Online BayesSim for Combined Simulator Parameter Inference and Policy Improvement

Possas, Rafael; Barcelos, Lucas; Oliveira, Rafael; Fox, Dieter; Ramos, Fábio

doi:10.1109/iros45743.2020.9341401

Cited by 9 publications

(13 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A variety of modelling representations have been explored, including Gaussian processes [8], neural network ensembles [7], Bayesian regression, and meta-learning [16]. Alternatively, the authors in [28,26] estimate posterior distributions of physical parameters for black-box simulators, given real-world observations.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Dual Online Stein Variational Inference for Control and Dynamics

Barcelos¹,

Lambert²,

Oliveira³

et al. 2021

Robotics: Science and Systems XVII

Self Cite

View full text Add to dashboard Cite

Model predictive control (MPC) schemes have a proven track record for delivering aggressive and robust performance in many challenging control tasks, coping with nonlinear system dynamics, constraints, and observational noise. Despite their success, these methods often rely on simple control distributions, which can limit their performance in highly uncertain and complex environments. MPC frameworks must be able to accommodate changing distributions over system parameters, based on the most recent measurements. In this paper, we devise an implicit variational inference algorithm able to estimate distributions over model parameters and control inputs on-the-fly. The method incorporates Stein Variational gradient descent to approximate the target distributions as a collection of particles, and performs updates based on a Bayesian formulation. This enables the approximation of complex multi-modal posterior distributions, typically occurring in challenging and realistic robot navigation tasks. We demonstrate our approach on both simulated and realworld experiments requiring real-time execution in the face of dynamically changing environments.

show abstract

Section: Related Workmentioning

confidence: 99%

“…Conversely, previous work has demonstrated that incorporating uncertainty in the evaluation of SOC estimates can improve performance [3], particularly when this uncertainty is periodically re-estimated [26]. Although this method is more robust to model mismatch and help address the sim-to-real gap, Step 0…”

Section: Introductionmentioning

confidence: 99%

Dual Online Stein Variational Inference for Control and Dynamics

Barcelos¹,

Lambert²,

Oliveira³

et al. 2021

Robotics: Science and Systems XVII

Self Cite

View full text Add to dashboard Cite

show abstract

“…BayesSim [7] is a likelihood-free method that has been applied to a variety of robotics problems [32], [8], [33], [34], [35]. It offers a principled way of obtaining posteriors over simulation parameters, and does not place restrictions on simulator type of properties, i.e.…”

Section: B Probabilistic Parameter Inferencementioning

confidence: 99%

A Bayesian Treatment of Real-to-Sim for Deformable Object Manipulation

Antonova¹,

Yang²,

Sundaresan³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Deformable object manipulation remains a challenging task in robotics research. Conventional techniques for parameter inference and state estimation typically rely on a precise definition of the state space and its dynamics. While this is appropriate for rigid objects and robot states, it is challenging to define the state space of a deformable object and how it evolves in time. In this work, we pose the problem of inferring physical parameters of deformable objects as a probabilistic inference task defined with a simulator. We propose a novel methodology for extracting state information from image sequences via a technique to represent the state of a deformable object as a distribution embedding. This allows to incorporate noisy state observations directly into modern Bayesian simulation-based inference tools in a principled manner. Our experiments confirm that we can estimate posterior distributions of physical properties, such as elasticity, friction and scale of highly deformable objects, such as cloth and ropes. Overall, our method addresses the real-to-sim problem probabilistically and helps to better represent the evolution of the state of deformable objects.

show abstract

“…Possas et al [19] take an alternative approach, where the simulator is treated as a black-box, and real-world data is used to infer the parameters of the simulator online as a robot is learning to execute a task. Antonova et al [2] propose a related method, where inference is done with Gaussian Processes with kernels learned from simulation data.…”

Section: On Bridging the Reality Gap With Improved Models And Real-wo...mentioning

confidence: 99%

“…This section summarizes the findings and conclusions from the workshop, which we classify as belonging to either a practitioner's or a researcher's perspective. For the practitioner's perspective, we aim to give concrete advice on applying existing Sim2Real techniques, their capabilities and their limitations 19 . For the researcher's perspective, we state fundamental research problems identified during the workshop, hoping to inspire future research endeavors.…”

Section: Key Directions Forwardmentioning

confidence: 99%

Perspectives on Sim2Real Transfer for Robotics: A Summary of the R:SS 2020 Workshop

Höfer,

Bekris,

Handa

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

This report presents the debates, posters, and discussions of the Sim2Real workshop held in conjunction with the 2020 edition of the "Robotics: Science and System" conference. Twelve leaders of the field took competing debate positions on the definition, viability, and importance of transferring skills from simulation to the real world in the context of robotics problems. The debaters also joined a large panel discussion, answering audience questions and outlining the future of Sim2Real in robotics. Furthermore, we invited extended abstracts to this workshop which are summarized in this report. Based on the workshop, this report concludes with directions for practitioners exploiting this technology and for researchers further exploring open problems in this area.

show abstract

Online BayesSim for Combined Simulator Parameter Inference and Policy Improvement

Cited by 9 publications

References 19 publications

Dual Online Stein Variational Inference for Control and Dynamics

Dual Online Stein Variational Inference for Control and Dynamics

A Bayesian Treatment of Real-to-Sim for Deformable Object Manipulation

Perspectives on Sim2Real Transfer for Robotics: A Summary of the R:SS 2020 Workshop

Contact Info

Product

Resources

About