.

Yanwei Wang, Tsun-Hsuan Wang, Jiayuan Mao et al.

Grounding the common-sense reasoning of Large Language Models (LLMs) in physical domains remains a pivotal yet unsolved problem for embodied AI. Whereas prior works have focused on leveraging LLMs directly for planning in symbolic spaces, this work uses LLMs to guide the search of task structures and constraints implicit in multi-step demonstrations. Specifically, we borrow from manipulation planning literature the concept of mode families, which group robot configurations by specific motion constraints, to serve as an abstraction layer between the high-level language representations of an LLM and the low-level physical trajectories of a robot. By replaying a few human demonstrations with synthetic perturbations, we generate coverage over the demonstrations' state space with additional successful executions as well as counterfactuals that fail the task. Our explanation-based learning framework trains an end-to-end differentiable neural network to predict successful trajectories from failures and as a by-product learns classifiers that ground low-level states and images in mode families without dense labeling. The learned grounding classifiers can further be used to translate language plans into reactive policies in the physical domain in an interpretable manner. We show our approach improves the interpretability and reactivity of imitation learning through 2D navigation and simulated and real robot manipulation tasks. Website: https://yanweiw.github.io/glide

Michael Hagenow, Michael Zinn, Terrence Fong et al.

Affordance Templates (ATs) are a method for parameterizing objects for autonomous robot manipulations. In this approach, instances of an object are registered by positioning a model in a 3D environment, which requires a large amount of user input. We instead propose a registration method which combines autonomy and user corrections. For selected objects, the system determines both the model and corresponding pose autonomously. The user makes corrections only if the model or pose is incorrect. This method increases the level of autonomy compared to existing approaches which can reduce user input and time on task. In this paper, we present an overview of existing methods, a description of our method, preliminary results, and planned future work.

Michael Hagenow, Bolun Zhang, Bilge Mutlu et al.

Manipulations of a constrained object often use a non-rigid grasp that allows the object to rotate relative to the end effector. This orientation slip strategy is often present in natural human demonstrations, yet it is generally overlooked in methods to identify constraints from such demonstrations. In this paper, we present a method to model and recognize prehensile orientation slip in human demonstrations of constrained interactions. Using only observations of an end effector, we can detect the type of constraint, parameters of the constraint, and orientation slip properties. Our method uses a novel hierarchical model selection method that is informed by multiple origins of physics-based evidence. A study with eight participants shows that orientation slip occurs in natural demonstrations and confirms that it can be detected by our method.

Guru Subramani, Michael Hagenow, Michael Gleicher et al.

Many physical tasks such as pulling out a drawer or wiping a table can be modeled with geometric constraints. These geometric constraints are characterized by restrictions on kinematic trajectories and reaction wrenches (forces and moments) of objects under the influence of the constraint. This paper presents a method to infer geometric constraints involving unmodeled objects in human demonstrations using both kinematic and wrench measurements. Our approach takes a recording of a human demonstration and determines what constraints are present, when they occur, and their parameters (e.g. positions). By using both kinematic and wrench information, our methods are able to reliably identify a variety of constraint types, even if the constraints only exist for short durations within the demonstration. We present a systematic approach to fitting arbitrary scleronomic constraint models to kinematic and wrench measurements. Reaction forces are estimated from measurements by removing friction. Position, orientation, force, and moment error metrics are developed to provide systematic comparison between constraint models. By conducting a user study, we show that our methods can reliably identify constraints in realistic situations and confirm the value of including forces and moments in the model regression and selection process.