A novel form-finding method via noise-tolerant neurodynamic model for symmetric tensegrity structure.

In this paper, the form-finding of symmetric tensegrity structure is discussed and investigated by means of force density method and neural network algorithm. An optimization model based on the rank deficiency maximization is established to solve feasible force density, and the nodal coordinates are determined by eigenvalue decomposition of force density matrix. Then, the Lagrangian multiplier method can transform constrained optimization problem into unconstrained optimization problem. To avoid high-dimensional Hessian matrix calculation, noise-tolerant neural algorithm (NTNA) is exploited to calculate the unconstrained optimization problem. Numerical results indicate that the proposed noise-tolerant neural-based quasi-Newton Broyden–Fletcher–Goldfarb–Shanno (NTN-QNBFGS) form-finding method enhances convergence speed. Furthermore, the consistency between the form-finding results and analytical solutions infers that the developed method can be effectively applied to the form-finding of tensegrity robot. [ABSTRACT FROM AUTHOR]