| Line |
Branch |
Exec |
Source |
| 1 |
|
|
// Copyright (c) 2018 CNRS |
| 2 |
|
|
// Authors: Joseph Mirabel |
| 3 |
|
|
// |
| 4 |
|
|
|
| 5 |
|
|
// Redistribution and use in source and binary forms, with or without |
| 6 |
|
|
// modification, are permitted provided that the following conditions are |
| 7 |
|
|
// met: |
| 8 |
|
|
// |
| 9 |
|
|
// 1. Redistributions of source code must retain the above copyright |
| 10 |
|
|
// notice, this list of conditions and the following disclaimer. |
| 11 |
|
|
// |
| 12 |
|
|
// 2. Redistributions in binary form must reproduce the above copyright |
| 13 |
|
|
// notice, this list of conditions and the following disclaimer in the |
| 14 |
|
|
// documentation and/or other materials provided with the distribution. |
| 15 |
|
|
// |
| 16 |
|
|
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS |
| 17 |
|
|
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
| 18 |
|
|
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR |
| 19 |
|
|
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT |
| 20 |
|
|
// HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, |
| 21 |
|
|
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
| 22 |
|
|
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, |
| 23 |
|
|
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY |
| 24 |
|
|
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
| 25 |
|
|
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
| 26 |
|
|
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH |
| 27 |
|
|
// DAMAGE. |
| 28 |
|
|
|
| 29 |
|
|
#include <hpp/constraints/manipulability.hh> |
| 30 |
|
|
|
| 31 |
|
|
namespace hpp { |
| 32 |
|
|
namespace constraints { |
| 33 |
|
✗ |
Manipulability::Manipulability(DifferentiableFunctionPtr_t function, |
| 34 |
|
✗ |
DevicePtr_t robot, std::string name) |
| 35 |
|
|
: DifferentiableFunction(function->inputSize(), |
| 36 |
|
|
function->inputDerivativeSize(), 1, name), |
| 37 |
|
✗ |
function_(function), |
| 38 |
|
✗ |
robot_(robot), |
| 39 |
|
✗ |
J_(function->outputDerivativeSize(), function->inputDerivativeSize()) { |
| 40 |
|
✗ |
activeParameters_ = function->activeParameters(); |
| 41 |
|
✗ |
activeDerivativeParameters_ = function->activeDerivativeParameters(); |
| 42 |
|
✗ |
cols_ = Eigen::BlockIndex::fromLogicalExpression(activeDerivativeParameters_); |
| 43 |
|
✗ |
J_JT_.resize(J_.rows(), J_.rows()); |
| 44 |
|
|
} |
| 45 |
|
|
|
| 46 |
|
✗ |
void Manipulability::impl_compute(LiegroupElementRef res, |
| 47 |
|
|
vectorIn_t arg) const { |
| 48 |
|
✗ |
assert(cols_.cols().size() > 0); |
| 49 |
|
|
|
| 50 |
|
✗ |
function_->jacobian(J_, arg); |
| 51 |
|
|
value_type logAbsDeterminant; |
| 52 |
|
|
|
| 53 |
|
|
// ------------ SVD --------------------------------------------------- // |
| 54 |
|
✗ |
J_JT_ = cols_.rview(J_); |
| 55 |
|
✗ |
Eigen::JacobiSVD<matrix_t> svd(J_JT_); |
| 56 |
|
✗ |
logAbsDeterminant = svd.singularValues() |
| 57 |
|
✗ |
.array() |
| 58 |
|
✗ |
.cwiseMax(std::numeric_limits<value_type>::min()) |
| 59 |
|
✗ |
.log10() |
| 60 |
|
✗ |
.sum(); |
| 61 |
|
|
|
| 62 |
|
|
/* |
| 63 |
|
|
// ------------ Other decomposition methods --------------------------- // |
| 64 |
|
|
|
| 65 |
|
|
// 1. Compute J * J^T |
| 66 |
|
|
if (cols_.cols().size() > 1) { |
| 67 |
|
|
typedef typename Eigen::ColBlockIndices::View<const matrix_t>::type |
| 68 |
|
|
MatrixView_t; MatrixView_t J (cols_.rview(J_)); |
| 69 |
|
|
//std::cout << J.eval() << std::endl; |
| 70 |
|
|
J_JT_.setZero(); |
| 71 |
|
|
for (MatrixView_t::block_iterator block (J); block.valid(); ++block) |
| 72 |
|
|
J_JT_.noalias() += J._block(block) * J._block(block).transpose(); |
| 73 |
|
|
} else { |
| 74 |
|
|
const segment_t& s = cols_.cols()[0]; |
| 75 |
|
|
//std::cout << J_.middleCols(s.first, s.second) << std::endl; |
| 76 |
|
|
J_JT_.noalias() = J_.middleCols(s.first, s.second) |
| 77 |
|
|
* J_.middleCols(s.first, s.second).transpose(); |
| 78 |
|
|
} |
| 79 |
|
|
|
| 80 |
|
|
// 2. Compute decomposition |
| 81 |
|
|
|
| 82 |
|
|
// 2.1 LDLT |
| 83 |
|
|
Eigen::LDLT<matrix_t> ldlt (J_JT_); |
| 84 |
|
|
logAbsDeterminant = ldlt.matrixL().nestedExpression().diagonal().array() |
| 85 |
|
|
.cwiseMax(std::numeric_limits<value_type>::min()) |
| 86 |
|
|
.log10() |
| 87 |
|
|
.sum(); |
| 88 |
|
|
logAbsDeterminant += ldlt.vectorD().array() |
| 89 |
|
|
.cwiseMax(std::numeric_limits<value_type>::min()) |
| 90 |
|
|
.log10() |
| 91 |
|
|
.sum(); |
| 92 |
|
|
// 2.2 QRs (FullPiv is more robust that ColPiv) |
| 93 |
|
|
//Eigen::ColPivHouseholderQR<matrix_t> qr (J_JT_); |
| 94 |
|
|
Eigen::FullPivHouseholderQR<matrix_t> qr (J_JT_); |
| 95 |
|
|
logAbsDeterminant = qr.logAbsDeterminant(); |
| 96 |
|
|
// */ |
| 97 |
|
|
|
| 98 |
|
|
// This funcion will be used as a cost function whose squared norm is to |
| 99 |
|
|
// be minimized. |
| 100 |
|
✗ |
res.vector()[0] = std::max(-logAbsDeterminant, 0.); |
| 101 |
|
|
} |
| 102 |
|
|
|
| 103 |
|
✗ |
void Manipulability::impl_jacobian(matrixOut_t jacobian, vectorIn_t arg) const { |
| 104 |
|
✗ |
finiteDifferenceCentral(jacobian, arg, robot_, 1e-8); |
| 105 |
|
|
} |
| 106 |
|
|
} // namespace constraints |
| 107 |
|
|
} // namespace hpp |
| 108 |
|
|
|