Analytical Delta-V Approximation for Nonlinear Programming of Multi-target Rendezvous and Flyby Trajectories

TL;DR

This paper introduces an analytical Delta-V approximation method for multi-target rendezvous and flyby trajectories that reduces computational effort and integrates with gradient-based nonlinear programming for efficient sequence optimization.

Contribution

It develops an analytical approach based on linear relative motion equations to approximate Delta-V, avoiding iterative Lambert's solutions and enabling faster, gradient-based trajectory optimization.

Findings

Analytical Delta-V approximation reduces computation time.

Gradient-based algorithms achieve similar results faster.

Errors remain acceptable for close orbit transfers.

Abstract

This study proposes an analytical Delta-V approximation of short-time transfers based on the linear relative motion and a gradient-based nonlinear programming model of multi-target rendezvous and flyby trajectories. In previous studies, the Lambert's solution is commonly used to evaluate Delta-V of short-duration transfers. In this study, to avoid the iteration process for obtaining the Lambert's solution and its gradient, the linear relative motion equations are applied to form an analytical two-point boundary value model for the near-circular orbit rendezvous problems. Although the relative motion equations are usually applicable when the two orbits are close enough, and the position and velocity errors would become more significant as the orbital differences increase, the errors of the velocity increments were proved acceptable in our simulations. Moreover, the analytical formula facilitates the calculation of the gradients to the start epoch and flight time, which are used to establish a nonlinear programming model for sequence optimization that gradient-based algorithms can easily solve. Simulation results demonstrated that the analytical Delta-V approximation requires much less calculation than the Lambert's solution, and the proposed gradient-based nonlinear programming algorithms can obtain similar results in less time than previous methods.