Developing intelligent tensegrity structures with stochastic search

Since structural control of civil structures was first proposed in 1972, most research and applications have focused on enhancing the safety of structures under extreme conditions. This paper introduces a new direction in structural control: the use of computational methods and explicitly defined knowledge to improve serviceability and maintenance of civil structures. The objectives of such structures, called intelligent structures, are to maintain and improve structural performance by recognizing changes in behaviors and loads, adapting the structural geometry to meet defined goals, and using past events to improve future performance. A computational framework based on intelligent control methodology is presented that combines reasoning from explicit knowledge, search, learning and planning to illustrate a vision for intelligently controlled civil structures. This is applied to enable global shape control of an adjustable tensegrity structure using a combination of simulated annealing search, dynamic relaxation analysis and structural measurements. Active tensegrity structures are modular, reusable cable structures that do not require expensive anchorages or steep slopes. This provides an innovative solution for temporary structures. Particularly, when used for exhibitions, they could become part of an exhibition, rather than just house one. As part of a multi-stage project, results presented in this paper demonstrate the feasibility of building an intelligent tensegrity structure.