| Line |
Branch |
Exec |
Source |
| 1 |
|
|
// |
| 2 |
|
|
// Copyright (c) 2017-2020 CNRS INRIA |
| 3 |
|
|
// |
| 4 |
|
|
|
| 5 |
|
|
#include "pinocchio/multibody/model.hpp" |
| 6 |
|
|
#include "pinocchio/multibody/data.hpp" |
| 7 |
|
|
#include "pinocchio/algorithm/jacobian.hpp" |
| 8 |
|
|
#include "pinocchio/algorithm/joint-configuration.hpp" |
| 9 |
|
|
#include "pinocchio/algorithm/rnea-second-order-derivatives.hpp" |
| 10 |
|
|
#include "pinocchio/algorithm/rnea-derivatives.hpp" |
| 11 |
|
|
#include "pinocchio/multibody/sample-models.hpp" |
| 12 |
|
|
|
| 13 |
|
|
#include <iostream> |
| 14 |
|
|
|
| 15 |
|
|
#include <boost/test/unit_test.hpp> |
| 16 |
|
|
#include <boost/utility/binary.hpp> |
| 17 |
|
|
|
| 18 |
|
|
BOOST_AUTO_TEST_SUITE(BOOST_TEST_MODULE) |
| 19 |
|
|
|
| 20 |
|
✗ |
BOOST_AUTO_TEST_CASE(test_rnea_derivatives_SO) |
| 21 |
|
|
{ |
| 22 |
|
|
using namespace Eigen; |
| 23 |
|
|
using namespace pinocchio; |
| 24 |
|
|
|
| 25 |
|
✗ |
Model model; |
| 26 |
|
✗ |
buildModels::humanoidRandom(model); |
| 27 |
|
|
|
| 28 |
|
✗ |
Data data(model), data_fd(model); |
| 29 |
|
|
|
| 30 |
|
✗ |
model.lowerPositionLimit.head<3>().fill(-1.); |
| 31 |
|
✗ |
model.upperPositionLimit.head<3>().fill(1.); |
| 32 |
|
|
|
| 33 |
|
✗ |
VectorXd q = randomConfiguration(model); |
| 34 |
|
✗ |
VectorXd v(VectorXd::Random(model.nv)); |
| 35 |
|
✗ |
VectorXd a(VectorXd::Random(model.nv)); |
| 36 |
|
|
|
| 37 |
|
|
// check with only q non-zero |
| 38 |
|
✗ |
Data::Tensor3x dtau2_dq(model.nv, model.nv, model.nv); |
| 39 |
|
✗ |
Data::Tensor3x dtau2_dv(model.nv, model.nv, model.nv); |
| 40 |
|
✗ |
Data::Tensor3x dtau2_dqdv(model.nv, model.nv, model.nv); |
| 41 |
|
✗ |
Data::Tensor3x dtau2_dadq(model.nv, model.nv, model.nv); |
| 42 |
|
✗ |
dtau2_dq.setZero(); |
| 43 |
|
✗ |
dtau2_dv.setZero(); |
| 44 |
|
✗ |
dtau2_dqdv.setZero(); |
| 45 |
|
✗ |
dtau2_dadq.setZero(); |
| 46 |
|
|
|
| 47 |
|
✗ |
ComputeRNEASecondOrderDerivatives( |
| 48 |
|
✗ |
model, data, q, VectorXd::Zero(model.nv), VectorXd::Zero(model.nv), dtau2_dq, dtau2_dv, |
| 49 |
|
|
dtau2_dqdv, dtau2_dadq); |
| 50 |
|
|
|
| 51 |
|
✗ |
Data::Tensor3x dtau2_dq_fd(model.nv, model.nv, model.nv); |
| 52 |
|
✗ |
Data::Tensor3x dtau2_dv_fd(model.nv, model.nv, model.nv); |
| 53 |
|
✗ |
Data::Tensor3x dtau2_dqdv_fd(model.nv, model.nv, model.nv); |
| 54 |
|
✗ |
Data::Tensor3x dtau2_dadq_fd(model.nv, model.nv, model.nv); |
| 55 |
|
✗ |
dtau2_dq_fd.setZero(); |
| 56 |
|
✗ |
dtau2_dv_fd.setZero(); |
| 57 |
|
✗ |
dtau2_dqdv_fd.setZero(); |
| 58 |
|
✗ |
dtau2_dadq_fd.setZero(); |
| 59 |
|
|
|
| 60 |
|
✗ |
MatrixXd drnea_dq_plus(MatrixXd::Zero(model.nv, model.nv)); |
| 61 |
|
✗ |
MatrixXd drnea_dv_plus(MatrixXd::Zero(model.nv, model.nv)); |
| 62 |
|
✗ |
MatrixXd drnea_da_plus(MatrixXd::Zero(model.nv, model.nv)); |
| 63 |
|
|
|
| 64 |
|
✗ |
MatrixXd temp1(MatrixXd::Zero(model.nv, model.nv)); |
| 65 |
|
✗ |
MatrixXd temp2(MatrixXd::Zero(model.nv, model.nv)); |
| 66 |
|
|
|
| 67 |
|
✗ |
VectorXd v_eps(VectorXd::Zero(model.nv)); |
| 68 |
|
✗ |
VectorXd q_plus(model.nq); |
| 69 |
|
✗ |
VectorXd v_plus(model.nv); |
| 70 |
|
|
|
| 71 |
|
✗ |
MatrixXd rnea_partial_dq(MatrixXd::Zero(model.nv, model.nv)); |
| 72 |
|
✗ |
MatrixXd rnea_partial_dv(MatrixXd::Zero(model.nv, model.nv)); |
| 73 |
|
✗ |
MatrixXd rnea_partial_da(MatrixXd::Zero(model.nv, model.nv)); |
| 74 |
|
|
|
| 75 |
|
✗ |
computeRNEADerivatives( |
| 76 |
|
✗ |
model, data, q, VectorXd::Zero(model.nv), VectorXd::Zero(model.nv), rnea_partial_dq, |
| 77 |
|
|
rnea_partial_dv, rnea_partial_da); |
| 78 |
|
|
|
| 79 |
|
✗ |
const double alpha = 1e-7; |
| 80 |
|
|
|
| 81 |
|
✗ |
for (int k = 0; k < model.nv; ++k) |
| 82 |
|
|
{ |
| 83 |
|
✗ |
v_eps[k] += alpha; |
| 84 |
|
✗ |
q_plus = integrate(model, q, v_eps); |
| 85 |
|
✗ |
computeRNEADerivatives( |
| 86 |
|
✗ |
model, data_fd, q_plus, VectorXd::Zero(model.nv), VectorXd::Zero(model.nv), drnea_dq_plus, |
| 87 |
|
|
drnea_dv_plus, drnea_da_plus); |
| 88 |
|
✗ |
temp1 = (drnea_dq_plus - rnea_partial_dq) / alpha; |
| 89 |
|
✗ |
for (int ii = 0; ii < model.nv; ii++) |
| 90 |
|
|
{ |
| 91 |
|
✗ |
for (int jj = 0; jj < model.nv; jj++) |
| 92 |
|
|
{ |
| 93 |
|
✗ |
dtau2_dq_fd(jj, ii, k) = temp1(jj, ii); |
| 94 |
|
|
} |
| 95 |
|
|
} |
| 96 |
|
✗ |
v_eps[k] -= alpha; |
| 97 |
|
|
} |
| 98 |
|
|
|
| 99 |
|
✗ |
Map<VectorXd> mq(dtau2_dq.data(), dtau2_dq.size()); |
| 100 |
|
✗ |
Map<VectorXd> mq_fd(dtau2_dq_fd.data(), dtau2_dq_fd.size()); |
| 101 |
|
✗ |
BOOST_CHECK(mq.isApprox(mq_fd, sqrt(alpha))); |
| 102 |
|
|
|
| 103 |
|
|
// Check with q and a non zero |
| 104 |
|
✗ |
dtau2_dq.setZero(); |
| 105 |
|
✗ |
dtau2_dv.setZero(); |
| 106 |
|
✗ |
dtau2_dqdv.setZero(); |
| 107 |
|
✗ |
dtau2_dadq.setZero(); |
| 108 |
|
✗ |
ComputeRNEASecondOrderDerivatives( |
| 109 |
|
✗ |
model, data, q, VectorXd::Zero(model.nv), a, dtau2_dq, dtau2_dv, dtau2_dqdv, dtau2_dadq); |
| 110 |
|
|
|
| 111 |
|
✗ |
rnea_partial_dq.setZero(); |
| 112 |
|
✗ |
rnea_partial_dv.setZero(); |
| 113 |
|
✗ |
rnea_partial_da.setZero(); |
| 114 |
|
|
|
| 115 |
|
✗ |
dtau2_dq_fd.setZero(); |
| 116 |
|
✗ |
dtau2_dadq_fd.setZero(); |
| 117 |
|
✗ |
drnea_dq_plus.setZero(); |
| 118 |
|
✗ |
drnea_dv_plus.setZero(); |
| 119 |
|
✗ |
drnea_da_plus.setZero(); |
| 120 |
|
|
|
| 121 |
|
✗ |
computeRNEADerivatives( |
| 122 |
|
✗ |
model, data, q, VectorXd::Zero(model.nv), a, rnea_partial_dq, rnea_partial_dv, rnea_partial_da); |
| 123 |
|
|
|
| 124 |
|
✗ |
for (int k = 0; k < model.nv; ++k) |
| 125 |
|
|
{ |
| 126 |
|
✗ |
v_eps[k] += alpha; |
| 127 |
|
✗ |
q_plus = integrate(model, q, v_eps); |
| 128 |
|
✗ |
computeRNEADerivatives( |
| 129 |
|
✗ |
model, data_fd, q_plus, VectorXd::Zero(model.nv), a, drnea_dq_plus, drnea_dv_plus, |
| 130 |
|
|
drnea_da_plus); |
| 131 |
|
✗ |
temp1 = (drnea_dq_plus - rnea_partial_dq) / alpha; |
| 132 |
|
✗ |
temp2 = (drnea_da_plus - rnea_partial_da) / alpha; |
| 133 |
|
✗ |
temp2.triangularView<Eigen::StrictlyLower>() = |
| 134 |
|
✗ |
temp2.transpose().triangularView<Eigen::StrictlyLower>(); |
| 135 |
|
✗ |
for (int ii = 0; ii < model.nv; ii++) |
| 136 |
|
|
{ |
| 137 |
|
✗ |
for (int jj = 0; jj < model.nv; jj++) |
| 138 |
|
|
{ |
| 139 |
|
✗ |
dtau2_dq_fd(jj, ii, k) = temp1(jj, ii); |
| 140 |
|
✗ |
dtau2_dadq_fd(jj, ii, k) = temp2(jj, ii); |
| 141 |
|
|
} |
| 142 |
|
|
} |
| 143 |
|
✗ |
v_eps[k] -= alpha; |
| 144 |
|
|
} |
| 145 |
|
✗ |
Map<VectorXd> maq(dtau2_dadq.data(), dtau2_dadq.size()); |
| 146 |
|
✗ |
Map<VectorXd> maq_fd(dtau2_dadq_fd.data(), dtau2_dadq_fd.size()); |
| 147 |
|
|
|
| 148 |
|
✗ |
BOOST_CHECK(mq.isApprox(mq_fd, sqrt(alpha))); |
| 149 |
|
✗ |
BOOST_CHECK(maq.isApprox(maq_fd, sqrt(alpha))); |
| 150 |
|
|
|
| 151 |
|
|
// Check with q,v and a non zero |
| 152 |
|
✗ |
dtau2_dq.setZero(); |
| 153 |
|
✗ |
dtau2_dv.setZero(); |
| 154 |
|
✗ |
dtau2_dqdv.setZero(); |
| 155 |
|
✗ |
dtau2_dadq.setZero(); |
| 156 |
|
✗ |
ComputeRNEASecondOrderDerivatives( |
| 157 |
|
|
model, data, q, v, a, dtau2_dq, dtau2_dv, dtau2_dqdv, dtau2_dadq); |
| 158 |
|
|
|
| 159 |
|
✗ |
rnea_partial_dq.setZero(); |
| 160 |
|
✗ |
rnea_partial_dv.setZero(); |
| 161 |
|
✗ |
rnea_partial_da.setZero(); |
| 162 |
|
✗ |
computeRNEADerivatives(model, data, q, v, a, rnea_partial_dq, rnea_partial_dv, rnea_partial_da); |
| 163 |
|
|
|
| 164 |
|
✗ |
dtau2_dq_fd.setZero(); |
| 165 |
|
✗ |
dtau2_dadq_fd.setZero(); |
| 166 |
|
✗ |
drnea_dq_plus.setZero(); |
| 167 |
|
✗ |
drnea_dv_plus.setZero(); |
| 168 |
|
✗ |
drnea_da_plus.setZero(); |
| 169 |
|
|
|
| 170 |
|
✗ |
for (int k = 0; k < model.nv; ++k) |
| 171 |
|
|
{ |
| 172 |
|
✗ |
v_eps[k] += alpha; |
| 173 |
|
✗ |
q_plus = integrate(model, q, v_eps); |
| 174 |
|
✗ |
computeRNEADerivatives( |
| 175 |
|
|
model, data_fd, q_plus, v, a, drnea_dq_plus, drnea_dv_plus, drnea_da_plus); |
| 176 |
|
✗ |
temp1 = (drnea_dq_plus - rnea_partial_dq) / alpha; |
| 177 |
|
✗ |
temp2 = (drnea_da_plus - rnea_partial_da) / alpha; |
| 178 |
|
✗ |
temp2.triangularView<Eigen::StrictlyLower>() = |
| 179 |
|
✗ |
temp2.transpose().triangularView<Eigen::StrictlyLower>(); |
| 180 |
|
✗ |
for (int ii = 0; ii < model.nv; ii++) |
| 181 |
|
|
{ |
| 182 |
|
✗ |
for (int jj = 0; jj < model.nv; jj++) |
| 183 |
|
|
{ |
| 184 |
|
✗ |
dtau2_dq_fd(jj, ii, k) = temp1(jj, ii); |
| 185 |
|
✗ |
dtau2_dadq_fd(jj, ii, k) = temp2(jj, ii); |
| 186 |
|
|
} |
| 187 |
|
|
} |
| 188 |
|
✗ |
v_eps[k] -= alpha; |
| 189 |
|
|
} |
| 190 |
|
✗ |
dtau2_dv_fd.setZero(); |
| 191 |
|
✗ |
dtau2_dqdv_fd.setZero(); |
| 192 |
|
✗ |
for (int k = 0; k < model.nv; ++k) |
| 193 |
|
|
{ |
| 194 |
|
✗ |
v_eps[k] += alpha; |
| 195 |
|
✗ |
v_plus = v + v_eps; |
| 196 |
|
✗ |
computeRNEADerivatives( |
| 197 |
|
|
model, data_fd, q, v_plus, a, drnea_dq_plus, drnea_dv_plus, drnea_da_plus); |
| 198 |
|
✗ |
temp1 = (drnea_dv_plus - rnea_partial_dv) / alpha; |
| 199 |
|
✗ |
temp2 = (drnea_dq_plus - rnea_partial_dq) / alpha; |
| 200 |
|
✗ |
for (int ii = 0; ii < model.nv; ii++) |
| 201 |
|
|
{ |
| 202 |
|
✗ |
for (int jj = 0; jj < model.nv; jj++) |
| 203 |
|
|
{ |
| 204 |
|
✗ |
dtau2_dv_fd(jj, ii, k) = temp1(jj, ii); |
| 205 |
|
✗ |
dtau2_dqdv_fd(jj, ii, k) = temp2(jj, ii); |
| 206 |
|
|
} |
| 207 |
|
|
} |
| 208 |
|
✗ |
v_eps[k] -= alpha; |
| 209 |
|
|
} |
| 210 |
|
✗ |
Map<VectorXd> mv(dtau2_dv.data(), dtau2_dv.size()); |
| 211 |
|
✗ |
Map<VectorXd> mv_fd(dtau2_dv_fd.data(), dtau2_dv_fd.size()); |
| 212 |
|
|
|
| 213 |
|
✗ |
Map<VectorXd> mqv(dtau2_dqdv.data(), dtau2_dqdv.size()); |
| 214 |
|
✗ |
Map<VectorXd> mqv_fd(dtau2_dqdv_fd.data(), dtau2_dqdv_fd.size()); |
| 215 |
|
|
|
| 216 |
|
✗ |
BOOST_CHECK(mq.isApprox(mq_fd, sqrt(alpha))); |
| 217 |
|
✗ |
BOOST_CHECK(maq.isApprox(maq_fd, sqrt(alpha))); |
| 218 |
|
✗ |
BOOST_CHECK(mv.isApprox(mv_fd, sqrt(alpha))); |
| 219 |
|
✗ |
BOOST_CHECK(mqv.isApprox(mqv_fd, sqrt(alpha))); |
| 220 |
|
|
|
| 221 |
|
✗ |
Data data2(model); |
| 222 |
|
✗ |
ComputeRNEASecondOrderDerivatives(model, data2, q, v, a); |
| 223 |
|
|
|
| 224 |
|
✗ |
Map<VectorXd> mq2(data2.d2tau_dqdq.data(), (data2.d2tau_dqdq).size()); |
| 225 |
|
✗ |
Map<VectorXd> mv2(data2.d2tau_dvdv.data(), (data2.d2tau_dvdv).size()); |
| 226 |
|
✗ |
Map<VectorXd> mqv2(data2.d2tau_dqdv.data(), (data2.d2tau_dqdv).size()); |
| 227 |
|
✗ |
Map<VectorXd> maq2(data2.d2tau_dadq.data(), (data2.d2tau_dadq).size()); |
| 228 |
|
|
|
| 229 |
|
✗ |
BOOST_CHECK(mq.isApprox(mq2, sqrt(alpha))); |
| 230 |
|
✗ |
BOOST_CHECK(mv.isApprox(mv2, sqrt(alpha))); |
| 231 |
|
✗ |
BOOST_CHECK(mqv.isApprox(mqv2, sqrt(alpha))); |
| 232 |
|
✗ |
BOOST_CHECK(maq.isApprox(maq2, sqrt(alpha))); |
| 233 |
|
|
} |
| 234 |
|
|
|
| 235 |
|
|
BOOST_AUTO_TEST_SUITE_END() |
| 236 |
|
|
|