Powered Descent Guidance via Sequential Convex Programming with Constraint Function Design

This paper proposes a method for imposing path constraints using a “constraint function” in sequential convex programming (SCP) and applies it to the powered descent guidance (PDG) of a reusable launch vehicle (RLV). During the landing burn phase, which begins at 1-5km altitude, it is necessary to limit the angle-of-attack in the regions of high dynamic pressure to prevent excessive aerodynamic loads. In addition, the attitude of the launch vehicle should be constrained near the landing point to ensure a vertical landing. To efficiently satisfy these requirements, path constraints should be imposed based on the states of the RLV, rather than fixed constants. To address this issue, we propose the constraint function method that allows constraints to be designed as functions of state-dependent or state-independent variables. This approach enables constraints to be shaped into desired forms, ensuring smooth and continuous changes as the inputs to the constraint function vary. Moreover, it expands the constraint design domain beyond independent variables, such as time or altitude, to include state variables like dynamic pressure, thereby enhancing flexibility in the design process.