TDOA_室内定位_matlab算法_chan_taylor_卡尔曼滤波_NLOS因素.rar

function X=drop(noi,noise,alp) rmse=0; e=[0;0;0;0]; threshold =6; %数据是否有用的阈值 %target trajectory :circle,(x,y) %theta = 0:0.1:2*pi+0.2; %N=length(theta); %x = 0.4*sin(theta)+0.5; %y = 0.4*cos(theta)+0.5; %plot(x,y,'-r');hold on; theta=0:4:100; N=length(theta); x=theta; y=0*theta+20; plot(x,y,'-r');hold on; %linearized mode A = [1 1 0 0;... 0 1 0 0;... 0 0 1 1;... 0 0 0 1]; C = cell(1,N); %observation data for k = 1 : N x(k) = x(k)+0*randn(1);%add system noise ,u(k) y(k) = y(k)+0*randn(1); z1(k) = sqrt(x(k)^2+y(k)^2); z2(k) = sqrt((x(k)-100)^2+y(k)^2); z3(k) = sqrt(x(k)^2+(y(k)-100)^2); z4(k) = sqrt((x(k)-100)^2+(y(k)-100)^2); end z = cell(1,N); for k = 1 : N if k<=45/4||k>65/4 z{k} = [z1(k);z2(k);z3(k);z4(k)] +noise*randn(4,1);%[1;1;1;1];%add observation noise v(k) else z{k} = [z1(k);z2(k);z3(k);z4(k)] +noise*randn(4,1)+[0;0;noi+0.1*randn(1);0]; end end %estimated state:x_s x_s = cell(1,N); x_s{1} = [x(1);0;y(1);0]; P = cell(1,N); P{1} = eye(4); Q = 0.01*eye(4); R = 0.01*eye(4); x_sTime = cell(1,N); for k = 1 : N-1 %C{k} = [x_s{k}(1)/sqrt(x_s{k}(1)^2+x_s{k}(3)^2) 0 x_s{k}(3)/sqrt(x_s{k}(1)^2+x_s{k}(3)^2) 0;... % (2*x_s{k}(1)-2)/sqrt(2*((x_s{k}(1)-1)^2+x_s{k}(3)^2)) 0 x_s{k}(3)/sqrt((x_s{k}(1)-1)^2+x_s{k}(3)^2) 0;... % x_s{k}(1)/sqrt((x_s{k}(3)-1)^2+x_s{k}(1)^2) 0 (2*x_s{k}(3)-2)/sqrt(2*((x_s{k}(3)-1)^2+x_s{k}(1)^2)) 0]; C{k} = [ 1/(x_s{k}(1)^2+x_s{k}(3)^2)^(1/2)*x_s{k}(1),0, 1/(x_s{k}(1)^2+x_s{k}(3)^2)^(1/2)*x_s{k}(3),0;... 1/2/(x_s{k}(1)^2-200*x_s{k}(1)+10000+x_s{k}(3)^2)^(1/2)*(2*x_s{k}(1)-200), 0, 1/(x_s{k}(1)^2-200*x_s{k}(1)+10000+x_s{k}(3)^2)^(1/2)*x_s{k}(3),0;... 1/(x_s{k}(1)^2+x_s{k}(3)^2-200*x_s{k}(3)+10000)^(1/2)*x_s{k}(1), 0,1/2/(x_s{k}(1)^2+x_s{k}(3)^2-200*x_s{k}(3)+10000)^(1/2)*(2*x_s{k}(3)-200),0;... 1/2/(x_s{k}(1)^2-200*x_s{k}(1)+20000+x_s{k}(3)^2-200*x_s{k}(3))^(1/2)*(2*x_s{k}(1)-200),0, 1/2/(x_s{k}(1)^2-200*x_s{k}(1)+20000+x_s{k}(3)^2-200*x_s{k}(3))^(1/2)*(2*x_s{k}(3)-200),0]; K{k} = P{k}*C{k}'*inv(R+C{k}*P{k}*C{k}'); z_e = [sqrt(x_s{k}(1)^2+x_s{k}(3)^2);sqrt((x_s{k}(1)-100)^2+x_s{k}(3)^2);sqrt(x_s{k}(1)^2+(x_s{k}(3)-100)^2);sqrt((x_s{k}(1)-100)^2+(x_s{k}(3)-100)^2)]; ling=z{k}-z_e; %e=(e*(k-1)+abs(ling))/k; for m=1:4 if ling(m)>=3 K{k}(:,m)= 0; elseif 3>ling(m)&&ling(m)>=0 K{k}(:,m)= K{k}(:,m)/alp; else K{k}(:,m)= K{k}(:,m)*alp; end end %if norm(z{k}-z_e)>=threshold % threshold 是否丢弃的阈值，要在程序开始时给出；例如threshold=5; % K{k}=0; % else % K{k} = P{k}*C{k}'*inv(R+C{k}*P{k}*C{k}'); % end x_sTime{k}=x_s{k}+ K{k}*(z{k}-z_e);%measurement update x_s{k+1} = A*x_sTime{k};%time update P{k+1} = A*P{k}*A' + Q - A*P{k}*C{k}'*inv(C{k}*P{k}*C{k}'+R)*C{k}*P{k}*A'; % x_e(k) = x_sTime{k}(1); % y_e(k) = x_sTime{k}(3); % plot(x(k),y(k),'-^r');hold on; % plot(x_e(k),y_e(k),'*k');pause(0.1);hold on; end for k = 1 : N-1 x_e(k) = x_sTime{k}(1); y_e(k) = x_sTime{k}(3); rmse=rmse+sqrt((x_e(k)-x(k))^2+( y_e(k)-y(k))^2); end X=rmse/(N-1); %X=[x_e',y_e']; %plot(x_e,y_e,'-b'); %axis([0 100 0 100]); %xlabel('x'); %ylabel('y'); title('Tracking Performance'); %legend('actual orbit','estimated orbit') % for k = 41 : 60 % z{k} = [z1(k);z2(k);z3(k);z4(k)] +[1;10;1;1];%add observation noise v(k) %end %for k = 61 : N % z{k} = [z1(k);z2(k);z3(k);z4(k)] +[1;1;1;1];%add observation noise v(k) %end