Research And Application Of Key Technology For Rotor UAVs Obstacle Avoidance Based On The Performance

The obstacle to unmanned aerial vehicles(UAVs)avoidance is becoming apparent with the increasing use of UAV,and its flight activities in urban and mountainous regions are increasing rapidly.On the one hand,UAV manufacturers can improve their obstacle avoidance ability by improving the performance of UAVs,especially the dynamic obstacle avoidance ability.UAV operators,on the other hand,can realize the static obstacle avoidance through the test of UAVs performance evaluation,building scientific collision model and carrying out reasonable obstacle avoidance path plan.It is the undoubtedly effective way to avoid the obstacle that dynamic obstacle avoidance technology based on the static path planning.The main problems in the obstacle avoidance path planning for UAVs are discussed in this thesis in the case of the commonly used four-rotor UAV as a typical model,and the campus environment as the flight environment.Firstly,based on the basic model of the four-motors UAV,the program of UAV performance test is designed,and rich flight record data is obtained although repeated test flight.Secondly,the flight path obtained from the test flight is compared with the set path,and the navigation accuracy error data is obtained,and its processing and analysis are carried out.Through the different distribution of hypothesis test,the error data,and the nonparametric models such as polynomial,gaussian,exponential fitting,the navigation precision of error distribution and probability density function types are got,and the vertical and horizontal precision error probability distribution function and the collision model are set up.Thirdly,the flight digital environment is established by using the grid method to construct the several digital environment modeling methods including the digital elevation,raster method and geometric method.Fourthly,the probability collision model,the digital flight environment and A*algorithm are combined to obtain the path planning algorithm for UAVs obstacle avoidance based on performance constraint and obstacle threat probability.Finally,the obstacle avoidance path from the planning algorithm is verified and the flight data is analyzed.It is indicated that that route planning method is effective,reliable and good adaptability.It is founded in this thesis that the four-motors UAVs horizontal navigation accuracy and vertical navigation precision errors is in accordance with the second order exponential distribution and the second-order gaussian distribution respectively through the analysis of the test flight data.The correlation coefficient of R~2 value between corresponding probability density function fitting formula and the real data is greater than 0.99,which means the credibility is very high.The model of collision probability is reasonable and reliable.The analysis of the flight data shows that the algorithm based on this model is reasonable and effective.At the same time,the collision probability model is also useful for the study of multi-UAVs safe distance calculation and dynamic obstacle avoidance algorithms.