KrisLibrary
1.0.0
|
A cost that measures quadratic tracking error and control cost. Note that the time must exist in the state. More...
#include <Objective.h>
Public Member Functions | |
QuadraticObjective (int timeIndex=0) | |
virtual const char * | TypeString () |
Subclasses: return an identifier for this goal type. | |
virtual Real | TerminalCost (const Vector &qend) |
virtual Real | DifferentialCost (const State &x, const ControlInput &u) |
Subclasses must override this. | |
Public Member Functions inherited from IntegratorObjectiveFunctional | |
IntegratorObjectiveFunctional (Real dt=0.1, int timeIndex=-1) | |
virtual Real | Domain (const ControlInput &u, const Interpolator *path) |
Subclasses should override this if timeIndex < 0. | |
virtual Real | IncrementalCost (const ControlInput &u, const Interpolator *path) |
This is implemented for you. | |
Public Member Functions inherited from ObjectiveFunctionalBase | |
virtual std::string | Description () |
Subclasses: return a string for printing (optional) | |
virtual Real | IncrementalCost (const Interpolator *path) |
virtual Real | IncrementalCost (const KinodynamicMilestonePath &path) |
virtual Real | TerminalCost (const Config &qend) |
Subclasses: return the cost of a terminal state. | |
virtual bool | PathInvariant () const |
Subclasses: planners may exploit path-invariant costs for faster performance. | |
virtual Real | PathCost (const MilestonePath &path) |
virtual Real | PathCost (const KinodynamicMilestonePath &path) |
virtual bool | SaveParams (AnyCollection &collection) |
Subclasses: read and write parameters to collection. | |
virtual bool | LoadParams (AnyCollection &collection) |
Public Attributes | |
InterpolatorPtr | desiredPath |
Math::Matrix | stateCostMatrix |
Math::Matrix | controlCostMatrix |
Math::Matrix | terminalCostMatrix |
Public Attributes inherited from IntegratorObjectiveFunctional | |
Real | dt |
int | timeIndex |
A cost that measures quadratic tracking error and control cost. Note that the time must exist in the state.
The error functional has differential cost L(x,u,t) = (x-xdes(t))^T P (x-xdes(t)) + u^T Q u with P = stateCostMatrix and Q = controlCost matrix.
Terminal cost is Phi(x) = (x-xdes(T))^T R (x-xdes(T)) with R = terminalCostMatrix.