Local Path Planning Using Optimal Control Techniques

摘要

The ability of an autonomous vehicle control system to plan a safe, collision-free local path from one vehicle position to another is one of the most important functions. In this thesis, it is shown how a safe obstacle-free local path can be planned using optimal control theory and optimization techniques. The problem is posed as a two point boundary value problem with various problem constraints which control the vehicle behavior in transversing from one point to another. The objective function being minimized is a control performance index which includes vehicle energy saving parameters. Numerous fixed and moving obstacles in the dive plane are introduced and successfully avoided using this technique. Three-dimensional path planning is also successfully demonstrated on a 12 state linear model of an underwater vehicle. This technique is shown to be feasible method for a local path planning applications. (KR)