A Robot Human-Like Learning Framework Applied to Unknown Environment Interaction

Algorithm 1

Framework of the adaptive optimal admittance scheme.

	Input: the set of initial feedback gain K0 and state variable ξ;
	Output: optimal feedback gain KK;
	phase 1: set fint0 = K0ξ as the initial input while the manipulator is contacting with the environment;
	Repeat: compute ξξ, Iξξ, and Iξu;
	Until: rank[Iξξ, Iξu] = m(m+1)/2 + mr;
	Repeat:
	phase 2: solve PK and KK+1 according to:
	Until: ||PK-PK-1 ||<ε;
	Return: KK;