Abe
|
13a4da21b2
|
misc bugfixes and cleanup from skydio
|
2015-05-13 22:26:24 -07:00 |
cbeall3
|
36a485169d
|
Refactor Ordering parameters. Now compiles and passes with gcc
|
2014-11-17 16:16:52 -05:00 |
Alex Cunningham
|
cb3cec3789
|
Replaced static variable counter inside ISAM2 update() with a member variable to avoid a bug where if multiple instances of ISAM2 are running in the same process, the counter gets incremented in each of them, resulting in very difficult to debug failures.
|
2014-03-27 16:15:29 -04:00 |
Richard Roberts
|
7192bd2f79
|
Removed namespace-functions for ISAM2 optimize* and gradient*, these were only there to try to maintain a consistent interface for linear but really were not correct anyway since ISAM2 is a nonlinear object. They instead caused confusion and made the code complicated since they used a complicated system for updating and caching various components that go into calculating the solution. Replaced all this with much simpler code directly inside updateDelta, which uses clearly-defined functions in the ISAM2::Impl class to calculate the components that go into the solution. Also removed the redundant deltaUpToDate flags - now just checks whether deltaReplacedMask is empty.
|
2014-02-22 16:46:38 -05:00 |
|
Richard Roberts
|
fa5d08d9a6
|
Option to reuse old factor slots in ISAM2
|
2013-11-18 19:23:09 +00:00 |
|
Richard Roberts
|
efc5ea8dce
|
Small cleanup
|
2013-11-12 17:02:46 +00:00 |
|
Richard Roberts
|
ec78d54f37
|
Removed old Index typedef and IndexFormatter
|
2013-11-08 16:35:28 +00:00 |
|
Richard Roberts
|
2dc40087d0
|
Fixed problem in ISAM2 marginalizeLeaves (with Luca, port from other repo)
|
2013-11-05 16:06:06 +00:00 |
Vadim Indelman
|
611fc14aaf
|
Fixed compile errors on some macs
|
2013-08-27 15:53:30 +00:00 |
|
Richard Roberts
|
0bc208e093
|
Ported a fix from Steve for keeping factor order consistent (I don't remember why this is needed)
|
2013-08-20 17:33:10 +00:00 |
|
Richard Roberts
|
6a89db28bb
|
Revert "Ported a fix from Steve for keeping factor order consistent (I don't remember why this is needed)"
This reverts commit 975d470b0fb7637c0f46e967301da4eda0f04b5a.
|
2013-08-20 17:07:40 +00:00 |
|
Richard Roberts
|
4fb7235310
|
Ported a fix from Steve for keeping factor order consistent (I don't remember why this is needed)
|
2013-08-20 16:53:44 +00:00 |
|
Richard Roberts
|
7ce91c7b27
|
Fixed two ISAM2 bugs
|
2013-08-19 22:07:31 +00:00 |
|
Richard Roberts
|
bd89c5fd45
|
Using FastVector (with tbb allocator) instead of vector
|
2013-08-15 17:21:20 +00:00 |
|
Richard Roberts
|
174c2da7d9
|
Switched a couple of data structures - ISAM2::Nodes and Scatter, to tbb versions for better performance
|
2013-08-14 22:54:00 +00:00 |
|
Richard Roberts
|
f9eb5b6e2c
|
Small data structure change in ISAM2::getAffectedFactors for a speedup
|
2013-08-14 19:28:45 +00:00 |
|
Richard Roberts
|
36b3e214b0
|
Fixed timing statement
|
2013-08-14 19:25:02 +00:00 |
|
Richard Roberts
|
845c5dfe43
|
Fixed some timing statements in ISAM2
|
2013-08-14 17:39:36 +00:00 |
|
Richard Roberts
|
46beb6c6a4
|
Re-added ISAM2::marginalizeLeaves optional output arguments from Steves changes in r18185
|
2013-08-14 15:21:10 +00:00 |
Chris Beall
|
d6501bcf9e
|
another fix for boost 1.49 on Linux
|
2013-08-13 14:21:12 +00:00 |
|
Chris Beall
|
38a895507d
|
fix warning about ambiguous parentheses
|
2013-08-13 14:17:05 +00:00 |
|
Richard Roberts
|
7ee1c48ce5
|
Fixed some compile errors with older boost
|
2013-08-13 03:13:14 +00:00 |
|
Richard Roberts
|
0b58b0166a
|
Fixed errors and warnings on linux
|
2013-08-12 21:32:51 +00:00 |
|
Richard Roberts
|
c2284a3d6b
|
Changed FastVector to vector in GFG and ISAM2
|
2013-08-12 18:21:23 +00:00 |
|
Richard Roberts
|
211542fc0e
|
Removed print statement
|
2013-08-11 00:39:44 +00:00 |
|
Richard Roberts
|
63e89656d2
|
Temporary debugging for marginalizeLeaves
|
2013-08-10 15:52:00 +00:00 |
|
Richard Roberts
|
67431ba8ad
|
Fix marginalizeLeaves bug
|
2013-08-10 15:51:52 +00:00 |
|
Richard Roberts
|
789f2bee97
|
ISAM2 compiling and fixed several issues but still some unit tests failing
|
2013-08-09 21:35:47 +00:00 |
|
Richard Roberts
|
a98180f84f
|
Working on converting ISAM2
|
2013-08-08 21:41:29 +00:00 |
|
Richard Roberts
|
c0ccec4656
|
Working on ISAM2
|
2013-08-07 02:56:39 +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 |
|
Richard Roberts
|
e39d100b6a
|
Made most global unit tests compile, includes dogleg, iterative, kalman filter, etc
|
2013-08-06 13:44:22 +00:00 |
|
Richard Roberts
|
ede0805fac
|
Global find+replace to remove Ordered
|
2013-08-05 22:31:44 +00:00 |
|
Richard Roberts
|
cb7eb1b510
|
Find+replace Ordered headers
|
2013-08-05 22:31:33 +00:00 |
|
Richard Roberts
|
34dc67c1af
|
Brought some relevant changes from aspn-imu-factor branch (is correct this time)
|
2013-08-02 16:04:17 +00:00 |
|
Richard Roberts
|
0d05bf5ac5
|
Started on HessianFactor, converted Nonlinear stuff and disabled a lot of more advanced classes so the library compiles with nonlinear using unordered.
|
2013-08-01 21:57:43 +00:00 |
|
Richard Roberts
|
321a286f02
|
Revert "Brought some relevant changes from aspn-imu-factor branch"
This reverts commit b08a11ffb27d14487115f20745f1cea17fcc55ee.
|
2013-07-31 15:26:48 +00:00 |
|
Richard Roberts
|
7345b40feb
|
Brought some relevant changes from aspn-imu-factor branch
|
2013-07-31 15:25:03 +00:00 |
|
Richard Roberts
|
3ade190128
|
Renamed old classes to "Ordered" and removed "Unordered" name from new classes
|
2013-07-29 23:55:40 +00:00 |
|
Richard Roberts
|
77a1e9a485
|
Added ISAM2::marginalCovariance function.
|
2013-06-29 02:19:03 +00:00 |
|
Richard Roberts
|
32066ca358
|
Comment fixes
|
2013-06-20 16:28:41 +00:00 |
|
Richard Roberts
|
84903d05c2
|
Fixed iSAM2 bug where assignment operator and copy constructor may cause null pointer exception when trying to clone cached linear factors, which become null by calling marginalizeLeaves with linear factor caching enabled.
|
2013-06-04 17:34:03 +00:00 |
|
Richard Roberts
|
e6993668ef
|
Disabled extra debug-mode consistency checks that may affect runtime complexity, these are now only enabled when GTSAM_EXTRA_CONSISTENCY_CHECKS is defined
|
2013-05-20 17:26:53 +00:00 |
Stephen Williams
|
9c4e19d141
|
Added optional iSAM2 parameter that forces extra keys into the re-eliminated set. This can be used to force certain keys to be re-ordered.
|
2013-04-08 18:05:48 +00:00 |
|
Richard Roberts
|
2c53a58c9f
|
Added number of factors recalculated to ISAM2Result
|
2013-04-02 17:36:50 +00:00 |
|
Richard Roberts
|
f773a34834
|
Renamed experimentalMarginalizeLeaves to marginalizeLeaves and added comment
|
2013-03-26 20:46:08 +00:00 |
|
Richard Roberts
|
e2aca2039e
|
Further fix in iSAM2 marginalization
|
2013-03-20 13:48:16 +00:00 |
|
Richard Roberts
|
596c5cdae8
|
Revamped ISAM2::marginalizeLeaves again, more tree algorithm instead of relying on sorted keys, hopefully correct this time.
|
2013-03-18 19:28:02 +00:00 |
|
Richard Roberts
|
4278e2b9dc
|
Including BayesTree.h instead of BayesTree-inl.h
|
2013-03-13 18:56:25 +00:00 |
|
Richard Roberts
|
72db4793c5
|
Fixed bugs in iSAM2 leaf marginalization found in HMF testing - set of factors to remove, adding new leaf clique from remaining variables. Also relies on bug fix in BayesTree::removeClique a couple of commits ago.
|
2013-03-05 05:47:27 +00:00 |