Learning convex bounds for linear quadratic control policy synthesis
This work addresses control policy synthesis for dynamic environments in fields like AI and robotics, but it is incremental as it builds on existing linear quadratic control frameworks.
The paper tackles the problem of learning control policies for unknown linear dynamical systems to maximize a quadratic reward function, presenting a method that optimizes expected reward over posterior system parameters and demonstrates strong performance and robustness in simulations and a real-world inverted pendulum stabilization.
Learning to make decisions from observed data in dynamic environments remains a problem of fundamental importance in a number of fields, from artificial intelligence and robotics, to medicine and finance. This paper concerns the problem of learning control policies for unknown linear dynamical systems so as to maximize a quadratic reward function. We present a method to optimize the expected value of the reward over the posterior distribution of the unknown system parameters, given data. The algorithm involves sequential convex programing, and enjoys reliable local convergence and robust stability guarantees. Numerical simulations and stabilization of a real-world inverted pendulum are used to demonstrate the approach, with strong performance and robustness properties observed in both.