Costmap configuration (static, obstacle, inflation) in ROS - Performance & Optimization

Performance: Costmap configuration (static, obstacle, inflation)

MEDIUM IMPACT

This affects the robot's navigation responsiveness and map update speed, impacting how quickly the robot reacts to obstacles and updates its path.

Configuring costmap layers for obstacle detection and inflation

ROS

static_layer:
  enabled: true
obstacle_layer:
  enabled: true
  observation_sources: laser_scan
  laser_scan: {sensor_frame: laser, data_type: LaserScan, topic: /scan, marking: true, clearing: true}
inflation_layer:
  enabled: true
  inflation_radius: 0.3
  cost_scaling_factor: 5.0
update_frequency: 2.0
publish_frequency: 1.0

Lower update frequency and smaller inflation radius reduce CPU load and speed up map updates without sacrificing obstacle safety.

📈 Performance GainReduces computation frequency by 80%, lowering CPU spikes and improving navigation responsiveness.

Configuring costmap layers for obstacle detection and inflation

ROS

static_layer:
  enabled: true
obstacle_layer:
  enabled: true
  observation_sources: laser_scan
  laser_scan: {sensor_frame: laser, data_type: LaserScan, topic: /scan, marking: true, clearing: true}
inflation_layer:
  enabled: true
  inflation_radius: 0.7
  cost_scaling_factor: 10.0
update_frequency: 10.0
publish_frequency: 5.0

High update frequency with large inflation radius causes frequent heavy computations, leading to navigation delays and CPU spikes.

📉 Performance CostTriggers multiple re-computations per second, causing CPU usage spikes and potential navigation lag.

Performance Comparison

Pattern	CPU Usage	Update Frequency	Inflation Radius	Verdict
High frequency, large inflation	High (80%+ CPU spikes)	10 Hz	0.7 meters	[X] Bad
Moderate frequency, smaller inflation	Moderate (30-40% CPU)	2 Hz	0.3 meters	[OK] Good

Rendering Pipeline

Costmap updates flow through sensor data acquisition, obstacle marking, inflation calculation, and map publishing stages. Each stage affects CPU and memory usage.

→Sensor Data Processing

→Obstacle Marking

→Inflation Calculation

→Map Publishing

⚠️ BottleneckInflation Calculation is most expensive due to expanding obstacle costs over multiple cells.

Optimization Tips

1Keep inflation radius as small as safely possible to reduce CPU load.

2Lower update frequency to balance responsiveness and performance.

3Enable only necessary observation sources to avoid extra processing.

Performance Quiz - 3 Questions

What is the main performance cost of a large inflation radius in costmap configuration?

A Higher memory usage for static map storage

B Slower sensor data acquisition

C Increased CPU usage due to more cells being updated during inflation

D Longer robot battery life

DevTools: ros2 topic echo /costmap_updates and ros2 run rqt_graph rqt_graph

How to check: Monitor the frequency of costmap update messages and observe CPU usage with system monitors while adjusting parameters.

What to look for: Look for stable CPU usage without spikes and update rates matching configured frequencies to confirm efficient costmap operation.

Practice

1. Which costmap layer in ROS is responsible for marking fixed obstacles from a pre-existing map?

easy

A. Obstacle layer

B. Inflation layer

C. Static layer

D. Sensor layer

Solution

1. Step 1: Understand the role of the static layer

   The static layer uses a pre-built map to mark fixed obstacles that do not change over time.

2. Step 2: Compare with other layers

   The obstacle layer updates obstacles dynamically from sensors, and the inflation layer adds safety buffers, so they do not mark fixed obstacles directly.

3. Final Answer:

   Static layer -> Option C

4. Quick Check:

   Fixed obstacles = Static layer [OK]

Hint: Static layer = fixed map obstacles [OK]

Common Mistakes:

• Confusing obstacle layer with static layer
• Thinking inflation layer marks obstacles
• Assuming sensor data is static

2. Which of the following YAML snippets correctly enables the inflation layer in a ROS costmap configuration?

easy

A. inflation:
enable: true
radius: 0.5

B. inflation_layer:
active: yes
radius: 1.0

C. inflation_layer:
enabled: false
inflation_radius: 0.5

D. inflation_layer:
enabled: true
inflation_radius: 0.5

Solution

1. Step 1: Check correct keys for inflation layer

   The correct keys are 'inflation_layer' with 'enabled' and 'inflation_radius' as parameters.

2. Step 2: Validate values and syntax

   inflation_layer: enabled: true inflation_radius: 0.5 uses 'enabled: true' and 'inflation_radius: 0.5', which is correct YAML syntax and valid keys.

3. Final Answer:

   inflation_layer: enabled: true inflation_radius: 0.5 -> Option D

4. Quick Check:

   Correct keys and true enabled = inflation_layer: enabled: true inflation_radius: 0.5 [OK]

Hint: Look for 'enabled: true' and 'inflation_radius' keys [OK]

Common Mistakes:

• Using wrong key names like 'enable' or 'active'
• Setting enabled to false when enabling
• Incorrect indentation or YAML syntax

3. Given this costmap configuration snippet:

obstacle_layer:
  observation_sources: laser_scan_sensor
  laser_scan_sensor:
    data_type: LaserScan
    topic: /scan
    marking: true
    clearing: true
inflation_layer:
  inflation_radius: 0.7
  cost_scaling_factor: 10.0

What effect does setting marking: true and clearing: true have on the obstacle layer?

medium

A. It marks obstacles detected by the laser and clears free space around them

B. It disables obstacle detection and only inflates existing obstacles

C. It marks obstacles but does not clear any free space

D. It clears obstacles but does not mark new ones

Solution

1. Step 1: Understand marking and clearing flags

   Setting 'marking: true' means detected obstacles are added to the costmap. 'clearing: true' means free space detected is cleared from the costmap.

2. Step 2: Apply to laser_scan_sensor

   The laser scan sensor will mark obstacles and clear free space dynamically based on sensor data.

3. Final Answer:

   It marks obstacles detected by the laser and clears free space around them -> Option A

4. Quick Check:

   marking + clearing true = mark and clear obstacles [OK]

Hint: marking=true adds obstacles, clearing=true removes free space [OK]

Common Mistakes:

• Assuming clearing disables marking
• Thinking marking disables clearing
• Confusing cost_scaling_factor with marking

4. You have this partial costmap config:

inflation_layer:
  inflation_radius: 0.6
  cost_scaling_factor: -5.0

What is the problem with this configuration?

medium

A. inflation_radius must be an integer, not a float

B. Negative cost_scaling_factor is invalid and causes inflation errors

C. Missing 'enabled: true' disables inflation layer

D. cost_scaling_factor should be zero to disable inflation

Solution

1. Step 1: Check cost_scaling_factor value

   The cost_scaling_factor controls how quickly costs drop off from obstacles. It must be positive; negative values are invalid.

2. Step 2: Understand impact of negative value

   A negative cost_scaling_factor causes errors or unexpected inflation behavior in the costmap.

3. Final Answer:

   Negative cost_scaling_factor is invalid and causes inflation errors -> Option B

4. Quick Check:

   cost_scaling_factor < 0 = error [OK]

Hint: cost_scaling_factor must be positive [OK]

Common Mistakes:

• Thinking inflation_radius must be integer
• Assuming missing enabled disables inflation
• Believing zero cost_scaling_factor disables inflation

5. You want to configure a ROS costmap that uses a static map, detects obstacles with a laser scanner, and inflates obstacles with a radius of 0.8 meters. Which of the following combined configurations is correct?

hard

A. static_layer:
map_topic: /map
obstacle_layer:
observation_sources: laser_scan
laser_scan:
data_type: LaserScan
topic: /scan
marking: true
clearing: true
inflation_layer:
enabled: true
inflation_radius: 0.8
cost_scaling_factor: 5.0

B. static_layer:
enabled: false
obstacle_layer:
observation_sources: laser
laser:
data_type: PointCloud
topic: /scan
marking: false
clearing: true
inflation_layer:
inflation_radius: 0.8
cost_scaling_factor: -1.0

C. static_layer:
map_topic: /map
obstacle_layer:
observation_sources: laser_scan
laser_scan:
data_type: LaserScan
topic: /scan
marking: false
clearing: false
inflation_layer:
enabled: true
inflation_radius: 1.0
cost_scaling_factor: 10.0

D. static_layer:
map_topic: /map
obstacle_layer:
observation_sources: laser_scan
laser_scan:
data_type: LaserScan
topic: /scan
marking: true
clearing: false
inflation_layer:
enabled: true
inflation_radius: 0.5
cost_scaling_factor: 5.0

Solution

1. Step 1: Verify static layer configuration

   static_layer: map_topic: /map obstacle_layer: observation_sources: laser_scan laser_scan: data_type: LaserScan topic: /scan marking: true clearing: true inflation_layer: enabled: true inflation_radius: 0.8 cost_scaling_factor: 5.0 correctly sets 'static_layer' with 'map_topic: /map' to use the static map.

2. Step 2: Check obstacle layer settings

   static_layer: map_topic: /map obstacle_layer: observation_sources: laser_scan laser_scan: data_type: LaserScan topic: /scan marking: true clearing: true inflation_layer: enabled: true inflation_radius: 0.8 cost_scaling_factor: 5.0 uses 'laser_scan' with correct data_type, topic, and both marking and clearing enabled for proper obstacle detection.

3. Step 3: Confirm inflation layer parameters

   static_layer: map_topic: /map obstacle_layer: observation_sources: laser_scan laser_scan: data_type: LaserScan topic: /scan marking: true clearing: true inflation_layer: enabled: true inflation_radius: 0.8 cost_scaling_factor: 5.0 enables inflation with radius 0.8 and a positive cost_scaling_factor, matching requirements.

4. Final Answer:

   static_layer: map_topic: /map obstacle_layer: observation_sources: laser_scan laser_scan: data_type: LaserScan topic: /scan marking: true clearing: true inflation_layer: enabled: true inflation_radius: 0.8 cost_scaling_factor: 5.0 -> Option A

5. Quick Check:

   All layers correct and enabled = static_layer: map_topic: /map obstacle_layer: observation_sources: laser_scan laser_scan: data_type: LaserScan topic: /scan marking: true clearing: true inflation_layer: enabled: true inflation_radius: 0.8 cost_scaling_factor: 5.0 [OK]

Hint: Check all layers enabled with correct parameters [OK]

Common Mistakes:

• Disabling static layer or obstacle marking
• Using negative cost_scaling_factor
• Setting inflation_radius different from 0.8