Research on UAV ad hoc network system design

Min Zhu, Yong-Jian Luo, Ming-shun Xing

Abstract: According to the special application environment and flexible and dynamic networking requirements of UAV network, a UAV network architecture design based on ad hoc network is proposed. This architecture provides a fully connected networking mode for UAV network, including the design of key technologies such as architecture, networking technology, routing protocol, etc., so as to meet the special application environment of UAV self-organizing network node moving fast and topological structure changing frequently, with high reliability and operability.

: Download (free)  CSE-PDF

Rui Xue, Shixian Zeng, Jiangbo Chen

Abstract: The ageing of the population and high-risk diseases associated with age have contributed to the focus on rehabilitation and exercise assist devices. In order to provide perfect control of the wearable knee orthosis at the appropriate time, proper Gait Phase Recognition (GPR) is necessary to reduce the incidence of knee collapse and fall events.In the study, the gait data of knee flexion angle, thigh segment angular velocity and acceleration were trained and tested by Logistic Model Decision Tree (LMDT) classifier.The Transition Sequence Verification and Correction (TSVC) algorithm is applied to correct the continuous class prediction model and improve the GPR performance.Finally, the sensitivity, specificity, accuracy and other correlation coefficients of the system are verified. Compared with other methods, the results show that the LMDT-GPR system can be used for real-time orthosis control with high feasibility.Combined with the TSVC algorithm, the complexity of the sensor system can be reduced. It is concluded that the gait phase identification system is designed to provide more convenience for users.

: Download (free)  CSE-PDF