Space Robot Parameter Estimation

Generating test cases

• The files present in test_case_data/{robot_make}/config_files have to pasted into ReDySim_Floating_Base/InvDyn
• From ReDySim_Floating_Base/InvDyn , execute run_me.m('/robot_make')
  The data files are generated and also copied to test_case_data/robot_make/sim_data
• From test_case_data/ , execute simdata_to_realdata('/robot_make', [x_coord, y_coord, z_coord].')
  The argument is the sensor coordinates seen in the base CoM frame (ReDySim frame)
  The transformed data is stored in test_case_data/{robot_make}/sim_real_data
  
• From src/ , execute run_me
  Before executing, make sure that the robot_make and the sensor coordinates are same as the ones used for simdata_to_realdata

.dat files indexing

Note: All the data is in SI units
In files, envar.dat, mtvar.dat, and statevar.dat, data in each row is observed at the time mentioned in the corresponding row in timevar.dat

• mtvar.dat: Contains momentum data
  C1, C2, C3: Linear momentum along x,y, and z axes respectively
  C4, C5, C6: Angular momentum along x,y, and z axes respectively
  

• statvar.dat: Contains the kinematic data of the base and the joints
  For n joints on the arms of the robot, then
  C1, C2, C3: x, y, and z positions of the base
  C4, C5, C6: Euler angles of the base (ZXY euler angle convention)
  C7 to C(6+n): Joint position
  C(7+n), C(8+n), C(9+n): x, y, and z velocities of the base
  C(10+n), C(11+n), C(12+n): Rates of euler angles of the base
  C(13+n) to C(13+2n-1): Joint velocities
  C(13+2n): Total energy of the system
  

Add common functions path to matlab search path

In the Matlab command window, find out the matlab root folder
matlabroot
In the terminal, move to the folder which has matlabrc.m file
cd matlabroot/toolbox/local
Allow user to edit matlabrc.m
Add the following command to the last line of the file
addpath(location_to_common_funtions_directory)

Generate case wise regressor matrices

Generate the configuration files for every case in the designated location
execute case_wise_reg_mat with the relevant inputs

Performance evaluation

Obain SPV of estimated MPV

Obtain the double form of PCM from ./symbolic_derivations/generate_param_coupling_mat
Save the PCM to param_coupling_mat.mat in the location, ./src/performance_evaluation
Save true MPV, estimated MPV, and true SPV to true_estimated_params.mat in the location, ./src/performance_evaluation
Run compute_est_spv.m

Generating validation joint trajectories

Set the config parameters in make_rand_tr_params.m in ./ReDySim_Floating_base/InvDyn_fourier and run it

Obtain state and torque data with true and estimated SPV

Set t_final and incr in ./ReDySim_Floating_base/InvDyn_fourier/initials.m as per the requirement

With true SPV

Copy the inputs file which constitute the true SPV to ./ReDySim_Floating_base/InvDyn_fourier
Goto to Getting the data files

With estimated SPV

Copy the inputs.m and input_param_vec.mat file from src/performance_evaluation to ./ReDySim_Floating_base/InvDyn_fourier
Goto to Getting the data files

Getting the data files

After copying the appropriate inputs file to the InvDyn folder, the following steps are to be followed for both true and estimaed SPV
Execute run_me('/dual_arm_articulate_validation')
Copy tor.dat and timevar.dat to ./src/performance_evaluation/robot_make
Set base_sensor_base_com_position_base_com to the value used for estimation
simdata_to_realdata('/dual_arm_articulate_validation', base_sensor_base_com_position_base_com, 0)
Copy ./test_case_data/dual_arm_articulate_validation/sim_real_data/statevar.dat to ./src/performance_evaluation/robot_make
For .dat files generated with true parameters and estimated, rename them to *_1.dat and *_2.dat respectively