Song

Papers

Total Citations

H-Index

About

Song’s research focuses on mobile robotics, state estimation, and adaptive filtering, with a particular emphasis on improving the performance of autonomous systems in constrained environments. His most significant contribution is the development of an adaptive unscented Kalman filter (UKF) algorithm for simultaneous state and parameter estimation of mobile robots. By leveraging the mismatch between innovation sequences and their corresponding covariance matrices as a cost function, Song designed an online adaptation mechanism—rooted in the MIT rule—that dynamically updates process noise covariance. This innovation compensates for the lack of prior knowledge about noise distributions, enhancing both estimation accuracy and convergence speed. The work, published in 2008 and cited 25 times, demonstrated clear advantages over conventional UKF through simulations with an omnidirectional mobile robot, establishing a foundation for robust real-time robot control. Song has also explored the kinematic properties of wheeled mobile robots navigating round ducts and pipes, a niche area with implications for pipeline inspection robotics. His adaptive filtering approach remains a reference point for researchers tackling state estimation under uncertain conditions, bridging theoretical rigor with practical robotic applications.