Method and apparatus for adaptive force and position control of manipulators

A hybrid controller in a system for controlling the positioning of, and force exerted by, a controlled manipulator in accordance with a received command signal, said system characterized in that the manipulator and the environment form a complex, dynamic model specified as having unknown parameter values, and the system is further characterized by:

• adaptive control means for issuing a real-time variable signal to said manipulator to be controlled, which signal controls the position and force of said manipulator, said adaptive control means further comprising;
• means for sensing the actual force and positioning of said manipulator as it is being controlled;
• and
• variable control signal generating means connected to said sensing means for modifying said received command by said sensed force and position in order to supply said variable control signal issued from said adaptive control means without knowledge of said parameter values of the system.

A controller for a manipulator wherein the manipulator and the environment form a complex, dynamic model with unknown parameter values, in which there exists cross-coupling between position and force variables, the improvement characterized by:

• position control means for controlling manipulator position;
• force control means for controlling manipulator force;
• and
• adaptive control means for compensating for said cross-coupling which exists between the position and force variables applied at said manipulator by said position and force control means.

A method for controlling a manipulator, wherein the manipulator and the environment form a complex, dynamic model with specifiable but unknown parameter values, the improvement characterized by:

• implementing an adaptive position control loop to compensate for the specifiable parameter values, which control loop operates from the standpont of Cartesian space and without knowledge of said parameter values based upon a position feedback signal which is representative of the manipulator's position in said Cartesian space;
• implementing an adaptive force control loop, which control loop operates from the standpoint of Cartesian space and without knowledge of said parameter values based upon a force feedback signal which is representative of the manipulator's force in said Cartesian space;
• and
• connecting both control loops to the manipulator to be controlled so that the manipulator's positioning and it's force is controlled in Cartesian space by both of said two control loops.

In a hybrid controller adapted to issue a control signal which moves a manipulator's position and force in an environment, wherein the manipulator and environment form a system exhibiting dynamics which are non-linear and have specifiable but unknown parameters, an improved control method comprising:

• controlling said manipulator in said environment in accordance with a desired position/force response;
• and
• adaptively varying said controlling step by adaptive position control and/or adaptive force control;
• and
• compensating in real-time for the system's nonlinearities in said adaptively varying control step without foreordained knowledge of

A controller for a manipulator wherein the manipulator and the environment form a complex, dynamic model, the improvement characterized by:

• adaptive position and adaptive force control means controlling a common load in accordance with the position and force control laws respectively identified by the following algorithms;
• F.sub.y (t)=f(t)+K.sub.p (t)E.sub.p (t)+K.sub.v (t)E.sub.p (t)+C(t)R(t)+B(t)R(t)+A(t)R(t)
• wherein the term R(t) is a desired position trajectory, E.sub.p (t) is a position tracking error term, and are terms contributed by feedback and feedforward adaptive position control loops;
• and
• wherein said force control law is characterized as:
• ##EQU38## wherein P.sub.r (t) is said desired force reference signal term, d(t) is an auxiliary force signal term emitted by said adaptation control, and the K.sub.I, K.sub.p, and K.sub.v are variable gain terms;
• E is an error term;
• and Z is a velocity term.