1: \begin{abstract}

2: %Graph rigidity is an important property for formation control and network localization. In this paper we consider a problem of control and maintenance of  the rigidity for a group of mobile robots. By using the rank condition of the rigidity matrix, we define symmetric weighted matrix and maintenance of the smallest positive eigenvalue of interest. An distributed way for estimating the eigenvector is proposed, which use the inverse power method for achieving adjustable and fast convergence speed. We further compare the method with the commonly used power method. Furthermore, the influence of the error from the estimation is analyzed, which shows that bounded error will not affect the stability. Additional bounded control input will not affect the rigidity preservation scheme. Results are evaluated by simulations.

3: %\end{abstract}

4: