Creation of Imu geometric model in Matlab In order to generate Synthetic data(Accelerometer and Gyroscope) 6DOF
12 次查看(过去 30 天)
显示 更早的评论
I am looking to create a geometric model(using the simulink blocks) of an IMU sensor that is available in Mathworks 'Sensor Fusion Toolbox'. I have to generate synthetic data of the 3-axis accelerometer,3-axis gyroscope(6DOF) in order to calculate the quaternions(to calculate the orientation) using an sensor fusion algorithm(Madgwick Filter or any similar one). I want to do this in Simulink without using real hardware sensors.I am stuck at this point how to go forward.
Looking for any suggestions or references of similar models.
Thanks in advance.
0 个评论
回答(2 个)
cui,xingxing
2024-10-10
@Revanth Kumar Adireddy Hi, I have given here the synthetic simulation data generated by imuSensor and then recovered the attitude data through the filter and simulated it.
%% 定义滤波器
SampleRate = 100;
fuse1 = imufilter(ReferenceFrame='NED',SampleRate=SampleRate,GyroscopeNoise=1e-4, GyroscopeDriftNoise=1e-6);
% fuse2 = ahrsfilter('SampleRate', SampleRate); % thiS filter must need magnetometer
fuse3 = complementaryFilter(ReferenceFrame='NED',HasMagnetometer=false,AccelerometerGain=0.01);
%% 定义人工生成的groundTruth
% 分别在2秒内绕x,y,z轴旋转360度,得到真值仿真数据
fs = SampleRate;
eachLoopNumSamples = fs*2;
totalNumSamples = 3*eachLoopNumSamples;
traj = kinematicTrajectory('SampleRate',fs);
accBody = zeros(totalNumSamples,3);
angVelBody = zeros(totalNumSamples,3);
angVelBody(1:eachLoopNumSamples,1) = (2*pi)/2;
angVelBody(eachLoopNumSamples+1:2*eachLoopNumSamples,2) = (2*pi)/2;
angVelBody(2*eachLoopNumSamples:end,3) = (2*pi)/2;
[~,orientationNED,~,accNED,angVelNED] = traj(accBody,angVelBody);
IMU = imuSensor('accel-gyro','SampleRate',fs);
% % 加点噪音模拟真实的IMU
% IMU.Accelerometer = accelparams(...
% MeasurementRange=20,...
% Resolution=0.00239,...
% TemperatureScaleFactor=0.008,...
% ConstantBias=0.1962,...
% TemperatureBias=0.0015,...
% NoiseDensity=0.1);
[accelReadings,gyroReadings] = IMU(accNED,angVelNED,orientationNED);
t = (0:(totalNumSamples-1))/SampleRate;
plot(t,accelReadings)
legend('X-axis','Y-axis','Z-axis')
ylabel('Acceleration (m/s^2)')
title('Accelerometer Readings')
%% 展示姿态解算恢复结果
fig = figure;
ax = axes(Parent=fig);
pp = poseplot(ax,"NED");
xlabel("North-x (m)")
ylabel("East-y (m)")
zlabel("Down-z (m)");
for idx = 1:size(accelReadings,1)
% Get the latest acceleration and angular velocity data
acc = accelReadings(idx,:);
gyr = gyroReadings(idx,:);
if ~isempty(acc) && ~isempty(gyr)
% acc = -[acc(2),acc(1),-acc(3)]; % if data from matlab mobile™ APP
% gyr = [gyr(2),gyr(1),gyr(3)];
% Use the imufilter to estimate orientation (quaternion)
orientation = fuse3(acc, gyr);
% Convert quaternion to rotation matrix
rotm = quat2rotm(orientation);
set(pp, "Orientation", orientation);
end
% Refresh the plot
% drawnow ;
pause(1/SampleRate);
end
0 个评论
Brian Fanous
2022-8-25
You can use waypointTrajectory and imuSensor to accomplish this in MATLAB. There is an equivalent Simulink block for imuSensor but you'd have to put the waypointTrajectory code in a MATLAB Function block to handle the trajectory generation portion:
% Simulink IMU data during a trajectory between two random orientations
wp = waypointTrajectory('Waypoints', zeros(2,3), 'Orientation', randrot(2,1), 'TimeOfArrival', [0 1].');
sampleRate = 10;
[~, orient, ~, acceleration, angvel] = lookupPose(wp, 0:(1/sampleRate):1);
imu = imuSensor('SampleRate', sampleRate);
[acc, gyro] = imu(acceleration, angvel, orient);
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Sensor Models 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!