《基于经典MUSIC的DOA估计Matlab仿真》由会员分享,可在线阅读,更多相关《基于经典MUSIC的DOA估计Matlab仿真(12页珍藏版)》请在金锄头文库上搜索。
1、基于经典MU S IC的DO A估计Ma t 1 a b仿真作者:日期:信息与通信工程学院阵列信号处理实验报告(基于经典MUSIC的DO A估计Matl ab仿真)学 号:XXXXXX专业:XXXXX X学生姓名:X XX任课教师:XXX201 5年X月?题目:基于经典MUSIC的DO A估计 Ma t lab仿真1 .算法简述:基于天线阵列协方差矩阵的特征分解类 DOA估计算法中,多重信号分类(M USIC)算法具有普遍适用性,只要已知天线阵的分布形式,无论直线阵还是圆阵, 不管阵元是否等间隔分布,都可以得到高分辨率的估计结果。阵列协方差矩阵R 可以划分为两个空间,即R UsUH UnnUN
2、。因导向矩阵的各矢量与噪声子 空间正交,可得到阵列 1 .空间谱函数Pmusic ( ) H使 变化,按照Pmusic()来搜索峰值来达aH( )UNUNa()到波达方向的估计值O2 .篌验内容与结果实验使用8阵元均匀线阵,阵元间距为信号波长的一半,输入信号为1个BP SK信号,1个非相干的单频干扰,设置载波频率 10 MHz、采样频率3 0MHz、 快拍数30、信干比0dB、信号方位角0;:、干扰方位角5,分析信噪比从1到20 dB估计均方误差。实验结果见下图。图1信噪比15dB情况下的波束图差误方均0。角测量x 100.5014161820x 100.5差误方均1012SNR(dB)5。角
3、测量2468101214161820SNR(dB)0图2不同信噪比测得的均方误差3 .仿真分析?由仿真结果可知,在实当前提条件下,经典 music算法的DO A估计可有效估计 出信号的来向。当信噪比较小时,估计得均方误差较大;随着信噪比的提高,估计 均方误差逐渐减小;当信噪比大于10dB以后,角度估计成功率100%。MUSI C算法在DOA估计中属于高精度算法,而算法的估计精度仍受阵元数量、信号角 度、信号数量、信号间距、信噪比等影响,本实验仿真分析了一定情况下估计精 度与信噪比的关系,得到了预期的实验结果。4 .程序c 1 ear all;cl o se al 1 ;c 1 c;% 线阵mu
4、sic DOA%参数设置5 ig n a l _ No =1 ;1 n te r f e re nce_N o = 1;S_No = Sig n a lN o + In t e r fer e nce N o;sensojNo = 8;azimuth = 0* p i /180 5*pi/180;Fs = 30 e6 ;Fc = 10e6;F1 = 8e6;t h eta = 10 45 52 93 4 3 *pi/18 0 ;RB = 2e6 ;M = 2;? ? ?% 二进制wavelength = 3e8/F c;d =wave lengt h/2;K = 300;? ? ?% 快拍数%
5、dBPs = -30;SNR = 10;SI R = 00;M_N o = 40;等于信息速率Data _No = M_No * F s /RB;t = 1/Fs: 1/Fs:Da t a_No/Fs;% for xunhuan=1: 1% mean1 = 0;% m ean2 = 0;% for SNR =1:20%干扰功率%码速率%功率转换Ps_l = 1 0A(P s /I 0 );Pi_l = 1 0 a (Ps-SIR) /10);Pn_l = 1 0 a (Ps-SNR)/ 1 0);% - - - - - - %信号生成b 1 t = ran d i nt ( 1 ,M_No);
6、?%产生信息序列bit s tre a m = 口;for i =1 :M_Noif b it (i) = 1b i tstream = b i t s t r eam, o n es (1,Fs/R B);elsebit s t r eam = bitstream, -o n es(1, F s/RB); endendCarri er _R = c o s(2 *pi*Fc* t);S_R = C ar r ier_R .* bit stream;CarrierI = sin(2*pi*Fc* t);S _ I = Car r i er_I . * bits t rea m;S ig n a
7、l_R = sq r t(P s _ 1 ) * S _R/sq rt ( (S_R* S _R/length(S_R);S i gna l_I = sqrt(Ps_l) * S_I /sq r t(S_I*S_I/length(S_I) ); %生成 BP S K复信号S ignal (1,:) = complex (Si g n al_R, S i gnal_I);%产生干扰%I_R = sqrt (2* Pi_l) * c o s (2 *pi*F 1 * t ); %I_I = sqrt (2 * P i _l)*s i n(2 * p i * F1*t);for i = 2:Inte
8、rf erence _No + 1S i gnal(i ,:) = sqrt(2* P i l) *c o mple x (sin(2*pi* F 1( i -1)*t+ t heta(i), sin ( 2 * p i*F1(i-1 ) *t+th e ta(i);end% - - - - -一 %模拟天线接收As = zeros (sensor_No, S_No);for i = 1:s e nsor_Nofo r ii = 1:S_NoAs (i, i i) = ex p(-1i* 2 *pi*sin(azimuth( i i ) )*d/wav e len g t h * (i- 1
9、);e ndendx = As* Signal;%加噪声n o i s e_R = ra ndn(sens or _ No,Data_No) *sqrt(P n l);n o ise_I = ran d n( s ensor_No,Data_No)*sqrt (Pn_l);noi s e = compl e x ( n oise_R ,n o is e I );s = x + n o ise;x1 = s(:,(l :K);R = x1*x1/K;V, D =eig(R);D 1 = d iag (D);Un=V (:, 1: s e nsor_No-S_No);Gn=U n *U n ;se
10、 arc hing_doa= - 9 0: 0. 1: 90;?% 线阵的搜索范围为-909 0 度for i=1:lengt h (s earching_doa)a_the ta=exp(-j*(0:se n sor_No-1)*2*pi*d*sin(p i *s e arching_doa(i)/180)/ wav ele ng th);Pmusic ( i ) = 1 ./abs(a_thet a)*Gn*a_ t heta); en dp l o t( s e arch i ng_ do a ,10* 1 o g(Pmusic),r );x l a bel(入射角 /degree );
11、y label (空间谱/d B);lege nd (Music Spectrum/);11 tle(经典 MUS I C估计);grid on;%查最大值% P m ax _db 1 =0;% for i=1:909% if Pmu s ic (i)Pmax _ db1%Pmax_d b1 = Pmusic ( i );%P max_p i t c h i n g1 = i;% e nd% end% Pmax_db2=0;% for i = 91 0 :1 8 01% i f Pmusic (i) Pmax _d b2%Pm a x_ d b 2 = P music( i );%Pm ax_
12、pit c hin g 2=i;% end% e nd% e rr1 (SNR) = (Pmax_p i tch i ng 1 -1)/10- 9 0 )A 2 ;% err2 (SNR) = (Pmax_pi t chi n g2-1 ) / 10-90 - 5 )人 2;% end% mean1 = me an 1 + err1;% mean 2 = mean2 + e r r 2;% end% mse1 = m e an1/100O;% mse2 = mean2/10 00;% su b plot (211) ; p 1 o t (m s e 1 );x 1 ab el (SNR( d B)% yl a be 1 (均方误差)% su b p 1 ot(21 2 ); plot(mse2);xl a be V SNR(dB)% ylabel (/均方误差)% f i g ure(2);% polar (-pi/2:pi/1800:pi/2 ),1 0 *l o g(Pm usi c )+33) ; f i g u r e(gcf );