As a key technology for sensing the physical world, embedded sensor networks attract wide research attention in recent years. Operating systems for sensor networks can provide a wide range of APIs for various applications, propelling future technological innovations. This work focuses on four key issues in traditional sensor network operating systems, i.e., user-friendly software programming, real-time task scheduling, efficient software update, and high bandwidth utilization. A novel operating system（SenSpire OS）is proposed to address these issues. SenSpire OS is based on a hybrid scheduling model and a real-time scheduling algorithm. Related works have been published in prestigious journals and conferences. Dr. Dong from College of Computer Science and Technology was nominated for the National Excellent Doctorial Dissertation Award.