Class ct::optcon::CTSystemModel< STATE_DIM, CONTROL_DIM, SCALAR >: this needs to get unified with the system-interface from optcon
Member ct::optcon::loadMpcSettings (const std::string &filename, mpc_settings &settings): make part of mpc_settings struct
Class ct::optcon::LQOCProblem< STATE_DIM, CONTROL_DIM, SCALAR >: Refactor the initializing methods such that const-references can be handed over.
Class ct::optcon::LQOCSolver< STATE_DIM, CONTROL_DIM, SCALAR >: uncouple from NLOptConSettings
Member ct::optcon::NLOCBackendBase< STATE_DIM, CONTROL_DIM, P_DIM, V_DIM, SCALAR, CONTINUOUS >::getControlTrajectory () const: this method currently copies the time array (suboptimal)
Member ct::optcon::NLOCBackendBase< STATE_DIM, CONTROL_DIM, P_DIM, V_DIM, SCALAR, CONTINUOUS >::getStateTrajectory () const: this method currently copies the time array (suboptimal)
Member ct::optcon::NLOCBackendBase< STATE_DIM, CONTROL_DIM, P_DIM, V_DIM, SCALAR, CONTINUOUS >::printSummary (): the printing of the smallest eigenvalue is hacky
Member ct::optcon::NLOCBackendBase< STATE_DIM, CONTROL_DIM, P_DIM, V_DIM, SCALAR, CONTINUOUS >::rolloutSingleShot (const size_t threadId, const size_t k, ControlVectorArray &u_ff_local, StateVectorArray &x_local, const StateVectorArray &x_ref_lqr, StateVectorArray &xShot, StateSubsteps &substepsX, ControlSubsteps &substepsU, std::atomic_bool *terminationFlag=nullptr) const: : here we override the state trajectory directly (as passed by reference). This is bad.
Member ct::optcon::NLOptConSettings::computeK (double timeHorizon) const: naming it K is confusing, should better be N.
Member ct::optcon::NLOptConSolver< STATE_DIM, CONTROL_DIM, P_DIM, V_DIM, SCALAR, CONTINUOUS >::getBackend (): this is not optimal, allows the user too much access)
Member ct::optcon::SingleShooting< STATE_DIM, CONTROL_DIM, P_DIM, V_DIM, SCALAR, CONTINUOUS >::finishMPCIteration () override: : in MPC always returning true. Unclear how user wants to deal with varying costs, etc.
Class ct::optcon::TermStateBarrier< STATE_DIM, CONTROL_DIM, SCALAR_EVAL, SCALAR >: implement sigmoid barriers
Class SummaryAllIterations< SCALAR >: also log settings here