Efficient Long-Range Lunar Descent Trajectory Generation with Continuous-Time Sequential Convex Programming

A continuous-time sequential convex programming (SCP) algorithm is applied to the minimum fuel, six degree-of-freedom powered descent guidance problem with free-final time and compound altitude-triggered hazard detection and avoidance constraints. Compared to standard discrete-time sequential convex programming algorithms, this reformulation ensures that state constraints are satisfied everywhere along the trajectory, independent of mesh density. We demonstrate that the recently developed continuous-time compound state-triggered constraints algorithm (CT-cSTC) has potential to solve the entire lunar powered descent and landing trajectory with a coarser discretization mesh than discrete-time algorithms, making it more computationally tractable to solve the entire trajectory in one-shot rather than solving each phase sequentially. In this paper, two scenarios are evaluated: a 4 km short-range descent and a 150 km long-range braking burn to landing profile. Three discrete-time SCP baseline methods--two Penalized Trust Region (PTR) and an adaptive trust-region variant--serve as benchmarks. We show that CT-cSTC is capable of solving the long-range trajectory with a coarser discretization and a quicker solve time than three discrete-time compound state-triggered constraint algorithms (DT-cSTC).