KDL::Binary< A, B, OpID > Class Template Reference

A type to specify the implementation of a binary operation. More...

#include <jacobianexpr.hpp>

Public Types
typedef MyOp::derivType	derivType
typedef MyOp::exprType	exprType
typedef BinaryOp< OpID, typename A::exprType, typename B::exprType >	MyOp
typedef MyOp::valueType	valueType
Public Member Functions
INLINE	Binary (const A &_a, const B &_b)
INLINE derivType	deriv (int i) const
INLINE JacobianBitset	getDerivs () const
INLINE bool	hasDeriv (int i) const
INLINE int	nrOfDeriv () const
INLINE valueType	value () const
Public Attributes
A	a
B	b
JacobianBitset	hasderiv

Detailed Description

template<class A, class B, class OpID>
class KDL::Binary< A, B, OpID >

A type to specify the implementation of a binary operation.

An implementation of a binary operation constructs a template specialization. You can see this as a compile-time lookup table that finds the correct operation depending on the operator and the type of its two arguments.

Template arguments: -A is the type of the value of the argument 1. -B is the type of the value of the argument 2.

The marker class OpID is a type that is created to specify a binary operator. For example, the binary + operator is specified by OpAdd. There can only be one specification type for each operator. E.g. defining OpSum also created with the + operator for other types A and B will lead to compile-time errors. This type can be created with an empty class declaration.

Specializations should define the following :

a type valueType defining the type of the value of the result of this unary operator.

a type derivType defining the type of the derivative of the result of this unary operator.

a method valueType value(const A& arg1,const B& arg2)
a method derivType derivVV(const A& arg1, const derivA& darg1, const B& arg2, const derivB& darg2).
additional methods derivVC, derivCV.

In the above method names, V stands for variable, C stands for constant (ie derivatives==0). B corresponds the type of the derivative of arg. (in general: GeomTraits<A>::derivType, you can fill in the specific type in these classes ). derivCC does not need to exist because it is known that in that case the derivatives are always zero.

For example

		template <>
		class BinaryOp<OpMult,double,double> {
		public:
			typedef double valueType;
			typedef double derivType;
			INLINE2 static double value(double a,double b)  {
				return a*b;
			}
			INLINE2 static double derivVV(double a,double da,double b,double db) {
				return a*db + da*b;
			}
			INLINE2 static derivType derivCV(double a,double b,double db) {
				return a*db;
			}
			INLINE2 static derivType derivVC(double a,double da,double b) {
				return da*b;
			}
		};

Definition at line 296 of file jacobianexpr.hpp.