Shape estimation of inflatable space structures using neural network

Inflatable space structures need to maintain in a desired shape in space in order to achieve satisfactory performance. The active shape control technique has shown its advantages in solving this problem. One difficulty to realize an active control system in space is how to establish a model that reflects the structure shapes under different environment and boundary tensions. This paper proposes a neural network scheme to estimate the shape of inflatable structures. A neural network is trained to map environment information and control tensions into the structure shape. After the neural network training completes, an estimation of the structure shape can be obtained by inputting the measured environment data and control variables to the neural network. Some validation studies have been conducted in laboratory on the estimation of the flatness of a rectangular Kapton membrane. The results showed the proposed scheme gave very good estimations of the membrane flatness