Adaptive Multiresolution Collocation-Based Sequential Convex Programming for Fuel-Optimal Low-Thrust Transfer Orbit Guidance

The minimum fuel transfer problem in low-thrust trajectory optimization remains a major challenge and is typically addressed using bang-bang control. A novel methodology integrating Adaptive Multiresolution Collocation (AMRC) and Sequential Convex Programming (SCP) to solve the minimum-fuel low-thrust trajectory optimization problem is proposed. First, the approach employs the cubic spline wavelet-like transform for mesh refinement, where wavelet coefficients serve as error indicators to dynamically concentrate nodes in regions of rapid state variation. Then, the nonlinear programming problem is convexified via control variable relaxation and small-perturbation linearization, reformulated as a second-order cone programming (SOCP) problem, and efficiently solved using convex optimization tools. Subsequently, progressive selection of the location points ensures rapid and accurate convergence to the optimal trajectory. Finally, numerical simulations of Earth–Mars and Earth–Venus transfer validate the effectiveness and accuracy of the AMRC-based method. Compared with conventional approaches, the proposed method achieves comparable optimality while markedly improving computational efficiency, precisely localizing switching times, and improving numerical precision, requiring only 29.7% of the nodes and 14.7% of the computation time of uniform-grid convex optimization, achieving fuel-optimal deviations within 0.07% of the indirect method and demonstrating accuracy improvements of 2–3 orders of magnitude over GPOPS.