localCornerControl

Solve optimal control problem for the corner states of the parallelotope

Contents

Syntax

[uTotal, xt]= localCornerControl(system,xf,vert,h,Q,R,steps,lenHorizon,Opts)

Description

This function solves an optimal control problem for all vertices of the initial parallelotope (see Lines 6-8 of Alg. 1 in [1]).

Input Arguments

system

object containing the system dynamics (class:; nonlinearSys)

xf

desired final state at the end of the control process (dimension: [nx, 1])

vert

vertices of the initial parallelotope. (dimension: [nu,2^nx])

h

length of one timestep of the multiple shooting algorithm

Q

weighting matrix for the final state of the optimal control problem (dimension: [nx,nx])

R

weighting matrix for the input term of the optimal control problem (dimension: [nu,nu])

steps

number of intermediate timesteps of the corner trajectories during one timestep of the center trajectory

lenHorizon

length of the optimization horizon in center trajectory time steps

Opts

a structure containing following options

.parallel

boolean value that determines if parallel computing should be used or not (0 or 1)

.useAcado

use ACADO toolbox for solving the optimal control problem. Fmincon is used otherwise (0 or 1)

.extHorizon.decay

decay function for the objective function of the optimization problem with extended optimization horizon ['uniform' / 'fall' / 'fall+end' /; 'fallLinear' / 'fallLinear+End' /; 'fallEqDiff' / 'FallEqDiff+End' /; 'rise' / 'quad' / 'riseLinear' /; 'riseEqDiff' / 'end']

Output Arguments

uTotal

optimal control inputs for the corner trajectories (dimension: [nu,2^nx,steps*lenHorizon])

xt

resulting corner trajectories (dimension: [nx,2^nx,steps*lenHorizon])

See Also

convexInterpolationControl, optimalControl, optimalControlFmincon

References


© 2018-2020 I6 Technische Universität München Website License

logoAroc logoCora logoChair logoTum