Multimodal Probabilistic Model-Based Planning for Human-Robot Interaction

Szczerbicki, Edward; Leung, Karen Ka Yan; Vollprecht, Wolf; Pavone, Marco

doi:10.48550/arxiv.1710.09483

Cited by 5 publications

(10 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One common approach for predicting human actions is supervised learning, where the current state and the history of the human's actions are used directly to predict future actions. Such approaches have enabled inference and planning around human arm motion [1,5,13,15,10], navigation [5], plans for multi-step tasks like assembly [10], and driving [20].…”

Section: A Human Modeling and Predictionmentioning

confidence: 99%

Probabilistically Safe Robot Planning with Confidence-Based Human Predictions

Fisac¹,

Bajcsy²,

Herbert³

et al. 2018

Robotics: Science and Systems XIV

View full text Add to dashboard Cite

Abstract-In order to safely operate around humans, robots can employ predictive models of human motion. Unfortunately, these models cannot capture the full complexity of human behavior and necessarily introduce simplifying assumptions. As a result, predictions may degrade whenever the observed human behavior departs from the assumed structure, which can have negative implications for safety. In this paper, we observe that how "rational" human actions appear under a particular model can be viewed as an indicator of that model's ability to describe the human's current motion. By reasoning about this model confidence in a real-time Bayesian framework, we show that the robot can very quickly modulate its predictions to become more uncertain when the model performs poorly. Building on recent work in provably-safe trajectory planning, we leverage these confidence-aware human motion predictions to generate assured autonomous robot motion. Our new analysis combines worst-case tracking error guarantees for the physical robot with probabilistic time-varying human predictions, yielding a quantitative, probabilistic safety certificate. We demonstrate our approach with a quadcopter navigating around a human.

show abstract

Section: A Human Modeling and Predictionmentioning

confidence: 99%

Probabilistically Safe Robot Planning with Confidence-Based Human Predictions

Fisac¹,

Bajcsy²,

Herbert³

et al. 2018

Robotics: Science and Systems XIV

View full text Add to dashboard Cite

show abstract

“…Lee et al [1] predict vehicle trajectories in environments with multiple counterparts, such as other vehicles and pedestrians. Ivanovic et al [4] and Schmerling et al [3] use CVAEs to predict human behavior for human-robot interaction. Zheng et al [20] use deep RNN models to recreate trajectories of a team of basketball players.…”

Section: Related Workmentioning

confidence: 99%

“…Many real world problems involve multiple agents operating in the same environment, including autonomous vehicles [1,2,3], navigation in presence of humans [4], learning communication [5], and multiplayer games [6,7]. In this work, we consider continuous control problems with two agents, a setting relevant when robots interact with other robots or with humans in their workspace.…”

Section: Introductionmentioning

confidence: 99%

“…In Mishra et al[12], the action and state sequences were explicit, X and U , and prediction consisted of two CVAEs: P (u) (U + |U − , Z) and P (x) (X + |X − , U + , U − , Z). Here, we unify the models into one CVAE 3. The distribution P (X + |X − , Z) is typically represented as a neural network that encodes the mean of a Gaussian, and the expectation in (2) is approximated with the deterministic prediction of the mean trajectory.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Multi-Agent Reinforcement Learning with Multi-Step Generative Models

Krupnik,

Mordatch,

Tamar

2019

Preprint

View full text Add to dashboard Cite

We consider model-based reinforcement learning (MBRL) in 2-agent, high-fidelity continuous control problems -an important domain for robots interacting with other agents in the same workspace. For non-trivial dynamical systems, MBRL typically suffers from accumulating errors. Several recent studies have addressed this problem by learning latent variable models for trajectory segments and optimizing over behavior in the latent space. In this work, we investigate whether this approach can be extended to 2-agent competitive and cooperative settings. The fundamental challenge is how to learn models that capture interactions between agents, yet are disentangled to allow for optimization of each agent behavior separately. We propose such models based on a disentangled variational auto-encoder, and demonstrate our approach on a simulated 2-robot manipulation task, where one robot can either help or distract the other. We show that our approach has better sample efficiency than a strong model-free RL baseline, and can learn both cooperative and adversarial behavior from the same data.

show abstract

“…But even if humans do use such tools in interaction, it is not at all clear that robots ought to as well. For robots and interaction, methods from all three paradigms exist: model-free [14]- [16], regular modelbased [17], and ToM-based [18]- [22].…”

Section: Introductionmentioning

confidence: 99%

On the Utility of Model Learning in HRI

Choudhury

Swamy

Hadfield-Menell

et al. 2019

2019 14th ACM/IEEE International Conference on Human-Robot Interaction (HRI)

View full text Add to dashboard Cite

Fundamental to robotics is the debate between model-based and model-free learning: should the robot build an explicit model of the world, or learn a policy directly? In the context of HRI, part of the world to be modeled is the human. One option is for the robot to treat the human as a black box and learn a policy for how they act directly. But it can also model the human as an agent, and rely on a "theory of mind" to guide or bias the learning (grey box). We contribute a characterization of the performance of these methods under the optimistic case of having an ideal theory of mind, as well as under different scenarios in which the assumptions behind the robot's theory of mind for the human are wrong, as they inevitably will be in practice. We find that there is a significant sample complexity advantage to theory of mind methods and that they are more robust to covariate shift, but that when enough interaction data is available, black box approaches eventually dominate.Index Terms-theory of mind, inverse RL, model-based RL, model-free RL, sample complexity

show abstract

Multimodal Probabilistic Model-Based Planning for Human-Robot Interaction

Cited by 5 publications

References 0 publications

Probabilistically Safe Robot Planning with Confidence-Based Human Predictions

Probabilistically Safe Robot Planning with Confidence-Based Human Predictions

Multi-Agent Reinforcement Learning with Multi-Step Generative Models

On the Utility of Model Learning in HRI

Contact Info

Product

Resources

About