Low-Thrust Transfer Orbit Optimization Using Sequential Convex Programming and Adaptive Mesh Refinement

In this paper, a mesh-adaptive Hermite–Simpson sequential convex programming (SCP) method is proposed for solving low-thrust orbit transfer problems efficiently and accurately. First, we develop a Hermite–Simpson convex programming method that utilizes quadratic polynomial control interpolation to improve discretization accuracy. Then, a mesh-adaptive Hermite–Simpson SCP method is proposed by incorporating an improved adaptive mesh refinement method into the Hermite–Simpson SCP method for solving optimal control accurately and efficiently. The proposed method is applied to two-dimensional time-optimal and three-dimensional fuel-optimal low-thrust orbit transfer problems to show its features, and its performance is compared against several existing methods. Numerical simulations show that the proposed method achieves more accurate solutions while being computationally more efficient compared to the trapezoidal SCP method. In comparison to the GPOPS-II software, the proposed method produces solutions of comparable accuracy but is more computationally efficient.