---

In this paper, online manual guidance of industrial robots using impedance control with singularity avoidance is studied. In this method, operator enters the robot workspace, physically holds the end-effector equipped with force sensor and manually guides the robot. In doing so, the operator generates the desired trajectory for applications like welding or painting. Robot singular configuration is possible during the process which makes it unsafe due to unexpected high velocity robot joints and the physical human-robot interaction. Therefore, real-time identification of singularity position and orientation must be evaluated during trajectory generation. The use of manipulability ellipsoid is suggested as a simple method for the singularity identification. By combining the manipulability ellipsoid and impedance control, a simple and new approach is proposed to warn operator before reaching singularity. Based on the proposed approach, effect of opposite force is exerted on the human hand in the predefined distance to singularity. Real-time implementation is the main advantage of the proposed approach because it keeps robot away from reaching singularity. Real-time experiments are performed using a SCARA robot. In the first experiment, the operator stops the trajectory generation process when an opposite force is produced. In the second experiment, the operator insists on entering the singular points. Experimental results show the effectiveness of the proposed approach in dealing with singularity problem during the trajectory generation by an operator for industrial robots.

---

This paper investigates mapping of stiffness, damping and mass matrices for joint and Cartesian space in robot impedance control. The stiffness mapping is studied more than others for serial and parallel robots by lots of researchers. But all the mapping methods are considered for small displacement or static problems. In fact there is no formulation for large displacement or dynamic problems. So in the presented paper, impedance mapping is studied by considering old methods against new one. The new proposed method is called Large Displacement Mapping Method. Two main methods are presented in researchers’ works and are studied beside the new proposed one in this paper. A test is designed and simulated by Simulink/MATLAB in order to analyze different methods clearly. Then the proposed test is applied on a SCARA robot practically. According to simulation and experimental results, only the proposed method can map the stiffness, damping and mass matrices in large displacement case. The results show, as robot is getting away from the initial position, more deviation is happened. Considering impedance control mapping results, two main differences are seen in stiffness and damping matrices while mass matrix mapping result is the same in all three methods.