-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmtg_functions.h
387 lines (307 loc) · 13.5 KB
/
mtg_functions.h
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
#ifndef MTG_FUNCTIONS
#define MTG_FUNCTIONS 1
/*This file contains all the functions used to calculated the time optimal gradient waveforms.
minTimeGradientRIV - Computes the rotationally invariant solution
RungeKutte_riv - Used to solve the ODE using RK4
beta - calculates sqrt (gamma^2 * smax^2 - k^2 * st^4) in the ODE. Used in RungeKutte_riv
minTimeGradientRV - Computes the rotationally variant solution
RungeKutte_rv - Used to solve the ODE using RK4
sdotdot - calculates the maximum possible value for d^2s/dt^t, used in RugeKutte_rv */
#include "spline.h"
#include "float.h"
#include "math.h"
inline double beta(double k, double st, double smax, double gamma = 4.25756) {
/* calculates sqrt (gamma^2 * smax^2 - k^2 * st^4) used in RK4 method for rotationally invariant ODE solver */
return sqrt(sqrt((gamma*gamma*smax*smax - k*k*st*st*st*st)*(gamma*gamma*smax*smax - k*k*st*st*st*st)));
}
inline double RungeKutte_riv(double ds, double st, double k[], double smax, double gamma = 4.25756) {
/* Solves ODE for rotationally invariant solution using Runge-Kutte method*/
double k1 = ds * (1/st) * beta(k[0], st, smax, gamma);
double k2 = ds * 1 / (st + k1/2) * beta(k[1], st + k1/2, smax, gamma);
double k3 = ds * 1 / (st + k2/2) * beta(k[1], st + k2/2, smax, gamma);
double k4 = ds * 1 / (st + k3/2) * beta(k[2], st + k3/2, smax, gamma);
return (k1/6 + k2/3 + k3/3 + k4/6);
}
inline void minTimeGradientRIV(double* x, double* y, double* z, int Lp, double g0, double gfin, double gmax, double smax,
double dt, double*& gx, double*& gy, double*& gz,
int &l_t, double ds = -1, double gamma = 4.25756) {
/*Finds the time optimal gradient waveforms for the rotationally invariant constraints case.
x - The input curve in x
y - , y
z - , and z dir.
Lp - row dimension of x,y,z
g0 - Initial gradient amplitude.
gfin - Gradient value at the end of the trajectory.
If given value is not possible
the result would be the largest possible amplitude.
gmax - Maximum gradient [G/cm] (4 default)
smax - Maximum slew [G/cm/ms] (15 default)
dt - Sampling time intervale [ms] (4e-3 default)
gx, gy, gz - pointers to gradient waveforms to be returned
kx, ky, kz - pointers to k-space trajectory after interpolation
l_t - Dimension of interpolated k-space trajectory (kx, ky, kz) needed for creating mex return arrays.
*/
int i = 0;
/* iflag used in spline method to signal error */
int iflag;
int last;
double *p = new double[Lp];
/* Representing the curve with parameter p */
for (i = 0; i < Lp; ++i)
p[i] = i;
/* Interpolation of curve for gradient accuracy, using cubic spline interpolation */
/* arrays used by spline function to store coefficients. */
double *c1x = new double[Lp];
double *c2x = new double[Lp];
double *c3x = new double[Lp];
double *c1y = new double[Lp];
double *c2y = new double[Lp];
double *c3y = new double[Lp];
double *c1z = new double[Lp];
double *c2z = new double[Lp];
double *c3z = new double[Lp];
spline(Lp, 0, 0, 1, 1, p, x, c1x, c2x, c3x, &iflag);
spline(Lp, 0, 0, 1, 1, p, y, c1y, c2y, c3y, &iflag);
spline(Lp, 0, 0, 1, 1, p, z, c1z, c2z, c3z, &iflag);
double dp = 0.1;
int num_evals = (int) floor((Lp-1) / dp)+ 1;
double toeval = 0;
/* converting to arc-length parameterization from p, using trapezoidal integration */
double *s_of_p = new double[num_evals];
s_of_p[0] = 0;
double *sop_num = new double[num_evals];
double sofar = 0;
double Cp_abs_pre = 0.;
for (i = 0; i < num_evals; ++i) {
toeval = (double) i * dp;
double Cpx = deriv(Lp, toeval, p, c1x, c2x, c3x, &last);
double Cpy = deriv(Lp, toeval, p, c1y, c2y, c3y, &last);
double Cpz = deriv(Lp, toeval, p, c1z, c2z, c3z, &last);
/* interpolated curve in p-parameterization */
double Cp_abs = sqrt(Cpx*Cpx + Cpy*Cpy + Cpz*Cpz);
sofar += (Cp_abs + Cp_abs_pre)/2;
s_of_p[i] = dp * sofar;
Cp_abs_pre = Cp_abs;
sop_num[i] = i * dp;
}
/* length of the curve */
double L = s_of_p[num_evals-1];
/* decide ds and compute st for the first point */
double stt0 = gamma*smax; /* always assumes first point is max slew */
double st0 = (stt0*dt)/2; /* start at half the gradient for accuracy close to g=0 */
double s0 = st0*dt;
if (ds < 0) { /* if a ds value was not specified */
// ds = s0/1.5; /* smaller step size for numerical accuracy */
ds = fabs(ds) * s0/1.5; /* pehses: to make it easier to scale ds */
}
int length_of_s = (int) floor(L/ds);
int half_ls = (int) floor(L/(ds/2));
double *s = new double[length_of_s];
double *sta = new double[length_of_s];
double *stb = new double[length_of_s];
for (i = 0; i<length_of_s; i++) {
s[i] = i*ds;
sta[i] = 0;
stb[i] = 0;
}
/* Convert from s(p) to p(s) and interpolate for accuracy */
double *a1x = new double[num_evals];
double *a2x = new double[num_evals];
double *a3x = new double[num_evals];
spline(num_evals, 0, 0, 1, 1, s_of_p, sop_num, a1x, a2x, a3x, &iflag);
double *s_half = new double[half_ls];
double *p_of_s_half = new double[half_ls];
for (i=0; i < half_ls; ++i) {
s_half[i] = (double)i*(ds/2);
p_of_s_half[i] = seval(num_evals, s_half[i], s_of_p, sop_num, a1x, a2x, a3x, &last);
}
delete[] a1x; delete[] a2x; delete[] a3x;
delete[] s_of_p; delete[] sop_num;
/* Csp is C(s(p)) = [Cx(p(s)) Cy(p(s)) Cz(p(s))] */
double *Cspx = new double[length_of_s];
double *Cspy = new double[length_of_s];
double *Cspz = new double[length_of_s];
double *p_of_s = new double[length_of_s];
for (i=0; i<length_of_s; ++i) {
p_of_s[i] = p_of_s_half[2*i];
Cspx[i] = seval(Lp, p_of_s[i], p, x, c1x, c2x, c3x, &last);
Cspy[i] = seval(Lp, p_of_s[i], p, y, c1y, c2y, c3y, &last);
Cspz[i] = seval(Lp, p_of_s[i], p, z, c1z, c2z, c3z, &last);
}
delete[] p_of_s_half;
/* arrays used by spline function to store coefficients. */
double *Csp1x = new double[length_of_s];
double *Csp2x = new double[length_of_s];
double *Csp3x = new double[length_of_s];
double *Csp1y = new double[length_of_s];
double *Csp2y = new double[length_of_s];
double *Csp3y = new double[length_of_s];
double *Csp1z = new double[length_of_s];
double *Csp2z = new double[length_of_s];
double *Csp3z = new double[length_of_s];
spline(length_of_s, 0, 0, 1, 1, s, Cspx, Csp1x, Csp2x, Csp3x, &iflag);
spline(length_of_s, 0, 0, 1, 1, s, Cspy, Csp1y, Csp2y, Csp3y, &iflag);
spline(length_of_s, 0, 0, 1, 1, s, Cspz, Csp1z, Csp2z, Csp3z, &iflag);
int size_k = half_ls+2; /* extend of k for RK4 */
double *k = new double[size_k]; /* k is the curvature along the curve */
for (i=0; i < half_ls; ++i) {
double kx = deriv2(length_of_s, s_half[i], s, Csp1x, Csp2x, Csp3x, &last);
double ky = deriv2(length_of_s, s_half[i], s, Csp1y, Csp2y, Csp3y, &last);
double kz = deriv2(length_of_s, s_half[i], s, Csp1z, Csp2z, Csp3z, &last);
k[i] = sqrt(kx*kx + ky*ky + kz*kz); /* the curvature, magnitude of the second derivative of the curve in arc-length parameterization */
}
k[size_k-2] = k[size_k-3];
k[size_k-1] = k[size_k-3];
delete[] s_half;
delete[] Cspx; delete[] Cspy; delete[] Cspz;
delete[] Csp1x; delete[] Csp1y; delete[] Csp1z;
delete[] Csp2x; delete[] Csp2y; delete[] Csp2z;
delete[] Csp3x; delete[] Csp3y; delete[] Csp3z;
/* computing geomtry dependent constraints (forbidden line curve) */
double *sdot = new double[half_ls];
/* Calculating the upper bound for the time parametrization */
/* sdot (which is a non scaled max gradient constaint) as a function of s. */
/* sdot is the minimum of gamma*gmax and sqrt(gamma*gmax / k) */
double gammagmax = gamma*gmax;
for (i=0; i< half_ls; ++i) {
double sdot2 = sqrt((gamma*smax) / (fabs(k[i]+(DBL_EPSILON))));
if (gammagmax < sdot2)
sdot[i] = gammagmax;
else
sdot[i] = sdot2;
}
double g0gamma = g0*gamma + st0;
if (g0gamma < gammagmax)
sta[0] = g0gamma;
else
sta[0] = gammagmax;
/* Solving ODE Forward */
for (i=1; i<length_of_s; ++i) {
double k_rk[3];
k_rk[0] = k[2*i-2];
k_rk[1] = k[2*i-1];
k_rk[2] = k[2*i];
double dstds = RungeKutte_riv(ds, sta[i-1], k_rk, smax, gamma);
double tmpst = sta[i-1] + dstds;
if (sdot[2*i+1] < tmpst)
sta[i] = sdot[2*i+1];
else
sta[i] = tmpst;
}
/*Solving ODE Backwards: */
double max;
if(gfin < 0 ) {
/*if gfin is not provided */
stb[length_of_s-1] = sta[length_of_s - 1];
} else {
if (gfin * gamma > st0)
max = gfin*gamma;
else
max = st0;
if (gammagmax < max)
stb[length_of_s-1] = gammagmax;
else
stb[length_of_s-1] = max;
}
for (i=length_of_s-2; i>-1; --i) {
double k_rk[3];
k_rk[0] = k[2*i+2];
k_rk[1] = k[2*i+1];
k_rk[2] = k[2*i];
double dstds = RungeKutte_riv(ds, stb[i+1], k_rk, smax, gamma);
double tmpst = stb[i+1] + dstds;
if (sdot[2*i] < tmpst)
stb[i] = sdot[2*i];
else
stb[i] = tmpst;
}
delete[] k;
delete[] sdot;
/*Final interpolation */
/* Converting to the time parameterization, t(s) using trapezoidal integration. t(s) = integral (1/st) ds */
double st_of_s;
if (sta[0] < stb[0])
st_of_s = sta[0];
else
st_of_s = stb[0];
double st_ds_i_pre = ds*(1/st_of_s);
double *t_of_s = new double[length_of_s];
t_of_s[0] = 0;
for (i=1; i < length_of_s; ++i) {
if (sta[i] < stb[i])
st_of_s = sta[i];
else
st_of_s = stb[i];
double st_ds_i = ds*(1/st_of_s); /* ds * 1/st(s) used in below calculation of t(s) */
t_of_s[i] = t_of_s[i-1] + (st_ds_i+ st_ds_i_pre)/2;
st_ds_i_pre = st_ds_i;
}
delete[] sta; delete[] stb;
l_t = (int) floor(t_of_s[length_of_s-1]/dt);
/* coefficient arrays for spline interpolation of t(s) to get s(t) */
double *t1x = new double[length_of_s];
double *t2x = new double[length_of_s];
double *t3x = new double[length_of_s];
spline(length_of_s, 0, 0, 1, 1, t_of_s, s, t1x, t2x, t3x, &iflag);
/* coefficient arrays for spline interpolation of p(s) with s(t) to get p(s(t)) = p(t) */
double *p1x = new double[length_of_s];
double *p2x = new double[length_of_s];
double *p3x = new double[length_of_s];
spline(length_of_s, 0, 0, 1, 1, s, p_of_s, p1x, p2x, p3x, &iflag);
double *p_of_t = new double[l_t];
for (i=0; i < l_t; ++i) {
double s_of_t = seval(length_of_s, i*dt, t_of_s, s, t1x, t2x, t3x, &last);
p_of_t[i] = seval(length_of_s, s_of_t, s, p_of_s, p1x, p2x, p3x, &last);
}
delete[] t1x; delete[] t2x; delete[] t3x;
delete[] t_of_s;
delete[] s; delete[] p_of_s;
delete[] p1x; delete[] p2x; delete[] p3x;
/* interpolated k-space trajectory */
double *Cx = new double[l_t];
double *Cy = new double[l_t];
double *Cz = new double[l_t];
for (i=0; i<l_t; i++) {
Cx[i] = seval(Lp, p_of_t[i], p, x, c1x, c2x, c3x, &last);
Cy[i] = seval(Lp, p_of_t[i], p, y, c1y, c2y, c3y, &last);
Cz[i] = seval(Lp, p_of_t[i], p, z, c1z, c2z, c3z, &last);
}
delete[] p_of_t; delete[] p;
delete[] c1x; delete[] c1y; delete[] c1z;
delete[] c2x; delete[] c2y; delete[] c2z;
delete[] c3x; delete[] c3y; delete[] c3z;
/* Final gradient waveforms to be returned */
gx = new double[l_t];
gy = new double[l_t];
gz = new double[l_t];
for (i=0; i< l_t - 1; ++i) {
gx[i] = (Cx[i+1] - Cx[i]) / (gamma * dt);
gy[i] = (Cy[i+1] - Cy[i]) / (gamma * dt);
gz[i] = (Cz[i+1] - Cz[i]) / (gamma * dt);
}
delete[] Cx; delete[] Cy; delete[] Cz;
gx[l_t-1] = 2*gx[l_t-2] - gx[l_t-3];
gy[l_t-1] = 2*gy[l_t-2] - gy[l_t-3];
gz[l_t-1] = 2*gz[l_t-2] - gz[l_t-3];
}
inline void calcTrajectory(double *gx, double *gy, double *gz, int l_t, double dt, double*& kx, double*& ky, double*& kz, double gamma = 4.25756) {
/* k-space trajectory to be returned (calculated by integrating gradient waveforms by trapezoidal integration) */
kx = new double[l_t];
ky = new double[l_t];
kz = new double[l_t];
double sofarx = 0;
double sofary = 0;
double sofarz = 0;
kx[0] = 0.;
ky[0] = 0.;
kz[0] = 0.;
for (int i=1; i < l_t; ++i) {
sofarx += (gx[i] + gx[i-1]) / 2;
sofary += (gy[i] + gy[i-1]) / 2;
sofarz += (gz[i] + gz[i-1]) / 2;
kx[i] = sofarx * dt * gamma;
ky[i] = sofary * dt * gamma;
kz[i] = sofarz * dt * gamma;
}
}
#endif