Abstract: According to the dynamic localization failure of the indoor mobile robot in network blind spots， a selfdynamic localization system was proposed which dynamically choose beacon node on the basis of WSN. The extended Kalman filter was applied to range management by RSSI. Then， maximum likelihood estimation was used to accomplish the localization. Finally， the localization errorcorrect was implemented by asynchronous Kalman filter. In order to correct and improve the localization accuracy， the classical Kalman filter with the other localization algorithms were integrated successfully using the proposed method， which could smooth and optimize the result of the algorithms. Especially in network blind spots， the asynchronous Kalman filter could provide optimal data. Simulation results showed that the accuracy， adaptability and robustness of the selfdynamic localization of mobile robot are good.

Key words: wireless sensor network(WSN), dynamic localization, improved maximum likelihood estimation, asynchronous Kalman filter algorithm, received signal strength index