Commit Graph

151 Commits (ed098eaec64d185fbf91923b89e9442383b18979)

Author SHA1 Message Date
Varun Agrawal 8280a082bd implemented residual functions for common M-estimators and added corresponding tests 2019-10-02 04:04:34 -04:00
Frank Dellaert a614e5efa1
Merge pull request #115 from borglab/fix/absolute-value
C++11 Absolute Value
2019-09-19 07:45:40 -04:00
Varun Agrawal ff1f37c26f replaced fabs with c++11 std::abs 2019-09-18 18:30:26 -04:00
Alice Anderson f72322c4a4 Added old code back wrapped in deprecated flag 2019-09-02 14:11:56 -07:00
Alice Anderson 2f8238a956 Fixed spelling for Welsch and added to gtsam.h 2019-08-28 17:28:10 -07:00
Abe fbe9aac41c squash local changes on top of gtsam upstream pull from 6/14/2016 2016-06-18 23:13:59 -07:00
yao a0661b3813 Almost done. Still need to figure out how to remove the remaining 3 occurrences. 2016-05-21 12:44:46 -04:00
Alex Hagiopol b6728ef620 Deprecated Vector ones(size_t n). All unit tests pass. 2016-04-15 17:30:54 -04:00
Alex Hagiopol 026a0f59fe Deprecated repeat() in Vector.h. 2016-03-11 23:10:18 -05:00
Alex Hagiopol 2f146e85ef Deprecated esqrt() from Vector.h. 2016-03-11 14:34:19 -05:00
Frank 0372a959ee Many small improvements, bug-fixes, and tests 2016-02-24 11:01:19 -08:00
Frank Dellaert 5afdc24ca7 Merged in feature/typedefPoint3 (pull request #225)
Point3 -> Vector3
2016-02-18 11:13:33 -08:00
Frank a10f462fef Fixed warning 2016-02-17 12:36:57 -08:00
Enrique Fernandez b96aa63370 Add Gaussian::Covariance NOTE about inverse 2016-02-12 15:39:28 -05:00
Enrique Fernandez b9c63ef5df Fix and optimize DCS 2015-08-20 16:03:56 -04:00
Enrique Fernandez 44ae7bfe01 Fix generalized Geman-McClure 2015-08-20 16:03:56 -04:00
Enrique Fernandez 6cdc1de268 Use fabs in Huber condition 2015-08-20 15:45:12 -04:00
Enrique Fernandez 2969d61519 Add Geman-McClure robust kernel 2015-08-18 22:46:36 -04:00
Enrique Fernandez 7d256ff2fb Add DCS robust kernel
DCS (Dynamic Covariance Scaling) from Robust Map Optimization (Agarwal13icra)
2015-08-18 22:46:36 -04:00
Luca f8205bfe02 Merge branch 'develop' into feature/SmartFactors3
Conflicts:
	gtsam/geometry/CalibratedCamera.h
	gtsam/geometry/tests/testCalibratedCamera.cpp
	gtsam/geometry/triangulation.h
	gtsam/nonlinear/ExpressionFactor.h
	gtsam/slam/SmartProjectionFactor.h
	gtsam_unstable/slam/SmartStereoProjectionFactor.h
	tests/testExpressionFactor.cpp
2015-06-19 16:55:30 -04:00
Frank Dellaert 7ec26ff323 Added Whiten 2015-06-10 18:27:10 -04:00
cbeall3 d8b531aaba attempt to fix strange MKL-related error 2015-05-15 00:19:39 -04:00
Abe 13a4da21b2 misc bugfixes and cleanup from skydio 2015-05-13 22:26:24 -07:00
Frank 4ba329c23b Fixed many warnings on Ubuntu 2015-05-12 13:46:24 -07:00
dellaert c5394da29e Better print of isotropic 2015-02-25 19:29:12 +01:00
dellaert c2feb239fb whitenInPlace 2015-02-22 21:50:58 +01:00
dellaert 64000bb747 Cleaned up handling of constraints, removed Constrained smart option as was not used, and discovered there was already an "isConstrained" method. 2014-11-25 13:45:59 +01:00
thduynguyen a61b49dafa remove augmented parts in constraint noise model hack 2014-10-17 15:25:40 -04:00
dellaert 8f493d6ee5 Formatting only 2014-05-31 22:10:51 -04:00
dellaert 89e6e27301 Additional unit test 2014-05-31 20:29:00 -04:00
dellaert e48fa756c0 Fixed bug in the n==0 case 2014-05-28 22:31:20 -04:00
dellaert 0c95959510 Changed to size_t, still fails some assertion 2014-05-28 22:31:20 -04:00
dellaert e730da95c4 Smarter noise models: Diagonal::Sigmas is now actually smart, and Gaussian::SqrtInformation now has a smart flag (default is true) 2014-05-28 09:17:31 -04:00
jing a2c3b8e083 remove changes in gtsam/linear/NoiseModel.cpp to fix unit test failures 2014-01-27 15:13:37 -05:00
dellaert 837085e5d4 Merge branch 'origin/release/2.4.0' 2014-01-25 11:45:59 -05:00
Alex Cunningham b08e889222 removed assertion that apparently fails in a bunch of places 2014-01-04 20:13:00 -05:00
Alex Cunningham dfb3e60171 Fixed issue with serialization by actively changing inf/nan values to zero in constrained noisemodel. Updated WhitenInPlace() and distance(). Apparently this caused a bunch of other unit tests to fail, though. 2014-01-04 20:01:05 -05:00
Frank Dellaert f8fbfaea50 Use precisions_ where possible 2014-01-04 10:14:59 -05:00
Frank Dellaert a3003a0736 Merge branch 'origin/release/2.4.0'
Mainly improvements in EssentialMatrix stuff, but also has the new precisions in the diagonal noise model.
2014-01-04 00:29:32 -05:00
Frank Dellaert 6f64b20a5b The precisions are used many many times in HessianFactor::updatATA, so I added them to the NoiseModel. The invsigmas are no longer optional and hence do not cause extra allocation at every invocation. Same with precision: returns const & 2014-01-02 18:05:44 -05:00
Frank Dellaert d5271dfff7 fixed warning 2013-10-30 12:08:58 +00:00
Richard Roberts 2e3520e0dd Fixed some warnings 2013-10-12 21:06:19 +00:00
Richard Roberts bf40956592 Merged from Hmf6_unordered 970392249a8c50153f24594c9c81acb740cedd06 2013-10-03 19:51:56 +00:00
Richard Roberts 2ad0cd4417 Fixed compile error on linux 2013-10-03 19:10:13 +00:00
Richard Roberts 19602bbb0b Fixed warning 2013-10-03 16:50:18 +00:00
Richard Roberts 95ac34aeeb Updating Hessian in-place instead of computing Hessian for each Jacobian 2013-10-03 16:50:16 +00:00
Duy-Nguyen Ta 63ff1b47c1 support augmented matrices in Matrix's vector_scale_in_place, Constrained's Whiten, WhitenInPlace, and unit(...). These are used in constraints SQP solver with Lagrange multipliers. 2013-10-02 03:52:08 +00:00
Richard Roberts 0bb5d1595d Added Cauchy noise model contributed by Jan Oberländer 2013-09-16 20:58:58 +00:00
Richard Roberts e2f8dd866d Merge branch 'svn/trunk'
Conflicts:
	gtsam/linear/GaussianBayesTree.cpp
	gtsam/linear/GaussianBayesTree.h
	gtsam/linear/VectorValues.h
	gtsam/linear/tests/testVectorValues.cpp
	gtsam_unstable/nonlinear/sequentialSummarization.cpp
2013-08-06 18:33:24 +00:00
Alex Cunningham cee0a234ac Fixing compile problems 2013-08-06 17:50:50 +00:00
Duy-Nguyen Ta ef99e1bf1a add some comments for Constrained model 2013-08-02 13:29:54 +00:00
Richard Roberts 941cd93bf6 Returning empty noise model from QR and fix for backwards-compatibility in old code 2013-07-12 22:27:44 +00:00
Richard Roberts 4718acdabc Put back returning Unit model instead of null from QR for compatibility with old code 2013-07-09 17:50:47 +00:00
Richard Roberts 316befbd2a Added tic 2013-07-01 20:19:28 +00:00
Duy-Nguyen Ta 6eafc9420e update namespace mEstimator for unittests 2013-02-08 00:47:52 +00:00
Duy-Nguyen Ta 25226602c3 change namespace "MEstimator" to "mEstimator" to wrap robust noise models to matlab. Add Tukey. 2013-02-07 19:19:33 +00:00
Richard Roberts 89b50e7679 Renamed tic -> gttic and toc -> gttoc to avoid conflict with PCL tic/toc 2012-10-02 20:18:41 +00:00
Richard Roberts 4876cc7ff7 Revamped timing statements - much easier to use, exception-safe (see email to frankcvs list) 2012-10-02 18:36:39 +00:00
Chris Beall 4297d24c96 changed tabs to spaces for consistent indentation in all of GTSAM 2012-10-02 14:40:07 +00:00
Richard Roberts c323f41e8f Fixed silently ignoring robust noise model when calling whiten on one - now throws exception. Also added function to whiten only the rhs vector in robust noise models. 2012-09-25 18:45:28 +00:00
Richard Roberts 7d02043855 Removed unused include 2012-08-27 13:21:48 +00:00
Richard Roberts 5ddd28ba18 Removed unused noise model cholesky 2012-08-22 22:40:20 +00:00
Richard Roberts b6ddb2197b Removed unused "zero_below_diagonal" flag in Matrix QR 2012-08-22 22:40:04 +00:00
Duy-Nguyen Ta 6f1ea87a00 clean up noise model: Remove Shared[NoiseModel] classes and headers, typedef for backward compatibility in NoiseModel.h. Fix all tests and examples to create shared noise models through static functions in noise model classes. Fix MATLAB wrapper and examples as well. Add tests for MATLAB examples 2012-06-22 19:36:49 +00:00
Alex Cunningham 18ba9bcb3d Removed sample() functions and global random number generator 2012-06-22 16:38:01 +00:00
Duy-Nguyen Ta 0a3d9975f3 adding smart flags for several static noise model creation function 2012-06-19 06:04:31 +00:00
Richard Roberts dacff98f9f Using boost timer instead of chrono to support older boost versions 2012-06-05 20:03:31 +00:00
Frank Dellaert 861d1d7dbd Fixed small formatting glitches 2012-05-21 05:17:35 +00:00
Richard Roberts 08b87be4b5 Fixed parenthesis typos 2012-04-09 13:21:14 +00:00
Alex Cunningham 0fe13ae3ca Fixed noisemodel compile error, moved remaining nonlinear constraints to NonlinearEquality 2011-11-10 02:05:03 +00:00
Alex Cunningham 98410ca5c9 Merge branch 'unified_constraints'
Conflicts:
	gtsam/linear/NoiseModel.cpp
	gtsam/linear/NoiseModel.h
	gtsam/linear/tests/testNoiseModel.cpp
2011-11-09 22:15:40 +00:00
Alex Cunningham 70afdfb7d3 Added mu to noisemodel 2011-11-09 22:09:52 +00:00
Frank Dellaert 5a7b2e0692 Added default string to print methods 2011-11-05 14:26:41 +00:00
Richard Roberts af3c12a7df Implemented N-way factor base class in NoiseModelFactor, added NonlinearFactor{3-6}, adapted NonlinearFactor1,2,3 and NonlinearConstraint1,2,3 to derive from NoiseModelFactor in a minimal way 2011-10-03 04:24:24 +00:00
Yong-Dian Jian b456733cd0 add comments and unit test 2011-09-11 04:39:28 +00:00
Frank Dellaert 24ea02fe69 Documentation and TODO 2011-09-09 21:16:57 +00:00
Yong-Dian Jian e9504bd470 add a null model for comparison 2011-08-29 04:51:17 +00:00
Yong-Dian Jian 25cd24409a add an option to reweight according to scalar or block 2011-08-28 02:47:17 +00:00
Yong-Dian Jian fd07d61a2e Add a robust noise model. Change SharedGaussian to SharedNoiseModel 2011-08-26 21:41:01 +00:00
Alex Cunningham a22586362b Changed the default Matrix to use column major, rather than row major. Removed Matrix-inl.h, as it isn't used 2011-06-04 22:15:23 +00:00
Alex Cunningham 2e942f08ac Removed some extraneous comments from eigen transition 2011-05-20 18:43:37 +00:00
Alex Cunningham e20561be73 Merge branch 'eigen' 2011-05-20 13:52:08 +00:00
Alex Cunningham fc3b668171 Removed boost.serialization export keys from noisemodel because they don't work outside of one test 2011-03-08 18:13:49 +00:00
Alex Cunningham 1c1e47ca1f Added serialization export keys directly to NoiseModel 2011-03-04 14:56:18 +00:00
Alex Cunningham d96b114618 Changed diagonal noisemodel to use boost.optional for inverse sigmas to avoid ever storing inf 2011-03-03 21:14:55 +00:00
Alex Cunningham a87a52035d Serialization for (some) nonlinear factors now works, added virtual destructors to factor classes that needed them. 2011-03-03 17:16:13 +00:00
Alex Cunningham 4865f1a64d Added more serialization functionality to noiseModel, but SharedGaussians segfault on test, so tests for Nonlinear graphs are commented out 2011-02-23 05:19:07 +00:00
Richard Roberts 3dc36369d9 Added a "matrix order" parameter to choleskyCareful so that it can be used to factor the frontal piece in choleskyPartial. choleskyCareful also now returns a bool indicating whether the input matrix is full-rank. Also added an additional Cholesky unit test. 2011-02-04 00:47:08 +00:00
Richard Roberts 4880257e69 Merging hessianfactor branch, Cholesky is now default and GaussianFactor is virtual (see email to frankcvs) 2011-01-20 22:22:00 +00:00
Richard Roberts a8a584d38e Rolling back last commit (undoing merge of hessianfactor branch) - does not compile on Linux, could be compiler or boost version problems. Also will probably break Alex's stuff so will work that out before merging. 2010-12-17 01:11:28 +00:00
Richard Roberts de1892016d Merged in hessianfactor branch, Cholesky is now default. This merge also includes improved timing statements with automatic outlining and low overhead 2010-12-17 00:51:51 +00:00
Richard Roberts fd381e97bc Slight speedup in Cholesky 2010-11-29 03:05:54 +00:00
Richard Roberts 5e00148315 Cholesky working but disabled by default 2010-11-29 01:31:51 +00:00
Richard Roberts 2fc5ab3bdb Extra debugging checks for Cholesky 2010-11-19 17:20:04 +00:00
Richard Roberts fd597762c7 In-progress integrating Cholesky into solving 2010-11-08 19:31:40 +00:00
Richard Roberts 9c5068e90a Removed bad assertion 2010-11-02 23:00:43 +00:00
Richard Roberts fca73852f1 Removed assert(false) that I thought was never called 2010-10-29 16:11:06 +00:00
Richard Roberts 078f07650f Added option to use LAPACK QR with -DUSE_LAPACK_QR 2010-10-29 14:45:20 +00:00
Richard Roberts a735b7cdeb Virtual equals and print methods in NonlinearFactor so that printing nonlinear factors works correctly 2010-10-26 20:21:52 +00:00
Alex Cunningham a46adeb1b4 Library now compiles with --enable-lapack off 2010-10-26 12:31:40 +00:00