文件列表:

CPPFiles/ (0, 2014-05-09)
CPPFiles/InvertMatrix.hpp (1370, 2014-05-09)
CPPFiles/cholesky.hpp (7030, 2014-05-09)
CPPFiles/cppUKFData.txt (19786, 2014-05-09)
CPPFiles/fmUKF.cpp (10026, 2014-05-09)
CPPFiles/inputData.cpp (9415, 2014-05-09)
CPPFiles/outputDataUKF.cpp (848, 2014-05-09)
CPPFiles/quadStateEst.cpp (6308, 2014-05-09)
Data/ (0, 2014-05-09)
Data/poseNPPOrient.txt (41618, 2014-05-09)
Data/poseNPPPos.txt (41202, 2014-05-09)
Data/processInputAcc.txt (41860, 2014-05-09)
Data/processInputOmega.txt (42196, 2014-05-09)
Data/quadData.mat (63368178, 2014-05-09)
Data/viconEuler.txt (127074, 2014-05-09)
Data/viconPos.txt (128436, 2014-05-09)
Data/viconTime.txt (43302, 2014-05-09)
MatlabFiles/ (0, 2014-05-09)
MatlabFiles/MatlabEquations/ (0, 2014-05-09)
MatlabFiles/MatlabEquations/14GEqn.mat (1261, 2014-05-09)
MatlabFiles/MatlabEquations/14GFunc.mat (11467, 2014-05-09)
MatlabFiles/MatlabEquations/14LFunc.mat (11203, 2014-05-09)
MatlabFiles/MatlabEquations/imuGEqn.mat (2434, 2014-05-09)
MatlabFiles/MatlabEquations/imuGFunc.mat (10782, 2014-05-09)
MatlabFiles/MatlabEquations/imuLFunc.mat (11157, 2014-05-09)
MatlabFiles/MatlabSource/ (0, 2014-05-09)
MatlabFiles/MatlabSource/RPYToRotMat.m (448, 2014-05-09)
MatlabFiles/MatlabSource/RotMatToRPY.m (205, 2014-05-09)
MatlabFiles/MatlabSource/atan3.m (123, 2014-05-09)
MatlabFiles/MatlabSource/calculateCovEquation.m (1572, 2014-05-09)
MatlabFiles/MatlabSource/final650.m (19055, 2014-05-09)
MatlabFiles/MatlabSource/imuUKF.m (14359, 2014-05-09)
MatlabFiles/MatlabSource/nPointPose.m (5146, 2014-05-09)
MatlabFiles/MatlabSource/outputSensorLog.m (5043, 2014-05-09)
MatlabFiles/MatlabSource/plotUKFcppData.m (586, 2014-05-09)
MatlabFiles/MatlabSource/quadUKF.m (4220, 2014-05-09)
MatlabFiles/MatlabSource/symbolicFuncs.m (8171, 2014-05-09)
presentationKF.pdf (4996236, 2014-05-09)
ukfROS/ (0, 2014-05-09)
... ...

##Kalman Filter Port ###Porting of Matlab Based Code for the State Estimation of a Quadrotor into C++/ROS framework. (UKF/EKF). ####State Vector for both Kalman Filter implementations is 14 Dimensional: ####[position, velocity, orientation, imu accelerometer bias, roll/pitch bias] ####Implementation uses Boost 1.49, C++11/STL, and ROS Hyrdo ######Information The Kalman Filter is an optimal estimator. If the noise of the system and observations can be modeled as Gaussian, then the Kalman Filter minimizes the mean square error of the estimate. In addition, the filter is recursive and can thus provide state estimates as new data becomes available. If you have a good estimate, then combining the filter with a pre-process step of gain learning can achieve a reliable system. The purpose of the project was to fly a quadrotor using either the Extended Kalman Filter or the Unscented Kalman Filter with an IMU and single camera serving as input to the system. Once a good state estimator was developed, this would be used in conjunction with a PD controller which would use the position and velocity estimates to calculate the required thrusts and moments to reach the desired position. The nanoplus quadrotor has an onboard attitude controller that runs at a higher frequency than the position and velocity controller. This means the orientation estimate would only be kept as an exercise of the kalman filter since the attitude controller requires a larger rate of input than is available (The IMU and Camera data were synchronized at 33Hz). Having obtained a good controller and good state estimate, the quadrotor could track a desired trajectory. If you imagine a situation where the quadrotor would be required to follow a desired trajectory as accurately as possible (say, surveillance duty) then it would be necessary to learn an optimal set of gains. Here, only the Kalman Filters are provided. ######Files [.cpp Source] (/CPPFiles) [Matlab Source] (/MatlabFiles/MatlabSource) [CPP Unscented Kalman Filter] (/CPPFiles/fmUKF.cpp) [Matlab Unscented Kalman Filter] (/MatlabFiles/MatlabSource/quadUKF.m) [CPP UKF Loop] (/CPPFiles/quadStateEst.cpp) [Matlab UKF Loop] (/MatlabFiles/MatlabSource/imuUKF.m) [Matlab EKF] (/MatlabFiles/MatlabSource/final650.m) [Matlab N-Point Pose Algorithm] (/MatlabFiles/MatlabSource/nPointPose.m) [Symbolic Functions] (/MatlabFiles/MatlabSource/symbolicFuncs.m) [Ouput Sensor Log Script] (/MatlabFiles/MatlabSource/outputSensorLog.m) [Read into STL Vectors] (/CPPFiles/inputData.cpp) [Boost LU Factorization Matrix Inversion] (/CPPFiles/InvertMatrix.hpp) [Boost Cholesky Decomposition] (/CPPFiles/cholesky.cpp) [Data Collected Wirelessly From Quadrotor] (/Data) ######Using RViz [ROS Variant] (/ukfROS) Open a terminal and run 'roscore'. In a second terminal, open rviz with 'rosrun rviz rviz &' and change the 'Fixed Frame' to '/quad_World' under 'Global Options'. Then add a marker display type, which should be subscribed to the 'visualization_marker' topic. Finally, in a third terminal, execute ./quadStateEstROS and you should see the path of the quadrotor in Rviz. The full command to compile and build the source is provided here for reference: g++ -Wall -c quadStateEstROS.cpp -o quadStateEstROS -std=c++11 -I/opt/ros/hydro/include -L/opt/ros/hydro/lib /opt/ros/hydro/lib/libroscpp_serialization.so -Wl,-rpath,/opt/ros/hydro/lib -lroscpp -lrosconsole -lrostime

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：