Integrated structure/control design of nonrigid robot arms for high-speed manipulation

In assembly of PC boards and VLSI's, there is a great need for ultrahigh speed robots that can significantly reduce the cycle time and thereby increase productivity. To meet this demand, a technical breakthrough is necessary for arm design, which substantially departs from the traditional design paradigm. Traditionally, arm linkage and drive mechanisms are designed first followed by control design. Structure design, while it is optimal with respect to kinematics and statics, is not always optimal from the control standpoint. There exists intricate interactions between mechanisms and controls, which must be taken into account in designing the system. To achieve high performance, a cohesive and comprehensive design method that integrates both structure and control design is necessary. First, flexibility of arm links is included in the model to consider the dynamic deflections in high speed manipulation. An actuator relocation method for improving flexible arm dynamics is presented. The endpoint control of a flexible arm has been known as a nonminimum phase system due to the noncollocated sensor and actuator. By relocating the actuator near the endpoint, the system can be changed to a minimum phase system. Pole-zero locations of a single-link flexible arm are obtained for the actuator types (force and torque) and the actuator locations. Pole-zero configurations and their effect on control are analyzed with regard to the location of the actuation point. The guidelines for the actuator locations are developed with respect to the operation bandwidth, stability, and controllability. An integrated approach to the concurrent design of arm structure and control is presented. To achieve high speed positioning, a technique is developed in which comprehensive design parameters describing arm link geometry, actuator locations, and feedback gains are optimized with respect to the settling time of the system. A two-link, nonrigid arm is analyzed and a simple dynamic model representing rapid positioning processes is obtained. Optimal feedback gains minimizing the settling times are obtained as functions of structural parameters involved in the dynamic model. The structural parameters are then optimized using a nonlinear programming technique in order to obtain an overall optimal performance. <The resultant arm design shows an outstanding performance, which is otherwise unattainable if one designs the structure and control separately.