Motion Control Algorithms for Sensor-Equipped Robots

This paper deals with the development of kinematic algorithms for the control of sensor-equipped robots. The kinematics is solved in the sensor coordinate system, which reduces the computation efforts, and allows the elimination of the first joint encoder. Simplification of the algorithms can be obtained when approximations are used to solve the inverse kinematics. Three control algorithms based on approxima­ tions are presented. However, with these algorithms, convergence to the target is not always guaranteed. A Theorem which specifies the sufficient conditions required for a trajectory to converge to a target point is proved. Based on this Theorem robot parameters can be selected in the design stage of the manipulator. This is illustrated for several types of manipulators. This paper deals with kinematical features of intelligent robots having a sensor located at the arm. A typical sensor might be a force-torque transducer or a camera located at the wrist section. With such robots, the motion commands are generated by the sensor and the robot must execute them in real time. This requires that the robot algorithms include con­ trol in sensor-oriented coordinates, rather than in world coor­ dinates as in conventional robots. It is claimed that control in sensor (or object) oriented coordinates may be comparable in complexity to that required in conventional robots [1], In this paper, it is shown that sensor-oriented control may be much simpler, and also much faster, compared with the control in world coordinate system. Based upon a kinematic analysis, the paper proposes mo­ tion algorithms for sensor-equipped robots. Simplification of the algorithms can be obtained when approximations are used to solve the inverse kinematics. However, when using the ap­ proximations the robot might diverge from the target point. Therefore, sufficient conditions which guarantee the con­ vergence in this case must be found. The kinematic approach in this paper is based upon the resolved motion rate control method [2, 3] in which the re­ quired velocity of the end effector, s, is related to the joint variables 0, by the equation