Clinial example 2: Heart biopsy
Optimization problem
In this example, the user can follow the instructions below to run the full optimization process through: path, sequential, and simultaneous optimization step.
The run files for each step are as follows. The user-defined parameter are the anatomical model (.ply format), starting point, target point and the number of control points, path points. A B-spline curve will be optimized by the function below.
import numpy as np
import scipy
from ctr_framework.design_method.path_opt import path_opt
# Initialize the number of control points and path points
num_cp = 25
num_pt = 100
# User-defined start point and target point
sp = np.array([-23,-8,-85])
fp = np.array([87,-27,-193])
# mesh .PLY file
filename = 'heart.ply'
path_opt(num_cp,num_pt,sp,fp,filename)
Once the collision-free path is found, the sequental optimization needs to be performed in order to get a better initial guesses for the simultaneous CTR optimization problem. The codes are as follow
import numpy as np
import scipy
from ctr_framework.design_method.seq_opt import seq_opt
#########################################
############## initialization ###########
#########################################
# number of waypoints
viapts_nbr=10
# number of links
num_nodes = 50
# Extract the waypoints from optimized path
pt = initialize_pt(viapts_nbr)
pt_pri = initialize_pt(viapts_nbr * 2)
pt_full = initialize_pt(100)
# initial robot configuration
# Tube 1(inner tube) ID, OD
d1 = 0.65
d2 = 0.88
# Tube 2
d3 = 1.076
d4 = 1.296
# Tube 3(outer tube)
d5 = 1.470
d6 = 2.180
# Tube curvature (kappa)
kappa_init = np.array([0.0061, 0.0131,0.0021]).reshape((1,3))
# The length of tubes
tube_length_init = np.array([200, 120,65]).reshape((1,3)) + 100
# The length of straight section of tubes
tube_straight_init = np.array([150, 80,35]).reshape((1,3)) + 50
# joint variables
alpha_init = np.zeros((k,3))
alpha_init[:,0] = -np.pi/2
alpha_init[:,1] = np.pi/1.5
alpha_init[:,2] = -np.pi/3
beta_init = np.zeros((k,3))
beta_init[:,0] = -280
beta_init[:,1] = -205
beta_init[:,2] = -155
# initial torsion
init_dpsi = np.random.random((k,3)) *0.01
rotx_ = 1e-10
roty_ = 1e-10
rotz_ = 1e-10
loc = np.ones((3,1)) * 1e-5
mdict = {'alpha':alpha_init, 'beta':beta_init,'kappa':kappa_init,
'tube_section_straight':tube_straight_init,'tube_section_length':tube_length_init,
'd1':d1, 'd2':d2, 'd3':d3, 'd4':d4, 'd5':d5, 'd6':d6, 'initial_condition_dpsi':init_dpsi,
'rotx':rotx_,'roty':roty_ ,'rotz':rotz_ , 'loc':loc,
}
scipy.io.savemat('initial.mat',mdict)
# Base frame
base = np.array([-23,-8,-57]).reshape((3,1))
rot = np.array([3.14,0,0]).reshape((3,1))
# mesh .PLY file
meshfile = 'heart.ply'
pathfile = 'path.mat'
seq_opt(num_nodes,viapts_nbr,base,rot,meshfile,pathfile)
The second step serves as an initial guesses for the final step, which is the patient-speific simultaneous optimization. In this step, the optimizer optimizes k robot configurations simultaneously to obtain a robot design and safe motion plan.
import numpy as np
from ctr_framework.design_method.sim_opt import sim_opt
# Initialize the number of number of links and waypoints
num_nodes = 50
k = 10
# robot initial pose
base = np.array([-23,-8,-57]).reshape((3,1))
rot = np.array([3.14,0,0]).reshape((3,1))
# mesh .PLY file
meshfile = 'heart.ply'
pathfile = 'path.mat'
# run simultaneous optimization
sim_opt(num_nodes,k,base,rot,meshfile,pathfile)