Robot Trajectory Optimization Using Approximate Inference

The general stochastic optimal control (SOC) problem in robotics scenarios is often too complex to be solved exactly and in near real time. A classical approximate solution is to first compute an opti