沉默杀手HomeArchivesTagsCategories
metrics机制
|激光SLAMCartographer源码解读|
Word count: 501|Reading time: 2 min

若要启用metrics机制，必须在launch文件里添加参数-collect_metrics true，这个本质也是gflag

源码的顺序如下

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
// node.cc中
if (collect_metrics)
{
  metrics_registry_ = absl::make_unique<metrics::FamilyFactory>();
  carto::metrics::RegisterAllMetrics(metrics_registry_.get());
}


// register.cc 中定义
void RegisterAllMetrics(FamilyFactory* registry) {
  mapping::constraints::ConstraintBuilder2D::RegisterMetrics(registry);
  mapping::GlobalTrajectoryBuilderRegisterMetrics(registry);
  mapping::LocalTrajectoryBuilder2D::RegisterMetrics(registry);
  mapping::PoseGraph2D::RegisterMetrics(registry);
  sensor::TrajectoryCollator::RegisterMetrics(registry);
  // 不需要可以注释掉
  mapping::constraints::ConstraintBuilder3D::RegisterMetrics(registry);
  mapping::LocalTrajectoryBuilder3D::RegisterMetrics(registry);
  mapping::PoseGraph3D::RegisterMetrics(registry);
}

以上所有RegisterMetrics在local_trajectory_builder_2d.cc, pose_graph_2d.cc, constraint_builder_2d.cc(排除3D情况)定义了所有metrics参数


以kLocalSlamCpuRealTimeRatio为例进行说明， 在local_trajectory_builder_2d.cc开始部分的声明：

1
static auto* kLocalSlamCpuRealTimeRatio = metrics::Gauge::Null();`

然后在LocalTrajectoryBuilder2D::RegisterMetrics中

1
2
3
4
5
auto* cpu_real_time_ratio = family_factory->NewGaugeFamily(
      "mapping_2d_local_trajectory_builder_cpu_real_time_ratio",
      "sensor duration / cpu duration.");

kLocalSlamCpuRealTimeRatio = cpu_real_time_ratio->Add({});

最后是赋值部分

1
2
3
4
5
6
7
8
if (last_thread_cpu_time_seconds_.has_value()) {
    const double thread_cpu_duration_seconds =
        thread_cpu_time_seconds - last_thread_cpu_time_seconds_.value();
    if (sensor_duration.has_value()) {
      kLocalSlamCpuRealTimeRatio->Set(
          common::ToSeconds(sensor_duration.value()) / thread_cpu_duration_seconds );
    }
  }

如果想查看metrics，只能rosservice call /read_metrics "{}"，一次调用的结果部分：

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
name: "mapping_constraints_constraint_builder_2d_constraints"
description: "Constraints computed"
metrics: 
    type: 0
    labels: 
      - 
        key: "matcher"
        value: "searched"
      - 
        key: "search_region"
        value: "global"
    value: 4206.0
    counts_by_bucket: []
  - 
    type: 0
    labels: 
      - 
        key: "matcher"
        value: "found"
      - 
        key: "search_region"
        value: "global"
    value: 0.0
    counts_by_bucket: []

- 
  name: "mapping_2d_pose_graph_constraints"
  description: "Current number of constraints in the pose graph"
  metrics: 
    - 
      type: 1
      labels: 
        - 
          key: "tag"
          value: "inter_submap"
        - 
          key: "trajectory"
          value: "different"
      value: 0.0
      counts_by_bucket: []

name: "mapping_constraints_constraint_builder_2d_num_submap_scan_matchers"
description: "Current number of constructed submap scan matchers"
metrics: 
- 
    type: 1
    labels: []
    value: 18.0
    counts_by_bucket: []
- 
name: "mapping_2d_pose_graph_constraints"
description: "Current number of constraints in the pose graph"
metrics: 
  - 
    type: 1
    labels: 
      - 
        key: "tag"
        value: "inter_submap"
      - 
        key: "trajectory"
        value: "different"
    value: 27.0
    counts_by_bucket: []
  - 
    type: 1
    labels: 
      - 
        key: "tag"
        value: "inter_submap"
      - 
        key: "trajectory"
        value: "same"
    value: 10.0
    counts_by_bucket: []
- 
name: "mapping_constraints_constraint_builder_2d_scores"
description: "Constraint scores built"
metrics: 
  - 
    type: 2
    labels: 
      - 
        key: "search_region"
        value: "local"

获知不同轨迹之间的查找次数和找到的约束数目，源码在ConstraintBuilder2D::ComputeConstraint中的
1
2
kGlobalConstraintsFoundMetric->Increment();
kGlobalConstraintScoresMetric->Observe(score);

模仿源码，可以根据自己的需要添加 metrics

如果不需要3D的metrics，可以到ConstraintBuilder3D::RegisterMetrics, LocalTrajectoryBuilder3D::RegisterMetrics, PoseGraph3D::RegisterMetrics中return