Environment stiffness estimation with multiple observers

When a compliant robotic manipulator interacts with the environment, forces are generated at the contact points, which are a function of environment parameters (such as stiffness, damping and friction). The resulting contact forces can cause a bouncing effect and eventually instability of the manipulator. This problem assumes a great importance when the environment is unknown or poorly structured, such as in robotic-assisted minimally invasive surgery (MIS). Stability has to be guaranteed when the robot interacts with stiff environments, such as touching a bone or another surgical instrument. To accomplish that, it is necessary to develop a mechanism that estimates in real time the relevant environment parameters to make the corresponding adaptation of control laws possible. In this paper, an algorithm for online stiffness estimation for compliant robotic manipulation is presented. This online algorithm receives the system command input and measured force, generating an estimation of the system stiffness. The method does not require persistent excitation, nor position and velocity measurements. It has low computational overhead, making it suitable for online operation. Simulation and experimental results are presented, validating the approach.