Research On Interactive Control Technology For Manned Vehicle-UAV Cooperative Mission

Along with the development of the technology associated with the Unmanned Aerial Vehicle (UAV) system, the applications are increasingly broad, the new combat pattern that manned vehicles accomplish the missions with UAVs will come true in the future war. The manned vehicle and the UAV take advantages of their own superiorities in collaboration with each other, which can achieve the better efficiency. It would powerfully ensure the success of the manned vehicle - UAV cooperative mission that the interaction system is designed effectively.The interaction system is designed for the manned vehicle - UAV cooperative missions based on methods in Natural Language Understanding in this thesis, the main work in the thesis includes:Firstly, the thesis analyzed the typical UAV missions in existence, the basic requests of the UAV mission control based on Ground Control System, the characteristics of the manned vehicle - UAV cooperative missions and the commands from human being, brought forward a concept model of the UAV mission control based on Natural Language Interaction, and presented the architecture of the Natural Language Command Understanding System.Considering the mistakes in the speech recognitions by the speech recognition system in existence, such as text disperse and the voice-character matching mistakes, the thesis brought forward a maximum matching segment algorithm aiming at recognition text. This algorithm effectively restrains the mistakes in the lexical layer. The thesis designed domain-based grammar and syntax tree structure, and understood the input command and output the relative parameters of mission in it, with bottom-to-up syntax analysis technology and case matching computing of domain-based syntax tree.Experiments on NLU module and NL-CUS were conducted, scaling their reliability and improved measure is given out. A series of experimental results show that the NL-CUS is effective.Considering the mistakes of the syllable confusion of the words, the thesis designed the syllable confusion of the words of commands, and improved the commands. A series of experimental results show that the new commands are better on speech recognition and understanding. For the sake of transforming the concept model of the UAV mission control based on Natural Language Interaction into a sequence of UAV tasks, based on the NN-based potential field path planning algorithm and rolling window-based path planning strategy, the thesis integrated the two algorithms togethe, and brought forward a new path-planning algorithm for UAV under the partially unknown static environment.