Research Of The Multi-target Detection Algorithm For Triangular FMCW Radar

The emergence of radar not only enhances the military power,but also facilitates people's lives significantly.As an important equipment in the field of navigation,the shipborne navigation radar plays a crucial guiding role in the sailing at night,in foggy conditions,into or out of ports,and in the collision avoidance,etc.The increasing complex sea conditions require high performance of ship navigation radar,promoting the development of continuous wave based marine radar instead of pulse based one.The linear frequency modulated continuous wave radar has been widely used due to its high resolution,full-scale range and other characteristics,thus the research on the beat signal processing algorithm with respect to triangle frequency modulated continuous wave radar system is of great importance.On the basis,this thesis mainly focuses on the clutter suppression processing,target identification,and ship classification,which are intensively studied in navigation field.Specifically,a multi-target detection algorithm is designed in this thesis,which can accurately detect multiple targets with noise and clutter,extract the distance and velocity information of detected targets,and achieve high-performance target classification.Firstly,after analyzing the factors that affect the measurement accuracy of the beat signal frequency,the signal preprocessing is used to reduce the influence of clutter.On the basis of study upon the characteristics of sea clutter,a sea clutter model is established,and then a clutter suppression algorithm taking advantage of autocorrelation function energy based empirical mode decomposition is proposed,which can overcome the filter cutoff order error when the clutter modal energy is greater than that of the intrinsic mode function energy in the signal mode.The simulation results show that the proposed algorithm can effectively inhibit the clutter.Secondly,according to error estimation problem when the signal peak frequency is within two frequency resolution values after FFT,an adaptive frequency correction algorithm is proposed,which selects different correction algorithms according to different frequencies,thus it avoids the problem that the single calibration algorithm is only locally optimal.The simulation results verify the validity of the algorithm and show that the proposed algorithm has higher frequency estimation accuracy and anti-noise ability.Thirdly,after analyzing characteristics of the ship targets,width and speed are chosen as features,and then a marine characteristic database is established according to the common ships at sea.The back propagation neural network and decision tree are used to classify the ship targets and make comparative analysis.The experimental results show that both the two algorithms are effective for target classification,and the classification accuracy is higher than the traditional K-nearest neighbor algorithm.Finally,based on above discussion,a multi-target detection algorithm namely the entire processing flow of the beat signal is designed.According to the process of the algorithm,the simulation analysis is conducted,which shows the effectiveness of the proposed algorithm.