On stochastic optimal control and reinforcement learning by approximate inference
We present a reformulation of the stochastic optimal control problem in terms of KL
divergence minimisation, not only providing a unifying perspective of previous approaches …
divergence minimisation, not only providing a unifying perspective of previous approaches …
Probabilistic planning with sequential monte carlo methods
In this work, we propose a novel formulation of planning which views it as a probabilistic
inference problem over future optimal trajectories. This enables us to use sampling methods …
inference problem over future optimal trajectories. This enables us to use sampling methods …
Toward asymptotically optimal motion planning for kinodynamic systems using a two-point boundary value problem solver
We present an approach for asymptotically optimal motion planning for kinodynamic
systems with arbitrary nonlinear dynamics amid obstacles. Optimal sampling-based …
systems with arbitrary nonlinear dynamics amid obstacles. Optimal sampling-based …
Space-time functional gradient optimization for motion planning
Functional gradient algorithms (eg CHOMP) have recently shown great promise for
producing locally optimal motion for complex many degree-of-freedom robots. A key …
producing locally optimal motion for complex many degree-of-freedom robots. A key …
A real-world application of Markov chain Monte Carlo method for Bayesian trajectory control of a robotic manipulator
Reinforcement learning methods are being applied to control problems in robotics domain.
These algorithms are well suited for dealing with the continuous large scale state spaces in …
These algorithms are well suited for dealing with the continuous large scale state spaces in …
Topology-based representations for motion planning and generalization in dynamic environments with interactions
V Ivan, D Zarubin, M Toussaint… - … Journal of Robotics …, 2013 - journals.sagepub.com
Motion can be described in several alternative representations, including joint configuration
or end-effector spaces, but also more complex topology-based representations that imply a …
or end-effector spaces, but also more complex topology-based representations that imply a …
Outcome-driven reinforcement learning via variational inference
While reinforcement learning algorithms provide automated acquisition of optimal policies,
practical application of such methods requires a number of design decisions, such as …
practical application of such methods requires a number of design decisions, such as …
Extended LQR: Locally-optimal feedback control for systems with non-linear dynamics and non-quadratic cost
J Van Den Berg - Robotics Research: The 16th International Symposium …, 2016 - Springer
We present Extended LQR, a novel approach for locally-optimal control for robots with non-
linear dynamics and non-quadratic cost functions. Our formulation is conceptually different …
linear dynamics and non-quadratic cost functions. Our formulation is conceptually different …
Exploiting variable physical damping in rapid movement tasks
Until now, design of variable physical impedance actuators (VIAs) has been limited mainly to
realising variable stiffness while other components of impedance shaping, such as damping …
realising variable stiffness while other components of impedance shaping, such as damping …
Black-box policy search with probabilistic programs
In this work we show how to represent policies as programs: that is, as stochastic simulators
with tunable parameters. To learn the parameters of such policies we develop connections …
with tunable parameters. To learn the parameters of such policies we develop connections …