Sensor Data Visualization

Understand your robot's perception through real-time data

---

Tutorial Overview

🎯 Learning Objectives

By the end of this tutorial, you will:

• ✅ Visualize LiDAR scans in real-time with RViz

• ✅ View camera images and understand image topics

• ✅ Monitor IMU data and interpret motion

• ✅ Track wheel odometry and observe drift

• ✅ Use rqt tools for multi-sensor visualization

• ✅ Record sensor data for offline analysis

• ✅ Understand coordinate frames and sensor fusion basics

• ✅ Identify sensor issues through visualization

⏱️ Time Required

• Reading & Setup: 15 minutes

• LiDAR Visualization: 20 minutes

• Camera & IMU: 20 minutes

• Odometry Analysis: 15 minutes

• Multi-Sensor Integration: 20 minutes

• Total: ~90 minutes

📚 Prerequisites

• ✅ Completed ROS2 Communication & Tools

• ✅ System setup complete (laptop ↔ robot communication)

• ✅ Robot powered, connected to WiFi

• ✅ Can see robot's topics from laptop

• ✅ RViz installed on laptop

🛠️ What You'll Need

• ✅ Beetlebot (powered, sensors active)

• ✅ Laptop on same network

• ✅ Wireless controller

• ✅ Open space (3m × 3m minimum)

• ✅ Various objects/obstacles for testing

• ✅ Measuring tape (optional, for validation)

---

Part 1: Understanding the Sensor Suite

Sensors on Beetlebot

Your robot has 5 sensor systems:

1. RPLidar C1 - 360° laser scanner

   • Range: 12m max, 8-10m typical indoors

   • Update rate: 10 Hz

   • Topic: /scan

2. Raspberry Pi Camera V1.3 - Visual perception

   • Resolution: 5MP (2592×1944), typically 1080p/720p

   • Update rate: 30 fps

   • Topics: /pi_camera/image_raw, /pi_camera/camera_info

3. LSM6DSRTR IMU - Motion sensing

   • 6-axis: 3-axis accelerometer + 3-axis gyroscope

   • Update rate: 100 Hz

   • Topics: /imu/data, /imu/data_raw

4. Wheel Encoders (4×) - Position feedback

   • Resolution: 1800 ticks/revolution

   • Update rate: 20 Hz

   • Topics: /wheel_rpm, /wheel_ticks

5. Battery Monitor - Power management

   • Voltage monitoring via ADC

   • Update rate: 10 Hz

   • Topic: /battery_voltage

---

Sensor Coordinate Frames

Every sensor has a position and orientation relative to the robot:

[PLACEHOLDER: Diagram showing sensor positions on robot]

Frame hierarchy:

odom (world frame)
  └─ base_link (robot center)
      ├─ lidar_link (8.5cm forward, 20.6cm up)
      ├─ camera_link (15cm forward, 8.8cm up)
      ├─ camera_optical_link (camera's view frame)
      └─ imu_link (center, 2cm down)

Why frames matter:

• Sensor data is reported in sensor's own frame

• Must transform to common frame for fusion

• RViz needs correct frames to display properly

---

Part 2: LiDAR Visualization

Launch Robot with LiDAR

On robot (via SSH):

# Check if LiDAR already running (auto-launch)
ros2 topic list | grep scan

# If /scan exists, LiDAR is already active
# If not, launch manually:
ros2 launch lyra_bringup lidar.launch.py

Verify LiDAR data:

# Check topic rate
ros2 topic hz /scan

# Should show:
# average rate: 10.xxx Hz

# Quick peek at data
ros2 topic echo /scan --once

---

Visualize LiDAR in RViz

On laptop:

# Launch RViz
rviz2

Configure RViz for LiDAR:

1. Set Fixed Frame:

   • Left panel: Global Options

   • Fixed Frame: Change "map" → "base_link"

2. Add LaserScan Display:

   • Bottom left: Click "Add"

   • By topic → /scan → LaserScan

   • Click OK

3. Configure LaserScan Display:

   • Expand LaserScan in left panel

   • Size (m): 0.05 (point size)

   • Style: Flat Squares (easier to see)

   • Color Transformer: Intensity or FlatColor

   • If Intensity: Color scheme = Rainbow

[PLACEHOLDER: Screenshot of RViz showing LiDAR scan]

What you see:

• Red/orange dots = close objects (< 1m)

• Green dots = medium distance (1-3m)

• Blue dots = far objects (3m+)

• Robot center at origin (coordinate axes)

---

Exercise 4.1: LiDAR Exploration

Task: Understand LiDAR capabilities

Step 1: Range Test

1. Place robot in clear area
2. Hold flat object (book, clipboard) at various distances
3. Move slowly toward robot from 5m away
4. Watch RViz - when do points appear?

Expected: Points appear around 8-10m indoors, 12m in ideal conditions

Step 2: Resolution Test

1. Place two thin objects (pens) 5cm apart at 2m distance
2. Can LiDAR distinguish them as separate objects?
3. Move to 5m - still separate?

Expected: At 2m, should see gap. At 5m, might merge into one blob.

Step 3: Material Test

1. Test different materials at 1m distance:
   - Flat cardboard (strong return)
   - Mirror/glass (weak/no return)
   - Black fabric (weak return)
   - White wall (strong return)

Expected: Shiny/dark/transparent surfaces give poor returns

What you learned:

• LiDAR range limitations

• Angular resolution (~0.5°)

• Material reflectivity affects detection

---

Advanced LiDAR Visualization

Add more information to RViz:

1. Robot Model:

Add → RobotModel
- Description Topic: /robot_description
- Now see robot chassis + LiDAR position

2. TF Frames:

Add → TF
- Shows coordinate frames as RGB axes
- See lidar_link position relative to base_link

3. LaserScan with Decay:

LaserScan settings:
- Decay Time: 5 seconds
- Shows last 5 seconds of scans (motion trail)
- Useful for seeing paths through environment

[PLACEHOLDER: Screenshot showing robot model + LiDAR + TF frames]

---

Exercise 4.2: Obstacle Detection

Task: Map your surroundings

Setup:

1. Place robot in center of room
2. Configure RViz:
   - Fixed Frame: "odom"
   - Add Odometry display (/odom)
   - Add LaserScan (/scan)
   - LaserScan Decay Time: 30 seconds

Action:

1. Drive robot in complete circle (360°)
2. Drive slowly (0.2 m/s)
3. Watch scan build up in RViz
4. Complete 2-3 circles for good coverage

Observe:

• Walls appear as solid lines

• Furniture shows as distinct blobs

• Doorways appear as gaps

• Moving objects (people) may appear inconsistent

Analysis Questions:

1. Can you identify room shape from scans?

2. Are corners clearly defined?

3. Any areas with no returns? (windows, mirrors)

---

Part 3: Camera Visualization

Camera Topics Overview

Available topics:

ros2 topic list | grep camera

# Output:
# /pi_camera/camera_info
# /pi_camera/image_raw
# /pi_camera/image_raw/compressed

Topic explanations:

/pi_camera/image_raw (sensor_msgs/Image)

• Uncompressed image data

• Large bandwidth (~30 MB/s at 30fps)

• Best quality

• Use on local network

/pi_camera/image_raw/compressed (sensor_msgs/CompressedImage)

• JPEG compressed

• Smaller bandwidth (~3 MB/s)

• Slight quality loss

• Better for remote viewing

/pi_camera/camera_info (sensor_msgs/CameraInfo)

• Camera calibration data

• Lens distortion parameters

• Used for 3D perception

---

View Camera Feed

Method 1: rqt_image_view (Simple)

# On laptop
ros2 run rqt_image_view rqt_image_view

# Window opens
# Dropdown: Select "/pi_camera/image_raw"
# Image appears!

[PLACEHOLDER: Screenshot of camera view in rqt_image_view]

---

Method 2: RViz (Integrated)

# In RViz (already running)
# Add → Camera
# Image Topic: /pi_camera/image_raw
# Camera Info Topic: /pi_camera/camera_info

# New window shows camera view

Benefits of RViz method:

• See camera + LiDAR simultaneously

• Overlay detection results

• 3D context visible

---

Exercise 4.3: Camera Exploration

Task: Understand camera capabilities

Test 1: Field of View

1. Place robot facing wall at 1m distance
2. Measure width of wall visible in camera
3. Calculate horizontal FOV

Expected: ~54° horizontal FOV
Math: FOV = 2 × arctan(width / (2 × distance))

Test 2: Focus Range

1. Hold object at various distances:
   - 10cm (very close)
   - 30cm (near)
   - 1m (medium)
   - 3m+ (far)
2. Which distances are in focus?

Expected: Best focus 30cm - 3m (Pi Camera V1.3 fixed focus)

Test 3: Lighting Conditions

1. Test in bright room (good)
2. Test in dim lighting (auto-adjusts exposure)
3. Test with backlight (camera facing window)
4. Test with point light source (flashlight)

Note effects on image quality

---

Image Topic Performance

Check image publishing rate:

# Should be 30 fps
ros2 topic hz /pi_camera/image_raw

# If much lower, check:
# 1. Network bandwidth
# 2. Compression settings
# 3. Camera configuration

Bandwidth usage:

# Check topic bandwidth
ros2 topic bw /pi_camera/image_raw

# Typical: 30-40 MB/s (uncompressed)
# Compressed: 2-5 MB/s

Use compressed for remote viewing!

---

Part 4: IMU Data Visualization

IMU Topics

Available topics:

ros2 topic list | grep imu

# Output:
# /imu/data          ← Filtered, calibrated data
# /imu/data_raw      ← Raw sensor readings

Difference:

• /imu/data_raw: Direct from sensor (noisy)

• /imu/data: After filtering and calibration (smoother)

Message contents:

orientation: [x, y, z, w]        ← Quaternion (not directly usable yet)
angular_velocity: [x, y, z]      ← Rotation rates (rad/s)
linear_acceleration: [x, y, z]   ← Accelerations (m/s²)

---

Visualize IMU in RViz

Add IMU display:

# In RViz
# Add → Imu
# Topic: /imu/data

What you see:

• Arrow shows acceleration vector

• Box shows orientation

• Color indicates magnitude

[PLACEHOLDER: Screenshot of IMU visualization in RViz]

---

Exercise 4.4: IMU Experimentation

Task: Understand IMU measurements

Test 1: Static Robot

1. Robot sitting still on level ground
2. Echo IMU data:

ros2 topic echo /imu/data --field linear_acceleration

# Expected:
# x: ~0.0 (no forward acceleration)
# y: ~0.0 (no sideways acceleration)
# z: ~9.81 (gravity pulling down!)

# Even stationary, accelerometer reads gravity

Test 2: Forward Acceleration

1. ARM robot, drive forward gently
2. Watch linear_acceleration.x
3. Should spike positive during acceleration
4. Return to ~0 at constant speed
5. Spike negative during braking

Test 3: Rotation

1. Spin robot in place (right stick)
2. Watch angular_velocity.z
3. Positive = counterclockwise
4. Negative = clockwise
5. Magnitude = rotation speed (rad/s)

---

Plotting IMU Data with rqt_plot

Real-time graphing:

# Launch rqt_plot
rqt_plot

# Add topics to plot:
# /imu/data/angular_velocity/z
# /imu/data/linear_acceleration/x

# Drive robot - see live graphs!

[PLACEHOLDER: Screenshot of rqt_plot showing IMU data]

Try these patterns:

• Forward → Stop (see accel spike)

• Turn left → Turn right (see angular velocity)

• Figure-8 pattern (see both signals)

---

Understanding IMU Noise

Compare raw vs filtered:

# Terminal 1: Raw IMU
ros2 topic echo /imu/data_raw --field angular_velocity

# Terminal 2: Filtered IMU
ros2 topic echo /imu/data --field angular_velocity

# Notice: Raw data is noisy, filtered is smoother

Why filter?

• Raw sensor has electrical noise

• Quantization errors

• Temperature drift

• Vibrations from motors

Filtering trade-off:

• Smoother signal (good for control)

• Slight delay (bad for fast response)

Next tutorial (IMU Signal Processing) covers filtering in depth!

---

Part 5: Wheel Odometry Visualization

Odometry Topics

Available topics:

ros2 topic list | grep odom

# Output:
# /odom                    ← Wheel-only odometry
# /wheel/odom             ← Alternative name (same data)
# /odometry/filtered      ← Fused odom (wheel + IMU via EKF)

Key difference:

• /odom: Based only on wheel encoders

• /odometry/filtered: Wheel encoders + IMU fusion (better!)

---

Visualize Odometry

Add to RViz:

# Already in RViz
# Add → Odometry
# Topic: /odometry/filtered

Configuration:

Odometry Display settings:
- Covariance: Keep Last 50
- Position Tolerance: 0.1
- Angle Tolerance: 0.1
- Color: 25, 255, 0 (green)

What you see:

• Green arrow = robot's pose (position + orientation)

• Ellipse = uncertainty covariance

• Trail = path history

[PLACEHOLDER: Screenshot of odometry trail in RViz]

---

Exercise 4.5: Odometry Drift Observation

Task: Measure odometry accuracy

Test 1: Straight Line

Setup:
1. Mark starting position with tape
2. Measure exactly 2 meters ahead
3. Mark target position

Procedure:
1. Reset odometry (power cycle robot)
2. Record starting odom:
   ros2 topic echo /odom --once
3. Drive exactly to 2m mark (use tape as guide)
4. Record ending odom
5. Compare:
   - Odometry says: X meters
   - Reality is: 2.000 meters
   - Error: ____%

Expected error: ±3-5% (1.94-2.06m reported)

Test 2: Square Path

Setup:
1. Mark 1m × 1m square on floor with tape

Procedure:
1. Start at corner
2. Drive square: 1m forward, turn 90° left, repeat 4×
3. Should return to start
4. Measure final position error

Expected: upto 50cm drift after 4m + 4 turns
Why drift? Wheel slip during turns, encoder quantization

Test 3: Rotation

Procedure:
1. Note starting orientation (use compass or fixed reference)
2. Spin 360° in place (count full rotations)
3. Check final orientation

Expected: ±5-20° error after full rotation
Why? Wheelbase calibration, floor friction variations

---

Comparing Wheel vs Fused Odometry

Plot both simultaneously:

# Terminal 1
rqt_plot /odom/pose/pose/position/x /odometry/filtered/pose/pose/position/x

# Drive robot forward slowly
# See how trajectories compare

Typically observe:

• /odometry/filtered is smoother (IMU helps)

• /odom may have small jumps (encoder quantization)

• Both drift over time, but fused drifts less

---

Part 6: Multi-Sensor Visualization

The Complete Perception Picture

Configure RViz with all sensors:

Displays to add:

1. ✅ RobotModel (/robot_description)

2. ✅ LaserScan (/scan)

3. ✅ Camera (/pi_camera/image_raw)

4. ✅ Imu (/imu/data)

5. ✅ Odometry (/odometry/filtered)

6. ✅ TF (coordinate frames)

Save configuration:

File → Save Config As
Save to: ~/beetlebot_full.rviz

# Next time:
rviz2 -d ~/beetlebot_full.rviz

[PLACEHOLDER: Screenshot of RViz with all sensors displayed]

---

Exercise 4.6: Multi-Sensor Correlation

Task: See how sensors relate during motion

Setup:

1. Launch full RViz config (all sensors)
2. Position robot facing open space with obstacles
3. Arrange RViz windows:
   - Main 3D view: Large, showing LiDAR + robot
   - Camera view: Smaller, top-right corner
   - IMU: Visible in 3D view

Test Scenario:

1. Drive toward obstacle
2. Observe simultaneously:
   - LiDAR: Points getting closer (red/orange)
   - Camera: Object getting larger
   - Odometry: Position advancing
   - IMU: Acceleration spike when starting

3. Stop before obstacle
4. Turn 90° left
5. Observe:
   - LiDAR: Scan rotates (walls move)
   - Camera: View rotates
   - IMU: angular_velocity.z spikes
   - Odometry: Orientation changes

6. Drive parallel to wall
7. Observe:
   - LiDAR: Wall appears as flat line at constant distance
   - Camera: May or may not see wall (depends on position)
   - Odometry: Smooth trajectory

Questions to answer:

1. Do all sensors agree on motion?

2. Which sensor updates fastest? (IMU: 100Hz)

3. Which is most accurate for distance? (LiDAR)

4. Which gives most information? (Camera: colors, textures)

---

Part 7: Recording Sensor Data

Why Record Data?

Use cases:

• Debugging: Replay issues offline

• Analysis: Process data in Python/MATLAB

• Sharing: Send to colleagues/support

• Comparison: Before/after calibration

• Development: Train ML models

• Documentation: Demonstrate capabilities

---

Recording with ros2 bag

Record all sensors:

# Record everything
ros2 bag record -a -o full_sensor_test

# Or record specific topics (smaller file):
ros2 bag record -o sensor_test \
  /scan \
  /pi_camera/image_raw/compressed \
  /imu/data \
  /odometry/filtered \
  /wheel_rpm \
  /battery_voltage

While recording:

• Drive predetermined path (square, circle)

• Test various speeds

• Include static periods

• Document what you're testing

Stop recording: Ctrl+C

---

Analyzing Recorded Data

Get bag info:

ros2 bag info sensor_test

# Shows:
# Duration: 45.3s
# Start: Jan 05 2026 15:23:10
# End: Jan 05 2026 15:23:55
# Messages: 8452
# Topics:
#   /scan: 453 messages (10 Hz)
#   /imu/data: 4530 messages (100 Hz)
#   ...

Play back data:

# Play at normal speed
ros2 bag play sensor_test

# Play at half speed (better for observation)
ros2 bag play sensor_test --rate 0.5

# Play specific time range
ros2 bag play sensor_test --start-offset 10 --duration 15

While playing back:

• RViz visualizes as if robot was live

• Can pause, rewind (by replaying)

• No robot needed!

---

Exercise 4.7: Data Collection Project

Task: Record and analyze driving patterns

Collect data:

# Test 1: Straight line at 0.3 m/s
ros2 bag record -o straight_slow /odom /wheel_rpm /imu/data
# Drive 3 meters straight
# Stop recording

# Test 2: Straight line at 0.8 m/s (faster)
ros2 bag record -o straight_fast /odom /wheel_rpm /imu/data
# Drive 3 meters straight (faster)
# Stop recording

# Test 3: Figure-8 pattern
ros2 bag record -o figure8 /odom /wheel_rpm /imu/data
# Drive smooth figure-8
# Stop recording

Analyze:

# Play back and compare in RViz
ros2 bag play straight_slow
# Note final position error

ros2 bag play straight_fast
# Compare error to slow speed

ros2 bag play figure8
# Observe complex trajectory

# Questions:
# 1. Which test had most odometry drift?
# 2. Did faster speed increase error?
# 3. How smooth were the curves in figure-8?

---

Part 8: Diagnostic Tools

rqt_multiplot - Advanced Plotting

Install (if not already):

sudo apt install ros-jazzy-rqt-multiplot

Launch:

rqt_multiplot

Create multi-panel plots:

Panel 1: IMU angular velocity (all axes)
- /imu/data/angular_velocity/x
- /imu/data/angular_velocity/y
- /imu/data/angular_velocity/z

Panel 2: Wheel speeds
- /wheel_rpm (all 4 wheels if available)

Panel 3: Battery voltage
- /battery_voltage

[PLACEHOLDER: Screenshot of rqt_multiplot with 3 panels]

Use case:

• Monitor all critical sensors at once

• Spot correlations between signals

• Detect anomalies

---

rqt_robot_monitor - System Health

rqt_robot_monitor

Shows:

• Battery status

• Motor health

• Sensor connectivity

• System warnings

Set up alerts:

• Low battery warning (<10.5V)

• Sensor timeout warnings

• Motor fault detection

---

Custom Monitoring Script

Create battery alert script:

nano ~/sensor_monitor.sh

Script content:

#!/bin/bash

while true; do
  # Battery
  BAT=$(ros2 topic echo /battery_voltage --once 2>/dev/null)
  
  # LiDAR rate
  LIDAR=$(ros2 topic hz /scan --window 10 2>/dev/null | grep average | awk '{print $3}')
  
  # IMU rate
  IMU=$(ros2 topic hz /imu/data --window 10 2>/dev/null | grep average | awk '{print $3}')
  
  echo "=== Sensor Status ==="
  echo "Battery: ${BAT}V"
  echo "LiDAR: ${LIDAR} Hz"
  echo "IMU: ${IMU} Hz"
  
  # Alert if any issues
  if (( $(echo "$BAT < 10.5" | bc -l) )); then
    echo "⚠️  LOW BATTERY!"
  fi
  
  if (( $(echo "$LIDAR < 8" | bc -l) )); then
    echo "⚠️  LiDAR rate low!"
  fi
  
  sleep 5
done

Run:

chmod +x ~/sensor_monitor.sh
./sensor_monitor.sh

---

Part 9: Troubleshooting Sensor Issues

LiDAR Not Publishing

Symptoms: /scan topic not appearing

Debug steps:

# 1. Check if LiDAR node running
ros2 node list | grep rplidar

# 2. Check USB connection
lsusb | grep CP2102
# Should show Silicon Labs CP2102 (LiDAR USB chip)

# 3. Check permissions
ls -l /dev/ttyUSB*
# Should show device exists

# 4. Restart LiDAR node
ros2 lifecycle set /rplidar_node configure
ros2 lifecycle set /rplidar_node activate

If still not working: ⚡ Power cycle robot

---

Camera Not Streaming

Symptoms: No image in rqt_image_view

Debug steps:

# 1. Check if camera detected
ls /dev/video*
# Should show /dev/video0

# 2. Test with v4l2
v4l2-ctl --list-devices

# 3. Check topic publishing
ros2 topic hz /pi_camera/image_raw

# 4. Check bandwidth (may be too slow over WiFi)
ros2 topic bw /pi_camera/image_raw

# Try compressed:
ros2 run rqt_image_view rqt_image_view /pi_camera/image_raw/compressed

If still not working: ⚡ Power cycle robot

---

IMU Data Frozen

Symptoms: IMU values not changing

Debug steps:

# 1. Check topic rate
ros2 topic hz /imu/data
# Should be ~100 Hz

# 2. Check for errors
ros2 topic echo /rosout | grep -i imu

# 3. Check I2C connection (on robot via SSH)
sudo i2cdetect -y 1
# Should show device at address 0x6B

If still not working: ⚡ Power cycle robot

---

Odometry Looks Wrong

Symptoms: Robot drifting wildly in RViz

Check:

# 1. Verify encoders working
ros2 topic echo /wheel_ticks

# Drive robot - ticks should increase

# 2. Check for reversed wheels
# Left wheels should have negative ticks going forward
# Right wheels should have positive ticks going forward

# 3. Verify wheel parameters
ros2 param get /lyra_node wheel_radius
ros2 param get /lyra_node wheelbase

# Should match: 0.065m radius, 0.295m wheelbase

If still not working: ⚡ Power cycle robot

---

General Sensor Troubleshooting

Most sensor issues fixed by:

1. ⚡ Power cycle robot (90% of issues)

2. Check topic rates (ros2 topic hz)

3. Check for error messages (ros2 topic echo /rosout)

4. Verify nodes running (ros2 node list)

5. Check network connectivity (WiFi signal strength)

---

Part 10: Knowledge Check

Concept Quiz

1. Which sensor has the highest update rate?

2. Why does odometry drift over time?

3. What's the difference between /imu/data and /imu/data_raw?

4. Why use compressed images instead of raw?

5. Which sensor is best for measuring distance to obstacles?

---

Hands-On Challenge

Task: Create sensor validation report

Requirements:

1. Record 2-minute rosbag with all sensors

2. Play back and screenshot each sensor view

3. Create report with:

   • LiDAR: Max range observed

   • Camera: Field of view measurement

   • IMU: Acceleration range during maneuvers

   • Odometry: Final drift after 5m × 5m square

Deliverable:

• Rosbag file

• Screenshots (4-5 images)

• Text report with measurements

---

Part 11: What You've Learned

✅ Congratulations!

You now understand:

Sensor Capabilities:

• ✅ LiDAR: Range, resolution, materials

• ✅ Camera: FOV, focus, lighting effects

• ✅ IMU: Acceleration, rotation sensing

• ✅ Odometry: Position tracking and drift

Visualization Tools:

• ✅ RViz configuration for all sensors

• ✅ rqt_image_view for camera

• ✅ rqt_plot for time-series data

• ✅ Multi-sensor displays

Data Management:

• ✅ Recording rosbags

• ✅ Playing back data

• ✅ Offline analysis

• ✅ Sharing sensor data

Troubleshooting:

• ✅ Diagnosing sensor issues

• ✅ Checking topic health

• ✅ Common problems and solutions

---

Next Steps

🎯 You're Now Ready For:

Immediate Next: → IMU Signal Processing - Filter noisy IMU data, extract orientation

Sensor Applications: → Teleoperation Control - Use sensors for feedback → SLAM Mapping - Build maps with LiDAR → Autonomous Navigation - Sensor-based obstacle avoidance

Advanced Topics:

• Sensor calibration

• Multi-sensor fusion

• Custom sensor integration

---

Quick Reference

Essential Visualization Commands

# --- RViz ---
rviz2                                    # Launch
rviz2 -d config.rviz                    # Load config

# --- Camera ---
ros2 run rqt_image_view rqt_image_view  # View images
# Topic: /pi_camera/image_raw/compressed (use compressed!)

# --- Plotting ---
rqt_plot /imu/data/angular_velocity/z   # Plot single topic
rqt_multiplot                            # Advanced plotting

# --- Recording ---
ros2 bag record -a                       # Record all
ros2 bag record /scan /odom /imu/data   # Record specific
ros2 bag play bagfile                    # Play back
ros2 bag info bagfile                    # Get info

# --- Topic Monitoring ---
ros2 topic hz /scan                      # Check rate
ros2 topic bw /pi_camera/image_raw      # Check bandwidth
ros2 topic echo /imu/data --once        # Quick peek

# --- Troubleshooting ---
ros2 node list                           # Check nodes
ros2 topic list                          # Check topics

---

Typical Sensor Rates

Sensor	Topic	Rate	Bandwidth
LiDAR	/scan	10 Hz	~50 KB/s
Camera (raw)	/pi_camera/image_raw	30 Hz	30-40 MB/s
Camera (compressed)	/pi_camera/image_raw/compressed	30 Hz	2-5 MB/s
IMU	/imu/data	100 Hz	~10 KB/s
Odometry	/odometry/filtered	20 Hz	~5 KB/s
Encoders	/wheel_rpm	20 Hz	~2 KB/s
Battery	/battery_voltage	10 Hz	<1 KB/s

---

Completed Sensor Data Visualization! 🎉

→ Continue to IMU Signal Processing → Or return to Tutorial Index

---

Last Updated: January 2026 Tutorial 4 of 11 - Intermediate Level Estimated completion time: 90 minutes