Torque-position transformer for task control of position controlled robots

• US 20060250101A1
• Filed: 04/13/2006
• Published: 11/09/2006
• Est. Priority Date: 04/13/2005
• Status: Active Grant

• Current Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Original Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Inventors: Thaulad, Peter, Khatib, Oussama, Park, Jaehoung
• Granted Patent: US 7,211,979 B2
• Time in Patent Office: Days
• US Class Current: 318/568.200
• CPC Class Codes: B25J 9/1633 compliant, force, torque control, e.g. combined with position control
• CPC Class Codes: G05B 2219/39206 Joint space position control
• CPC Class Codes: G05B 2219/39261 Calculate driving torque from dynamic model, computed torque method variant

Abstract

Torque control capability is provided to a position controlled robot by calculating joint position inputs from transformation of the desired joint torques. This is based on calculating the transfer function 1/E(s), which relates the desired joint torque to joint position. Here E(s) is a servo transfer function D(s) or an effective servo transfer function D*(s). The use of an effective servo transfer function D*s) is helpful in cases where joint nonlinearities are significant. The effective servo transfer function D*(s) is defined with respect to an ideal joint transfer function G*(s)=1/(I_effs²+b_effs) , where I_eff is an effective moment of inertia and b_eff is an effective damping coefficient.

Claims

1. A robotic control method for a position controlled robot, the method comprising:

a) determining a closed loop position transfer function T(s) of a selected joint of the robot relating a commanded input joint position Q_in(s) to an actual joint position Q_out(s) via Q_out(s)=T(S)Q_in(S);

b) selecting a dynamic open loop joint torque to position transfer function H(s) relating to the selected joint;

c) determining a corresponding effective servo transfer function E(s) of the selected joint according to E(s)=T(s)/(H(s)−

T(s)H(s));

d) providing a desired dynamic torque τ

_des(S) of the selected joint; and

e) determining Q_in(s) of the selected joint corresponding to τ

_des(S) via Q_in(S)=τ

_des(S)/(E(s)(1−

T(s)).

2. The method of claim 1, wherein said transfer function H(s) is a physical transfer function of said selected joint.

3. The method of claim 1, wherein said transfer function H(s) is an idealized transfer function of said selected joint.

4. The method of claim 3, wherein H(s) is given by H(s)=1/(I_effs²+b_effs), where I_eff is an effective moment of inertia of said selected joint, and b_eff is an effective damping coefficient of said selected joint.

5. The method of claim 4, wherein said transfer function H(s) differs substantially from a physical transfer function G(s) of said selected joint.

6. A robotic control method for a position controlled robot, the method comprising:

selecting two or more joints of the robot; and

providing desired torque inputs to each of the selected joints as corresponding position inputs according to the method of claim 1.

7. A robotic control system comprising:

a) a position controlled robot having a closed loop position transfer function T(s) of a selected joint of the robot relating a commanded input joint position Q_in(s) to an actual joint position Q_out(s) via Q_out(s)=T(s)Q_in(s);

b) a processor including a torque to position converter, wherein said torque to position converter is capable of providing Q_in(s) corresponding to a desired dynamic torque τ

_des(s) of the selected joint according to a method comprising;

selecting a dynamic open loop joint torque to position transfer function H(s) relating to the selected joint;

determining a corresponding effective servo transfer function E(s) of the selected joint according to E(s)=T(s)/(H(s)-T(s)H(s)); and

determining Q_in(s) of the selected joint corresponding to τ

_des(s) via Q_in(s)=τ

_des(s)/(E(s) (1−

T(s))

8. The system of claim 7, wherein said transfer function H(s) is a physical transfer function of said selected joint.

9. The system of claim 7, wherein said transfer function H(s) is an idealized transfer function of said selected joint.

10. The system of claim 9, wherein H(s) is given by H(s)=1/(I_effs²+b_effs), where I_eff is an effective moment of inertia of said selected joint, and b_eff is an effective damping coefficient of said selected joint.

11. The system of claim 10, wherein said transfer function H(s) differs substantially from a physical transfer function G(s) of said selected joint.