Fixed-wing UAVs are excellent platforms for persistent surveillance type applications. Their superiority with respect to range, payload, and endurance.
It may be rightly argued that path planning to a goal state from an arbitrary initial condition has been exhaustively studied, as evidenced by comprehensive surveys by LaValle and Choset. However, the complete coverage requirement of the persistent surveillance application greatly increases the problem complexity. The consequence is that the designer must be willing to tolerate approximate or heuristic solutions to the general coverage problem. Practical design space remains for applications that must run on embedded systems.
Parametric polynomials provide an alternate approach that retains algorithmic fidelity to underlying physics of flight in a more computationally attractive framework than the typical equations of motion used in analyses. We will discuss techniques that originate in computer graphics and animation applications and outline how they can be tailored for autonomous vehicle path planning. While offering great advantages, these techniques also have a few pitfalls for the unsuspecting developer, so the pros and cons will be discussed. An overview of a number of spline-based approaches will be documented.