Integrating structure, information architecture and control design: Application to tensegrity systems. (December 2021)

• Record Type: Journal Article
• Authors: Goyal, Raman; Majji, Manoranjan; Skelton, Robert E.
• Is Part Of: Mechanical systems and signal processing. Volume 161(2021)
• Publication Date: 2021-12
• Subjects: Integrated system design -- Sensor/actuator placement -- Multistable systems -- Tensegrity structure -- Minimum-order representation -- LMIs -- Convexification; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2021.107913
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: A system design approach by merging structure, actuator/sensor precision and control. An LMI based convexification scheme to solve the complete non-convex problem. A bounded covariance control system design example for a tensegrity structure. An important step to finding optimal set of structure parameters & actuators/sensors. Minimal-order representation for the tensegrity by removing the uncontrollable modes. Abstract: A novel unified approach to jointly optimize structural design parameters, actuator and sensor precision, and controller parameters is presented in this paper. The joint optimization problem is posed as a covariance control problem, where feasibility is achieved by bounding the covariance of the output as well as that of the control signals. The formulation is used to design a tensegrity system, where the initial prestress parameters, sensor and actuator precisions, and the control law are jointly optimized. Tensegrity system dynamics models linearized about an equilibrium point are used for system design, where minimality is ensured by constraint projection. The feedback loop is assumed to have a full-order dynamic compensator with its characteristic matrices chosen as optimization variables. The suboptimal solution of this non-convex system design problem is found by iterating over an approximated convex problem through the use of a convexifying potential function that enables the convergence to a stationary point. It is shown that for a linear dynamical system, the approximated joint optimization problem can be formulated using Linear Matrix Inequalities (LMIs).