Abstract
Successive convex programming is a promising technique for onboard applications thanks to its speed and guaranteed convergence. Hence it can be an enabler for future missions where spacecraft autonomy plays a key role. The definition of a good value of the trust region plays a vital role in the successful convergence of SCVX algorithms. This work presents an improved trust region algorithm based on a differential algebra technique that relies on the information given by the nonlinearities of the constraints and does not depend on the user for the initialization of the trust region.