Paper: Dual Online Stein Variational Inference for Control and Dynamics
RSS (Robotics: Science and Systems Conference) is arguably one of the most prestigious and selective conferences in robotics. This year we had 2 out of 2 submitted papers accepted to RSS!
One of them is DuSt- MPC: Dual Online Stein Variational Inference for Control and Dynamics
With Lucas Barcelos leading the show, this is the outcome of a multi-institution research collaboration between the University of Sydney, Georgia Tech, CSIRO, University of Washington and NVIDIA.
The code is available here: https://github.com/lubaroli/dust
—
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 real-world experiments requiring real-time execution in the face of dynamically changing environments.
Lucas Barcelos (University of Sydney, CSIRO), Alexander Lambert (Georgia Tech), Rafael Oliveira (University of Sydney), Paulo Borges (CSIRO), Byron Boots (University of Washington), Fabio Ramos (NVIDIA, University of Sydney), “Dual Online Stein Variational Inference for Control and Dynamics”, in Robotics: Science and Systems, 2021.
Download the full paper here.
The code is available here: https://github.com/lubaroli/dust