-
Notifications
You must be signed in to change notification settings - Fork 56
/
local_trajectory_builder_3d.cc
1182 lines (1047 loc) · 45.6 KB
/
local_trajectory_builder_3d.cc
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
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
* Copyright 2016 The Cartographer Authors
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "cartographer/mapping/internal/3d/local_trajectory_builder_3d.h"
#include <memory>
#include "cartographer/common/make_unique.h"
#include "cartographer/common/time.h"
#include "cartographer/mapping/internal/3d/scan_matching/rotational_scan_matcher.h"
#include "cartographer/mapping/proto/3d/local_trajectory_builder_options_3d.pb.h"
#include "cartographer/mapping/internal/3d/gravity_factor/gravity_factor.h"
#include <chrono>
#include "glog/logging.h"
namespace cartographer {
namespace mapping {
using namespace std;
using namespace chrono;
static auto* kLocalSlamLatencyMetric = metrics::Gauge::Null();
static auto* kRealTimeCorrelativeScanMatcherScoreMetric =
metrics::Histogram::Null();
static auto* kCeresScanMatcherCostMetric = metrics::Histogram::Null();
static auto* kScanMatcherResidualDistanceMetric = metrics::Histogram::Null();
static auto* kScanMatcherResidualAngleMetric = metrics::Histogram::Null();
LocalTrajectoryBuilder3D::LocalTrajectoryBuilder3D(
const mapping::proto::LocalTrajectoryBuilderOptions3D& options,
const std::vector<std::string>& expected_range_sensor_ids)
: options_(options),
active_submaps_(options.submaps_options()),
motion_filter_(options.motion_filter_options()),
real_time_correlative_scan_matcher_(
common::make_unique<scan_matching::RealTimeCorrelativeScanMatcher3D>(
options_.real_time_correlative_scan_matcher_options())),
ceres_scan_matcher_(
common::make_unique<scan_matching::CeresScanMatcher3D>(
options_.ceres_scan_matcher_options())),
accumulated_range_data_{Eigen::Vector3f::Zero(), {}, {}},
range_data_synchronizer_(expected_range_sensor_ids) {
scan_period_ = options_.scan_period();
eable_mannually_discrew_ = options_.eable_mannually_discrew();
frames_for_static_initialization_ = options_.frames_for_static_initialization();
frames_for_dynamic_initialization_ = options_.frames_for_dynamic_initialization();
g_est_win_size_ = options_.frames_for_online_gravity_estimate();
const float imuAccNoise = options_.imu_options().acc_noise();
const float imuGyrNoise = options_.imu_options().gyr_noise();
const float imuAccBiasN = options_.imu_options().acc_bias_noise();
const float imuGyrBiasN = options_.imu_options().gyr_bias_noise();
const float imuGravity = options_.imu_options().gravity();
const float prior_pose_n = options_.imu_options().prior_pose_noise();
const float prior_gravity_noise = options_.imu_options().prior_gravity_noise();
const float ceres_pose_n_t = options_.imu_options().ceres_pose_noise_t();
const float ceres_pose_n_r = options_.imu_options().ceres_pose_noise_r();
const float ceres_pose_n_t_1 =
options_.imu_options().ceres_pose_noise_t_drift();
const float ceres_pose_n_r_1 =
options_.imu_options().ceres_pose_noise_r_drift();
preint_param_ = gtsam::PreintegrationParams::MakeSharedU(imuGravity);
preint_param_->accelerometerCovariance = gtsam::Matrix33::Identity(3,3)
* pow(imuAccNoise, 2);
preint_param_->gyroscopeCovariance =
gtsam::Matrix33::Identity(3,3) * pow(imuGyrNoise, 2);
preint_param_->integrationCovariance =
gtsam::Matrix33::Identity(3,3) * pow(1e-4, 2);
prior_pose_noise_ = gtsam::noiseModel::Diagonal::Sigmas(
(gtsam::Vector(6) << prior_pose_n, prior_pose_n, prior_pose_n,
prior_pose_n, prior_pose_n, prior_pose_n).finished());
prior_vel_noise_ = gtsam::noiseModel::Isotropic::Sigma(3, 1e4);
prior_bias_noise_ = gtsam::noiseModel::Isotropic::Sigma(6, 1e-2);
//TODO:
prior_gravity_noise_ = gtsam::noiseModel::Diagonal::Sigmas(
(gtsam::Vector(2) << prior_gravity_noise, prior_gravity_noise).finished());
correction_noise_ = gtsam::noiseModel::Diagonal::Sigmas(
(gtsam::Vector(6) <<
ceres_pose_n_t, ceres_pose_n_t, ceres_pose_n_t,
ceres_pose_n_r, ceres_pose_n_r, ceres_pose_n_r).finished());
correction_noise_2_ = gtsam::noiseModel::Diagonal::Sigmas(
(gtsam::Vector(6) <<
ceres_pose_n_t_1, ceres_pose_n_t_1, ceres_pose_n_t_1,
ceres_pose_n_r_1, ceres_pose_n_r_1, ceres_pose_n_r_1).finished());
noise_model_between_bias_ = (gtsam::Vector(6) <<
imuAccBiasN, imuAccBiasN, imuAccBiasN,
imuGyrBiasN, imuGyrBiasN, imuGyrBiasN).finished();
// for vins initial integrator initialization
imu_noise_.ACC_N = imuAccNoise;
imu_noise_.ACC_W = imuAccBiasN;
imu_noise_.GYR_N = imuGyrNoise;
imu_noise_.GYR_W = imuGyrBiasN;
init_integrator_.reset(new IntegrationBase(INIT_BA, INIT_BW, imu_noise_));
InitCircularBuffers();
}
LocalTrajectoryBuilder3D::~LocalTrajectoryBuilder3D() {}
pcl::PointCloud<pcl::PointXYZI>::Ptr LocalTrajectoryBuilder3D::cvtPointCloud(
const cartographer::sensor::TimedPointCloudOriginData& point_cloud){
pcl::PointCloud<pcl::PointXYZI>::Ptr pcl_cloud;
pcl_cloud.reset(new pcl::PointCloud<pcl::PointXYZI>());
pcl_cloud->resize(point_cloud.ranges.size());
for (size_t i = 0; i < point_cloud.ranges.size(); ++i) {
const auto& range = point_cloud.ranges.at(i);
pcl_cloud->points[i].x = range.point_time[0];
pcl_cloud->points[i].y = range.point_time[1];
pcl_cloud->points[i].z = range.point_time[2];
pcl_cloud->points[i].intensity = range.point_time[3];
}
return pcl_cloud;
}
pcl::PointCloud<pcl::PointXYZI>::Ptr LocalTrajectoryBuilder3D::
cvtPointCloudAndDiscrew(
const cartographer::sensor::TimedPointCloudOriginData& point_cloud,
const transform::Rigid3d& rel_trans
/*rel_trans is the transform from last frame to current frame*/){
pcl::PointCloud<pcl::PointXYZI>::Ptr pcl_cloud;
pcl_cloud.reset(new pcl::PointCloud<pcl::PointXYZI>());
pcl_cloud->resize(point_cloud.ranges.size());
const double sample_period = scan_period_;
transform::Rigid3d tracking_pose_in_start, tracking_pose_in_end;
for (size_t i = 0; i < point_cloud.ranges.size(); ++i) {
const auto& range = point_cloud.ranges.at(i);
double s = (sample_period + range.point_time[3]) / sample_period;
// pose in last frame
InterpolatePose(s, rel_trans, tracking_pose_in_start);
// transform to current frame
tracking_pose_in_end = rel_trans.inverse() * tracking_pose_in_start;
Eigen::Vector3d pt, pt_in_end;
pt << range.point_time[0], range.point_time[1],range.point_time[2];
pt_in_end = tracking_pose_in_end * pt;
pcl_cloud->points[i].x = pt_in_end[0];
pcl_cloud->points[i].y = pt_in_end[1];
pcl_cloud->points[i].z = pt_in_end[2];
pcl_cloud->points[i].intensity = range.point_time[3];
}
return pcl_cloud;
}
void LocalTrajectoryBuilder3D::AddImuData(const sensor::ImuData& imu_data) {
if(!imu_initialized_){
if(options_.enable_ndt_initialization() && init_integrator_){
double imu_time = common::ToSecondsStamp(imu_data.time);
double dt = (last_imu_time_ini_ < 0)
? (1.0 / 500.0) : (imu_time - last_imu_time_ini_);
last_imu_time_ini_ = imu_time;
init_integrator_->push_back(
dt, imu_data.linear_acceleration, imu_data.angular_velocity);
}else{
init_imu_buffer_opt_.push_back(imu_data);
}
}
//由于外层调用时需要加锁，因此IMU的添加与点云的添加其实是交替进行的
if(imu_initialized_ && imu_integrator_opt_){
// imu_que_opt_.push_back(imu_data);
double imu_time = common::ToSecondsStamp(imu_data.time);
double dt = (last_imu_time_opt_ < 0)
? (1.0 / 500.0) : (imu_time - last_imu_time_opt_);
last_imu_time_opt_ = imu_time;
// integrate this single imu message
imu_integrator_opt_->integrateMeasurement(
gtsam::Vector3(imu_data.linear_acceleration.x(),
imu_data.linear_acceleration.y(),
imu_data.linear_acceleration.z()),
gtsam::Vector3(imu_data.angular_velocity.x(),
imu_data.angular_velocity.y(),
imu_data.angular_velocity.z()), dt);
gtsam::NavState current_state =
imu_integrator_opt_->predict(prev_state_, prev_bias_);
// 存储递推值用于纠正点云／用于scan matching的初始值
predicted_states_.push_back({imu_data.time, current_state});
}
}
void LocalTrajectoryBuilder3D::InitializeStatic(){
Eigen::Vector3d accel_accum;
Eigen::Vector3d gyro_accum;
int num_readings = 0;
accel_accum.setZero();
gyro_accum.setZero();
for(const auto& entry : init_imu_buffer_opt_){
accel_accum += entry.linear_acceleration;
gyro_accum += entry.angular_velocity;
num_readings++;
}
Eigen::Vector3d accel_mean = accel_accum / num_readings;
Eigen::Vector3d gyro_mean = gyro_accum / num_readings;
g_vec_ << 0.0, 0.0, -options_.imu_options().gravity();
P_.setZero();
V_.setZero();
//frame I to frame G
R_ = Eigen::Quaternion<double>::FromTwoVectors(accel_mean, -g_vec_);
Ba_ = R_.transpose() * g_vec_ + accel_mean;
Bg_ = gyro_mean;
InitializeIMU();
}
void LocalTrajectoryBuilder3D::InitilizeByNDT(
const common::Time& time,
const sensor::TimedPointCloudOriginData& synchronized_data){
Rigid3dWithVINSPreintegrator plt;
double stamp = common::ToSecondsStamp(time);
if(!last_scan_){
if(last_imu_time_ini_ < 0) return; // still no imu come in
last_stamp_ = stamp;
pcl::PointCloud<pcl::PointXYZI>::Ptr cur_scan = cvtPointCloud(
synchronized_data);
last_scan_ = cur_scan;
linear_velocity_ << 0., 0., 0.;
angular_velocity_ << 0., 0., 0.;
INIT_BA << 0,0,0;
INIT_BW << 0,0,0;
voxel_filter_pre_.setLeafSize(0.2, 0.2, 0.2);
voxel_filter_cur_.setLeafSize(0.2, 0.2, 0.2);
// Setting scale dependent NDT parameters
// Setting minimum transformation difference for termination condition.
ndt_.setTransformationEpsilon(0.01);
// Setting maximum step size for More-Thuente line search.
ndt_.setStepSize(0.1);
//Setting Resolution of NDT grid structure (VoxelGridCovariance).
ndt_.setResolution(1.0);
// Setting max number of registration iterations.
ndt_.setMaximumIterations(35);
// plt.transform = transform::Rigid3d::Identity();
// plt.pre_integration = nullptr;
CHECK(buffered_scan_count_ == 0)
<< "buffered_scan_count_ must be zero here!";
all_laser_transforms_[buffered_scan_count_++] = plt;
init_integrator_->resetIntegration(INIT_BA, INIT_BW, imu_noise_);
return;
}else{
dt_ = stamp - last_stamp_;
Eigen::Matrix3f R;
Eigen::Vector3f t;
// set initial alignment estimate
// Eigen::Quaternionf init_rotation_q =
// transform::AngleAxisVectorToRotationQuaternion(
// Eigen::Vector3f(angular_velocity_ * dt_)).normalized();
Eigen::Quaternionf init_q = init_integrator_->deltaQij().cast<float>();
Eigen::Matrix3f init_R(init_q);
Eigen::Vector3f init_t(linear_velocity_ * dt_);
Eigen::Matrix4f init_guess = Eigen::Matrix4f::Identity();
init_guess.block(0,0,3,3) = init_R;
init_guess.block(0,3,3,1) = init_t;
// discrewing points or not?
// transform::Rigid3d rel_trans = transform::Rigid3d(
// linear_velocity_.cast<double>() * dt_, init_integrator_->delta_q);
// pcl::PointCloud<pcl::PointXYZI>::Ptr cur_scan = cvtPointCloudAndDiscrew(
// synchronized_data, rel_trans);
pcl::PointCloud<pcl::PointXYZI>::Ptr cur_scan = cvtPointCloud(
synchronized_data);
MatchByNDT(last_scan_, cur_scan, init_guess, R, t);
last_scan_ = cur_scan;
// insert laser transform pair
plt.transform = transform::Rigid3d(t.cast<double>(),
Eigen::Quaterniond(all_laser_transforms_[buffered_scan_count_ - 1]
.transform.rotation().toRotationMatrix()
* R.cast<double>()));
std::shared_ptr<IntegrationBase> imu_int_init_pt;
imu_int_init_pt.reset(new IntegrationBase(*init_integrator_));
plt.pre_integration = imu_int_init_pt;
all_laser_transforms_[buffered_scan_count_++] = plt;
init_integrator_->resetIntegration(INIT_BA, INIT_BW, imu_noise_);
// update velocity and stamp
linear_velocity_ = t / dt_;
angular_velocity_ = transform::RotationQuaternionToAngleAxisVector(
Eigen::Quaternionf(R)) / dt_;
last_stamp_ = stamp;
}
if(buffered_scan_count_ == frames_for_dynamic_initialization_ + 1){
if(!AlignWithWorld()){
LOG(ERROR)<<"Initialization failed! Perform re-initialization...";
buffered_scan_count_ = 1;
InitCircularBuffers();
}else{
// init succeeded, initialize the reference frame's state
P_ = Ps_[frames_for_dynamic_initialization_];
V_ = Vs_[frames_for_dynamic_initialization_];
R_ = Rs_[frames_for_dynamic_initialization_];
Ba_ = Bas_[frames_for_dynamic_initialization_];
Bg_ = Bgs_[frames_for_dynamic_initialization_];
InitializeIMU();
}
}
}
void LocalTrajectoryBuilder3D::InitializeIMU(){
// setting up the IMU integration for IMU prediction
gtsam::imuBias::ConstantBias prior_imu_bias((gtsam::Vector(6)
<< Ba_[0], Ba_[1], Ba_[2], Bg_[0], Bg_[1], Bg_[2]).finished());
imu_integrator_opt_ =
new gtsam::PreintegratedImuMeasurements(preint_param_, prior_imu_bias);
Eigen::Quaterniond q(R_);
gtsam::Pose3 pose_start = gtsam::Pose3(
gtsam::Rot3::Quaternion(q.w(), q.x(), q.y(), q.z()),
gtsam::Point3(P_.x(), P_.y(), P_.z()));
gtsam::Velocity3 v_start = gtsam::Velocity3(V_[0], V_[1], V_[2]);
LOG(INFO)<<"Initialization paras: ";
LOG(INFO)<<"P: "<<P_.x()<<","<<P_.y()<<","<<P_.z();
LOG(INFO)<<"V: "<<V_[0]<<","<<V_[1]<<","<<V_[2];
LOG(INFO)<<"Q: "<<q.w()<<","<<q.x()<<","<<q.y()<<","<<q.z();
LOG(INFO)<<"Ba: "<<Ba_[0]<<","<<Ba_[1]<<","<<Ba_[2];
LOG(INFO)<<"Bg: "<<Bg_[0]<<","<<Bg_[1]<<","<<Bg_[2];
prev_state_ = gtsam::NavState(pose_start, v_start);
prev_bias_ = prior_imu_bias;
prev_state_odom_ = prev_state_;
prev_bias_odom_ = prev_bias_;
imu_initialized_ = true;
}
std::unique_ptr<LocalTrajectoryBuilder3D::MatchingResult>
LocalTrajectoryBuilder3D::AddRangeData(
const std::string& sensor_id,
const sensor::TimedPointCloudData& unsynchronized_data) {
sensor::TimedPointCloudOriginData synchronized_data
= range_data_synchronizer_.AddRangeData(
sensor_id, unsynchronized_data, eable_mannually_discrew_);
if (synchronized_data.ranges.empty()) {
// LOG(INFO) << "Range data collator filling buffer.";
return nullptr;
}
const common::Time& time = synchronized_data.time;
time_point_cloud_ = time;
if(!imu_initialized_){
if(options_.enable_ndt_initialization()){
InitilizeByNDT(time, synchronized_data);
}else{
if(accumulated_frame_num++ > frames_for_static_initialization_){
InitializeStatic();
}
}
return nullptr;
}
CHECK(!synchronized_data.ranges.empty());
CHECK_LE(synchronized_data.ranges.back().point_time[3], 0.1f);
const common::Time time_first_point = time
+ common::FromSeconds(synchronized_data.ranges.front().point_time[3]);
if (num_accumulated_ == 0) {
accumulation_started_ = std::chrono::steady_clock::now();
}
std::vector<sensor::TimedPointCloudOriginData::RangeMeasurement> hits =
sensor::VoxelFilter(0.5f * options_.voxel_filter_size())
.Filter(synchronized_data.ranges);
std::vector<transform::Rigid3f> hits_poses;
transform::Rigid3d tmp_pose;
hits_poses.reserve(hits.size());
bool warned = false;
//即使是插入子地图的第一帧，也是经过初始化步骤的，同样可以进行相对运动的矫正
first_scan_to_insert_ = false;
if(first_scan_to_insert_){
hits_poses = std::vector<transform::Rigid3f>(
hits.size(), PoseFromGtsamNavState(prev_state_).cast<float>());
first_scan_to_insert_ = false;
}else{
size_t idx0, idx1;
idx0 = idx1 = 0;
transform::Rigid3d cur_state_pre, rel_trans;
// FindIdxFromOdomQuene(time, idx0, idx1);
// if(idx0 != idx1){
// InterpolatePose(time, predicted_states_.at(idx0).first,
// predicted_states_.at(idx1).first, predicted_states_.at(idx0).second,
// predicted_states_.at(idx1).second, cur_state_pre);
// }else{
// if(idx0 == 0 ){
// LOG(WARNING)<<"Maybe IMU quene is empty?";
// return nullptr;
// }
// cur_state_pre = PoseFromGtsamNavState(predicted_states_.at(idx0).second);
// }
if(predicted_states_.empty()) return nullptr;
cur_state_pre = PoseFromGtsamNavState(predicted_states_.back().second);
rel_trans = PoseFromGtsamNavState(prev_state_).inverse() * cur_state_pre;
if(std::abs(hits.front().point_time[3]) < 1e-3){//没有单点的时间戳
hits_poses = std::vector<transform::Rigid3f>(
hits.size(), cur_state_pre.cast<float>());
LOG(WARNING)<<"Not discrewing!";
}else{
const double sample_period = scan_period_;
for (const auto& hit : hits) {
common::Time time_point = time + common::FromSeconds(hit.point_time[3]);
//注意，hit.point_time[3]是小于０的值
//当前帧的最后一个点为０，帧的时间戳为最后一个点的采集时间
double s = (sample_period + hit.point_time[3]) / sample_period;
InterpolatePose(s, rel_trans, tmp_pose);
hits_poses.push_back(
(PoseFromGtsamNavState(prev_state_) * tmp_pose).cast<float>());
}
}
}
TrimStatesCache(time);
if (num_accumulated_ == 0) {
// 'accumulated_range_data_.origin' is not used.
accumulated_range_data_ = sensor::RangeData{{}, {}, {}};
}
for (size_t i = 0; i < hits.size(); ++i) {
const Eigen::Vector3f hit_in_local =
hits_poses[i] * hits[i].point_time.head<3>();
const Eigen::Vector3f origin_in_local =
hits_poses[i] * synchronized_data.origins.at(hits[i].origin_index);
const Eigen::Vector3f delta = hit_in_local - origin_in_local;
const float range = delta.norm();
if (range >= options_.min_range()) {
if (range <= options_.max_range()) {
accumulated_range_data_.returns.push_back(hit_in_local);
} else {
// We insert a ray cropped to 'max_range' as a miss for hits beyond the
// maximum range. This way the free space up to the maximum range will
// be updated.
accumulated_range_data_.misses.push_back(
origin_in_local + options_.max_range() / range * delta);
}
}
}
++num_accumulated_;
if (num_accumulated_ >= options_.num_accumulated_range_data()) {
num_accumulated_ = 0;
transform::Rigid3f current_pose = hits_poses.back();
const sensor::RangeData filtered_range_data = {
current_pose.translation(),
sensor::VoxelFilter(options_.voxel_filter_size())
.Filter(accumulated_range_data_.returns),
sensor::VoxelFilter(options_.voxel_filter_size())
.Filter(accumulated_range_data_.misses)};
return AddAccumulatedRangeData(
time, current_pose.cast<double>(),
sensor::TransformRangeData(filtered_range_data, current_pose.inverse()));
}
return nullptr;
}
std::unique_ptr<LocalTrajectoryBuilder3D::MatchingResult>
LocalTrajectoryBuilder3D::AddAccumulatedRangeData(
const common::Time time,
const transform::Rigid3d& pose_prediction,
const sensor::RangeData& filtered_range_data_in_tracking) {
if (filtered_range_data_in_tracking.returns.empty()) {
LOG(WARNING) << "Dropped empty range data.";
return nullptr;
}
std::shared_ptr<const mapping::Submap3D> matching_submap =
active_submaps_.submaps().front();
transform::Rigid3d initial_ceres_pose =
matching_submap->local_pose().inverse() * pose_prediction;
sensor::AdaptiveVoxelFilter adaptive_voxel_filter(
options_.high_resolution_adaptive_voxel_filter_options());
const sensor::PointCloud high_resolution_point_cloud_in_tracking =
adaptive_voxel_filter.Filter(filtered_range_data_in_tracking.returns);
if (high_resolution_point_cloud_in_tracking.empty()) {
LOG(WARNING) << "Dropped empty high resolution point cloud data.";
return nullptr;
}
if (options_.use_online_correlative_scan_matching()) {
// We take a copy since we use 'initial_ceres_pose' as an output argument.
const transform::Rigid3d initial_pose = initial_ceres_pose;
double score = real_time_correlative_scan_matcher_->Match(
initial_pose, high_resolution_point_cloud_in_tracking,
matching_submap->high_resolution_hybrid_grid(), &initial_ceres_pose);
kRealTimeCorrelativeScanMatcherScoreMetric->Observe(score);
}
transform::Rigid3d pose_observation_in_submap;
ceres::Solver::Summary summary;
sensor::AdaptiveVoxelFilter low_resolution_adaptive_voxel_filter(
options_.low_resolution_adaptive_voxel_filter_options());
const sensor::PointCloud low_resolution_point_cloud_in_tracking =
low_resolution_adaptive_voxel_filter.Filter(
filtered_range_data_in_tracking.returns);
if (low_resolution_point_cloud_in_tracking.empty()) {
LOG(WARNING) << "Dropped empty low resolution point cloud data.";
return nullptr;
}
ceres_scan_matcher_->Match(
(matching_submap->local_pose().inverse() * pose_prediction).translation(),
initial_ceres_pose,
{{&high_resolution_point_cloud_in_tracking,
&matching_submap->high_resolution_hybrid_grid()},
{&low_resolution_point_cloud_in_tracking,
&matching_submap->low_resolution_hybrid_grid()}},
&pose_observation_in_submap, &summary);
kCeresScanMatcherCostMetric->Observe(summary.final_cost);
double residual_distance = (pose_observation_in_submap.translation() -
initial_ceres_pose.translation())
.norm();
kScanMatcherResidualDistanceMetric->Observe(residual_distance);
double residual_angle = pose_observation_in_submap.rotation().angularDistance(
initial_ceres_pose.rotation());
kScanMatcherResidualAngleMetric->Observe(residual_angle);
transform::Rigid3d pose_estimate =
matching_submap->local_pose() * pose_observation_in_submap;
WindowOptimize(pose_estimate, false);
auto opt_pose = PoseFromGtsamNavState(prev_state_);
const Eigen::Quaterniond gravity_alignment = opt_pose.rotation();
sensor::RangeData filtered_range_data_in_local = sensor::TransformRangeData(
filtered_range_data_in_tracking, opt_pose.cast<float>());
std::unique_ptr<InsertionResult> insertion_result = InsertIntoSubmap(
time, filtered_range_data_in_local, filtered_range_data_in_tracking,
high_resolution_point_cloud_in_tracking,
low_resolution_point_cloud_in_tracking, opt_pose, gravity_alignment);
auto duration = std::chrono::steady_clock::now() - accumulation_started_;
kLocalSlamLatencyMetric->Set(
std::chrono::duration_cast<std::chrono::seconds>(duration).count());
return common::make_unique<MatchingResult>(MatchingResult{
time, opt_pose, std::move(filtered_range_data_in_local),
std::move(insertion_result)});
}
void LocalTrajectoryBuilder3D::AddOdometryData(
const sensor::OdometryData& odometry_data) {
if (extrapolator_ == nullptr) {
// Until we've initialized the extrapolator we cannot add odometry data.
LOG(INFO) << "Extrapolator not yet initialized.";
return;
}
extrapolator_->AddOdometryData(odometry_data);
}
std::unique_ptr<LocalTrajectoryBuilder3D::InsertionResult>
LocalTrajectoryBuilder3D::InsertIntoSubmap(
const common::Time time,
const sensor::RangeData& filtered_range_data_in_local,
const sensor::RangeData& filtered_range_data_in_tracking,
const sensor::PointCloud& high_resolution_point_cloud_in_tracking,
const sensor::PointCloud& low_resolution_point_cloud_in_tracking,
const transform::Rigid3d& pose_estimate,
const Eigen::Quaterniond& gravity_alignment) {
if (motion_filter_.IsSimilar(time, pose_estimate)) {
return nullptr;
}
// Querying the active submaps must be done here before calling
// InsertRangeData() since the queried values are valid for next insertion.
std::vector<std::shared_ptr<const mapping::Submap3D>> insertion_submaps;
for (const std::shared_ptr<mapping::Submap3D>& submap :
active_submaps_.submaps()) {
insertion_submaps.push_back(submap);
}
active_submaps_.InsertRangeData(filtered_range_data_in_local,
gravity_alignment);
const Eigen::VectorXf rotational_scan_matcher_histogram =
scan_matching::RotationalScanMatcher::ComputeHistogram(
sensor::TransformPointCloud(
filtered_range_data_in_tracking.returns,
transform::Rigid3f::Rotation(gravity_alignment.cast<float>())),
options_.rotational_histogram_size());
return common::make_unique<InsertionResult>(
InsertionResult{std::make_shared<const mapping::TrajectoryNode::Data>(
mapping::TrajectoryNode::Data{
time,
gravity_alignment,
{}, // 'filtered_point_cloud' is only used in 2D.
high_resolution_point_cloud_in_tracking,
low_resolution_point_cloud_in_tracking,
rotational_scan_matcher_histogram,
pose_estimate}),
std::move(insertion_submaps)});
}
void LocalTrajectoryBuilder3D::RegisterMetrics(
metrics::FamilyFactory* family_factory) {
auto* latency = family_factory->NewGaugeFamily(
"mapping_internal_3d_local_trajectory_builder_latency",
"Duration from first incoming point cloud in accumulation to local slam "
"result");
kLocalSlamLatencyMetric = latency->Add({});
auto score_boundaries = metrics::Histogram::FixedWidth(0.05, 20);
auto* scores = family_factory->NewHistogramFamily(
"mapping_internal_3d_local_trajectory_builder_scores",
"Local scan matcher scores", score_boundaries);
kRealTimeCorrelativeScanMatcherScoreMetric =
scores->Add({{"scan_matcher", "real_time_correlative"}});
auto cost_boundaries = metrics::Histogram::ScaledPowersOf(2, 0.01, 100);
auto* costs = family_factory->NewHistogramFamily(
"mapping_internal_3d_local_trajectory_builder_costs",
"Local scan matcher costs", cost_boundaries);
kCeresScanMatcherCostMetric = costs->Add({{"scan_matcher", "ceres"}});
auto distance_boundaries = metrics::Histogram::ScaledPowersOf(2, 0.01, 10);
auto* residuals = family_factory->NewHistogramFamily(
"mapping_internal_3d_local_trajectory_builder_residuals",
"Local scan matcher residuals", distance_boundaries);
kScanMatcherResidualDistanceMetric =
residuals->Add({{"component", "distance"}});
kScanMatcherResidualAngleMetric = residuals->Add({{"component", "angle"}});
}
/******************************************************************************/
void LocalTrajectoryBuilder3D::FindIdxFromOdomQuene(
const common::Time& time, size_t& former, size_t& later){
//Maybe a simple implementation is ok.
size_t size = predicted_states_.size();
size_t idx = size - 1;
for(size_t i = size - 1; i >=0; --i){
if(common::ToSeconds(predicted_states_.at(i).first - time) < 0){
break;
}else{
idx = i;
}
}
if(idx == (size - 1)){
former = later = idx;
}else if(idx == 0){
former = later = 0;
}else{
former = idx - 1;
later = idx;
}
}
void LocalTrajectoryBuilder3D::ResetGTSAM(){
gtsam::ISAM2Params optParameters;
optParameters.relinearizeThreshold = 0.1;
optParameters.relinearizeSkip = 1;
optimizer_ = gtsam::ISAM2(optParameters);
gtsam::NonlinearFactorGraph new_graph_factors_;
graph_factors_ = new_graph_factors_;
gtsam::Values new_graph_values_;
graph_values_ = new_graph_values_;
}
void LocalTrajectoryBuilder3D::ResetParams(){
done_first_opt_ = false;
gtsam_initialized_ = false;
}
void LocalTrajectoryBuilder3D::WindowOptimize(
const transform::Rigid3d& matched_pose, bool is_drift){
double currentCorrectionTime = common::ToSecondsStamp(time_point_cloud_);
float p_x = matched_pose.translation().x();
float p_y = matched_pose.translation().y();
float p_z = matched_pose.translation().z();
float r_x = matched_pose.rotation().x();
float r_y = matched_pose.rotation().y();
float r_z = matched_pose.rotation().z();
float r_w = matched_pose.rotation().w();
//TODO(wz): check this
bool degenerate = is_drift;
// bool degenerate = (int)odomMsg->pose.covariance[0] == 1 ? true : false;
gtsam::Pose3 lidarPose = gtsam::Pose3(
gtsam::Rot3::Quaternion(r_w, r_x, r_y, r_z), gtsam::Point3(p_x, p_y, p_z));
// 0. initialize system
if (gtsam_initialized_ == false){
ResetGTSAM();
// pop old IMU message
TrimOldImuData(time_point_cloud_);
// initial pose
// prev_state_ and prev_bias_ were initialized in 'InitializeIMU'
prev_pose_ = prev_state_.pose();
gtsam::PriorFactor<gtsam::Pose3> priorPose(
X(0), prev_pose_, prior_pose_noise_);
graph_factors_.add(priorPose);
// initial velocity
prev_vel_ = prev_state_.velocity();
gtsam::PriorFactor<gtsam::Vector3> priorVel(
V(0), prev_vel_, prior_vel_noise_);
graph_factors_.add(priorVel);
// initial bias
gtsam::PriorFactor<gtsam::imuBias::ConstantBias> priorBias(
B(0), prev_bias_, prior_bias_noise_);
graph_factors_.add(priorBias);
// add values
graph_values_.insert(X(0), prev_pose_);
graph_values_.insert(V(0), prev_vel_);
graph_values_.insert(B(0), prev_bias_);
// optimize once
optimizer_.update(graph_factors_, graph_values_);
graph_factors_.resize(0);
graph_values_.clear();
imu_integrator_opt_->resetIntegrationAndSetBias(prev_bias_);
key_ = 1;
gtsam_initialized_ = true;
return;
}
// reset graph for speed
if (key_ == options_.submaps_options().num_range_data()) {
// get updated noise before reset
gtsam::noiseModel::Gaussian::shared_ptr updatedPoseNoise =
gtsam::noiseModel::Gaussian::Covariance(
optimizer_.marginalCovariance(X(key_-1)));
gtsam::noiseModel::Gaussian::shared_ptr updatedVelNoise =
gtsam::noiseModel::Gaussian::Covariance(
optimizer_.marginalCovariance(V(key_-1)));
gtsam::noiseModel::Gaussian::shared_ptr updatedBiasNoise =
gtsam::noiseModel::Gaussian::Covariance(
optimizer_.marginalCovariance(B(key_-1)));
// reset graph
ResetGTSAM();
// add pose
gtsam::PriorFactor<gtsam::Pose3> priorPose(
X(0), prev_pose_, updatedPoseNoise);
graph_factors_.add(priorPose);
// add velocity
gtsam::PriorFactor<gtsam::Vector3> priorVel(
V(0), prev_vel_, updatedVelNoise);
graph_factors_.add(priorVel);
// add bias
gtsam::PriorFactor<gtsam::imuBias::ConstantBias> priorBias(
B(0), prev_bias_, updatedBiasNoise);
graph_factors_.add(priorBias);
// gravity factor for the last pose
if(options_.enable_gravity_factor()){
if(EstimateGravity()){
auto ng_G = g_vec_est_G_.normalized();
gtsam::Unit3 g_z(ng_G[0], ng_G[1], ng_G[2]);
gtsam::Unit3 g_ref_B(0., 0., -1.);
gtsam::Pose3GravityFactor g_factor(
X(0), g_z, prior_gravity_noise_, g_ref_B);
graph_factors_.add(g_factor);
}
}
// add values
graph_values_.insert(X(0), prev_pose_);
graph_values_.insert(V(0), prev_vel_);
graph_values_.insert(B(0), prev_bias_);
// optimize once
optimizer_.update(graph_factors_, graph_values_);
graph_factors_.resize(0);
graph_values_.clear();
key_ = 1;
}
// integrate imu data and optimize（Done in AddImuData）
// add imu factor to graph
const gtsam::PreintegratedImuMeasurements& preint_imu = dynamic_cast<
const gtsam::PreintegratedImuMeasurements&>(*imu_integrator_opt_);
gtsam::ImuFactor imu_factor(
X(key_ - 1), V(key_ - 1), X(key_), V(key_), B(key_ - 1), preint_imu);
graph_factors_.add(imu_factor);
// add imu bias between factor
graph_factors_.add(
gtsam::BetweenFactor<gtsam::imuBias::ConstantBias>(
B(key_ - 1), B(key_), gtsam::imuBias::ConstantBias(),
gtsam::noiseModel::Diagonal::Sigmas(
sqrt(imu_integrator_opt_->deltaTij()) * noise_model_between_bias_)));
// add pose factor
gtsam::Pose3 curPose = lidarPose;
gtsam::PriorFactor<gtsam::Pose3> pose_factor(X(key_), curPose,
degenerate ? correction_noise_2_ : correction_noise_);
graph_factors_.add(pose_factor);
// gravity factor for the last pose
if(options_.enable_gravity_factor()){
if(EstimateGravity() && (key_ - g_est_win_size_) >= 0){
auto ng_G = g_vec_est_G_.normalized();
gtsam::Unit3 g_z(ng_G[0], ng_G[1], ng_G[2]);
gtsam::Unit3 g_ref_B(0., 0., -1.);
gtsam::Pose3GravityFactor g_factor(
X(key_ - g_est_win_size_), g_z, prior_gravity_noise_, g_ref_B);
graph_factors_.add(g_factor);
}
}
// insert predicted values
gtsam::NavState propState_ = imu_integrator_opt_->predict(
prev_state_, prev_bias_);
graph_values_.insert(X(key_), propState_.pose());
graph_values_.insert(V(key_), propState_.v());
graph_values_.insert(B(key_), prev_bias_);
// optimize
optimizer_.update(graph_factors_, graph_values_);
optimizer_.update();
graph_factors_.resize(0);
graph_values_.clear();
// Overwrite the beginning of the preintegration for the next step.
gtsam::Values result = optimizer_.calculateEstimate();
prev_pose_ = result.at<gtsam::Pose3>(X(key_));
prev_vel_ = result.at<gtsam::Vector3>(V(key_));
prev_state_ = gtsam::NavState(prev_pose_, prev_vel_);
prev_bias_ = result.at<gtsam::imuBias::ConstantBias>(B(key_));
// Reset the optimization preintegration object.
imu_integrator_opt_->resetIntegrationAndSetBias(prev_bias_);
TrimOldImuData(time_point_cloud_);
// check optimization
if (FailureDetection(prev_vel_, prev_bias_)){
ResetParams();
return;
}
++key_;
done_first_opt_ = true;
}
transform::Rigid3d LocalTrajectoryBuilder3D::PoseFromGtsamNavState(
const gtsam::NavState& pose_in){
return transform::Rigid3d(pose_in.position(), pose_in.quaternion());
}
void LocalTrajectoryBuilder3D::InterpolatePose(
const double s,
const transform::Rigid3d& relative_transform,
transform::Rigid3d& pose_t){
Eigen::Quaterniond q = Eigen::Quaterniond::Identity().slerp(
s, relative_transform.rotation());
Eigen::Vector3d t = s * relative_transform.translation();
pose_t = transform::Rigid3d(t, q);
}
void LocalTrajectoryBuilder3D::InterpolatePose(
const common::Time& t,
const common::Time& t1, const common::Time& t2,
const gtsam::NavState& pose1, const gtsam::NavState& pose2,
transform::Rigid3d& pose_t){
CHECK_LE(t1, t2)<<"t1 must be smaller than t2!";
double s = common::ToSeconds(t - t1) / common::ToSeconds(t2 - t1);
transform::Rigid3d trans1 = PoseFromGtsamNavState(pose1);
transform::Rigid3d trans2 = PoseFromGtsamNavState(pose2);
transform::Rigid3d relative_transform = trans1.inverse()*trans2;
transform::Rigid3d inter_pose_in_1;
InterpolatePose(s, relative_transform, inter_pose_in_1);
pose_t = trans1 * inter_pose_in_1;
}
bool LocalTrajectoryBuilder3D::FailureDetection(
const gtsam::Vector3& velCur, const gtsam::imuBias::ConstantBias& biasCur){
Eigen::Vector3f vel(velCur.x(), velCur.y(), velCur.z());
if (vel.norm() > 30){
LOG(WARNING)<<"Large velocity, reset IMU-preintegration!";
return true;
}
Eigen::Vector3f ba(biasCur.accelerometer().x(),
biasCur.accelerometer().y(), biasCur.accelerometer().z());
Eigen::Vector3f bg(biasCur.gyroscope().x(),
biasCur.gyroscope().y(), biasCur.gyroscope().z());
if (ba.norm() > 1.0 || bg.norm() > 1.0){
LOG(WARNING)<<"Large bias, reset IMU-preintegration!";
return true;
}
return false;
}
void LocalTrajectoryBuilder3D::TrimStatesCache(const common::Time& time){
while(1){
if(predicted_states_.empty()) break;
if(common::ToSeconds(predicted_states_.front().first - time) < 0){
predicted_states_.pop_front();
}else{
break;
}
}
}
void LocalTrajectoryBuilder3D::TrimOldImuData(const common::Time& time){
while (!imu_que_opt_.empty()){
if (common::ToSeconds(imu_que_opt_.front().time - time) < 0) {
last_imu_time_opt_ = common::ToSecondsStamp(imu_que_opt_.front().time);
imu_que_opt_.pop_front();
}
else
break;
}
}
void LocalTrajectoryBuilder3D::MatchByICP(
pcl::PointCloud<pcl::PointXYZI>::Ptr pre_scan,
pcl::PointCloud<pcl::PointXYZI>::Ptr cur_scan){
// ICP Settings
static pcl::IterativeClosestPoint<pcl::PointXYZI, pcl::PointXYZI> icp;
icp.setMaxCorrespondenceDistance(3);
icp.setMaximumIterations(100);
icp.setTransformationEpsilon(1e-6);
icp.setEuclideanFitnessEpsilon(1e-6);
icp.setRANSACIterations(0);
// Align clouds
auto start = system_clock::now();
icp.setInputSource(cur_scan);
icp.setInputTarget(pre_scan);
pcl::PointCloud<pcl::PointXYZI>::Ptr unused_result(
new pcl::PointCloud<pcl::PointXYZI>());
icp.align(*unused_result);
auto end = system_clock::now();
auto duration = duration_cast<microseconds>(end - start);
LOG(WARNING) << "ICP Cost " << double(duration.count())
* microseconds::period::num / microseconds::period::den << "s.";
LOG(INFO)<<icp.hasConverged()<<", "<<icp.getFitnessScore();
if (icp.hasConverged() == false || icp.getFitnessScore() > 0.3)
return;
// Get pose transformation
float x, y, z, roll, pitch, yaw;
Eigen::Affine3f correctionLidarFrame;
correctionLidarFrame = icp.getFinalTransformation();
float noiseScore = icp.getFitnessScore();
}
void LocalTrajectoryBuilder3D::MatchByNDT(
pcl::PointCloud<pcl::PointXYZI>::Ptr pre_scan,
pcl::PointCloud<pcl::PointXYZI>::Ptr cur_scan,
const Eigen::Matrix4f& init_guess,
Eigen::Matrix3f& R, Eigen::Vector3f& t){
// Filtering input scan to roughly 10% of original size
// to increase speed of registration.
filtered_cloud_pre_.reset(new pcl::PointCloud<pcl::PointXYZI>);
voxel_filter_pre_.setInputCloud(pre_scan);
voxel_filter_pre_.filter(*filtered_cloud_pre_);
auto start = system_clock::now();
filtered_cloud_cur_.reset(new pcl::PointCloud<pcl::PointXYZI>);
voxel_filter_cur_.setInputCloud (cur_scan);
voxel_filter_cur_.filter (*filtered_cloud_cur_);
// Setting point cloud to be aligned.
ndt_.setInputSource(filtered_cloud_cur_);
// Setting point cloud to be aligned to.
ndt_.setInputTarget(filtered_cloud_pre_);
// Calculating required rigid transform to align the input cloud
// to the target cloud.
pcl::PointCloud<pcl::PointXYZI>::Ptr output_cloud(
new pcl::PointCloud<pcl::PointXYZI>);
ndt_.align(*output_cloud, init_guess);
auto trans = ndt_.getFinalTransformation();
R = trans.block(0,0,3,3).cast<float>();
t = trans.block(0,3,3,1).cast<float>();
// lins::Transform transform_to_start;