GuSTO: Guaranteed Sequential Trajectory Optimization via Sequential Convex Programming
This work addresses trajectory optimization for control systems, offering improved guarantees and performance, though it is incremental as it builds on existing SCP methods.
The authors tackled the lack of rigorous performance guarantees in Sequential Convex Programming (SCP) for trajectory optimization by introducing GuSTO, a framework for control-affine systems with drift that generalizes earlier methods and provides convergence guarantees to stationary points. Numerical experiments showed that GuSTO outperforms state-of-the-art approaches in success rates, solution quality, and computation times.
Sequential Convex Programming (SCP) has recently seen a surge of interest as a tool for trajectory optimization. However, most available methods lack rigorous performance guarantees and they are often tailored to specific optimal control setups. In this paper, we present GuSTO (Guaranteed Sequential Trajectory Optimization), an algorithmic framework to solve trajectory optimization problems for control-affine systems with drift. GuSTO generalizes earlier SCP-based methods for trajectory optimization (by addressing, for example, goal-set constraints and problems with either fixed or free final time) and enjoys theoretical convergence guarantees in terms of convergence to, at least, a stationary point. The theoretical analysis is further leveraged to devise an accelerated implementation of GuSTO, which originally infuses ideas from indirect optimal control into an SCP context. Numerical experiments on a variety of trajectory optimization setups show that GuSTO generally outperforms current state-of-the-art approaches in terms of success rates, solution quality, and computation times.