Setup your Radar

A professional traffic logger with weatherproof infrastructure deployment that keeps data local, requires no cameras, and helps you advocate for safer streets.

Difficulty: Intermediate • Time: 4-6 hours • Cost: $563-596

In this guide: Parts List • Build Steps • Generate Reports • Troubleshooting

Introduction #

Measuring vehicle speeds is the first step toward safer streets. Without data, convincing city officials to address speeding can be challenging.

This guid will show you how to build your own privacy-first traffic radar using open-source software and off-the-shelf Doppler technology (the same sensors municipalaties use). No cameras, no license plates, just local speed data that produces professional traffic reports.

This weatherproof infrastructure deployment gives community advocates, parents, and civic-minded makers the evidence they need to drive change, with professional-grade hardware designed for permanent outdoor installations.

Who This Guide Is For #

• Community advocates: Get professional data for traffic calming proposals
• Parents: Prove speeding near schools with evidence, not emotion
• Data enthusiasts: Build useful civic tech with open hardware
• Local officials: Validate commercial traffic studies with independent data

Not sure? This project takes 4-6 hours and costs $563-596. If you care about street safety and need a permanent monitoring solution, you’ll find it worthwhile.

Before You Begin #

Skills required:

• Basic Linux command line (SSH, file editing, navigation)
• Basic hardware assembly (connecting cables, mounting)
• Patience for troubleshooting (sensor configuration can be finicky)

Tools needed:

• Computer for flashing SD card and SSH access
• Screwdrivers for assembly
• Optional: Multimeter for troubleshooting connections

No soldering required • No coding required • No prior radar experience needed

Privacy & Legal Considerations #

What This System Measures #

• ✅ Collected: Vehicle speed, direction, timestamp
• ❌ Not collected: License plates, vehicle photos, driver identity
• ❌ Not transmitted: All data stays on your device

Is This Legal? #

In most jurisdictions, yes. You’re measuring public behaviour on public streets, similar to what traffic engineers and academic researchers do.

Generally allowed:

• Monitoring streets visible from your property
• Temporary studies (1-4 weeks) for community advocacy
• Presenting findings to local government
• Sharing aggregate statistics (PDF reports)

May require permission:

• Mounting on utility poles (contact utility company)
• Long-term installations (>1 month)
• School zones or government property

Not allowed:

• Monitoring private property
• Selling data commercially
• Creating safety hazards

Disclaimer: Laws vary. When in doubt, consult local authorities or an attorney.

Understanding Speed: The Physics Behind Street Safety #

Speed isn’t just a number on a sign: it’s physics! And physics always wins.

The kinetic energy of a moving object follows this formula:

$$K_E = \frac{1}{2} m v^2$$

Where $m$ is mass and $v$ is velocity. The key insight: energy scales with the square of velocity.

Real-world impact:

• A 3,000 lb sedan at 40 mph carries four times the crash energy of the same car at 20 mph
• At 50 mph, that energy jumps to 6.25 times what it was at 20 mph
• Even a 5 mph difference (say, 30 mph vs 35 mph) increases crash energy by 36%

For anyone outside the vehicle (pedestrians, cyclists, kids) this exponential relationship is the difference between walking away and never walking again.

Streets designed for 25 mph but driven at 40? That’s not just a little faster: it’s 2.56× the destructive force on impact.

Your radar measures what matters: actual speeds, not posted limits. You’ll capture the real behaviour, quantify the risk, and have data that speaks louder than feelings.

What You’ll Build #

This guide walks you through building a professional, weatherproof traffic monitoring system:

• A Raspberry Pi radar logger that captures vehicle speeds via Doppler radar
• A SQLite database that stores detections locally (no cloud)
• A live web dashboard with real-time speeds, histograms, and time-of-day patterns
• Professional PDF reports with traffic engineering metrics (p50, p85, p98)
• Weatherproof infrastructure designed for permanent outdoor deployment

Privacy by design: No cameras, license plates, or identifying information—just velocity measurements.

[PLACEHOLDER: Image showing completed infrastructure deployment (weatherproof enclosure mounted on utility pole)]

Parts and Tools List #

New to radar sensors? The OmniPreSense OPS7243-A-CW-R2 is recommended for infrastructure deployment. It’s weatherproof (IP67), has 100m range, and handles outdoor conditions reliably.

Bill of Materials #

Tools Required #

• Basic screwdrivers, drill, adhesive
• Computer for flashing SD card and SSH access
• 5/16" nut driver (for steel bands)
• Optional: Multimeter for testing connections

Step-by-Step Build Guide #

Build overview (total time: 4-6 hours):

1. Connect Sensor to Raspberry Pi — 10-15 minutes
2. Configure Sensor Output Mode — 10 minutes
3. Verify Data Stream — 5 minutes
4. Install Software — 30-60 minutes
5. Access the Web Dashboard — 5 minutes
6. Mount the Radar Sensor — 15-30 minutes
7. Generate PDF Reports — After data collection

Step 1: Connect the Sensor to the Raspberry Pi #

Estimated time: 30-45 minutes

The OPS7243-A-CW-R2 sensor (RS232 interface, designated R2) requires a serial HAT for connection to the Raspberry Pi 4.

1. Flash Raspberry Pi OS to your microSD card:

   • Use Raspberry Pi Imager
   • Choose “Raspberry Pi OS Lite” (64-bit recommended)
   • Configure WiFi and SSH in advanced settings before writing

2. Install serial HAT on Raspberry Pi 4:

   • Power off Pi completely
   • Attach Waveshare RS232/485 HAT to 40-pin GPIO header
   • Ensure all pins are aligned and fully seated

3. Wire sensor to HAT:

[PLACEHOLDER: Diagram showing RS232 wiring connections between OPS7243 sensor and Waveshare HAT, with colour-coded wires and pin labels]

Critical: RS232 uses RX↔TX crossover. Sensor TX connects to HAT RX, and vice versa.

4. Configure Pi serial port:

# Disable serial console to enable serial for sensor
sudo raspi-config
# Navigate to: Interface Options → Serial Port
# - "Login shell over serial?" → NO
# - "Serial port hardware enabled?" → YES

5. Enable serial HAT (add to /boot/config.txt):

# Edit boot configuration
sudo nano /boot/config.txt

# Add these lines at the end:
dtoverlay=uart0
enable_uart=1

6. Reboot and verify:

# Reboot to apply changes
sudo reboot

# After reboot, verify serial device exists
ls -l /dev/serial0
# Should show link to ttyAMA0 or ttyS0

Success criteria: /dev/serial0 exists and links to a serial device

Power considerations:

• RS232 sensor typical draw: 300-440mA at 5V (~2.2W)
• Power from dedicated supply (not Pi’s 5V pin) for stability
• Use low-voltage disconnect if running on battery/solar

Platform note: These commands are for Linux. On macOS, use stty -f instead of stty -F

Step 2: Configure Sensor Output Mode #

Estimated time: 10 minutes

The OmniPreSense OPS243 sensor ships with CSV output by default, but this software expects JSON output.

1. Connect via serial terminal:

   # Set serial port parameters
   stty -F /dev/ttyUSB0 19200 cs8 -parenb -cstopb
   
   # Connect to sensor (press Ctrl+A then K to exit)
   screen /dev/ttyUSB0 19200

   Note: On macOS, use stty -f instead of stty -F

2. Configure sensor (type these commands in the terminal):

   OJ    # Enable JSON output mode
   UM    # Set units to Meters per second
   OM    # Enable magnitude reporting
   A!    # Save settings permanently
   

   Press Enter after each command. The sensor should respond with O.

3. Verify you see JSON output (lines with curly braces {}):

   { "magnitude": 1.2, "speed": 3.4 }

[PLACEHOLDER: Screenshot of terminal showing sensor configuration commands and JSON output verification]

Success criteria: You see JSON output (not CSV) when vehicles pass

If it’s not working:

• Still seeing CSV? → Type OJ again
• No response? → Verify baud rate is 19200
• Garbled output? → Check serial port settings

Need more sensor info? Type ?? to see module information or ?V for firmware version.

Full documentation: OmniPreSense Support

Step 3: Verify Data Stream #

Estimated time: 5 minutes

Confirm the sensor is streaming data correctly:

# View raw sensor output
screen /dev/ttyUSB0 19200

Success looks like: You see JSON output with vehicle detections:

{ "magnitude": 1.2, "speed": 3.4 }

When vehicles pass, you’ll see more detailed information. The key is that you see JSON-formatted output (curly braces {}), not comma-separated values.

If it’s not working:

• No output at all? → Check baud rate (19200) and port (/dev/ttyUSB0 or /dev/serial0)
• Garbled text? → Verify sensor is in JSON mode (type OJ command from Step 2)
• CSV format? → Reconfigure sensor with OJ command from Step 2
• Permission denied? → Add your user to dialout group: sudo usermod -a -G dialout $USER then log out/in

Step 4: Install Software #

Estimated time: 30-60 minutes

On your Raspberry Pi:

# Clone the repository
git clone https://github.com/banshee-data/velocity.report.git
cd velocity.report

# Build the deployment tool and server binary
make build-deploy
make build-radar-linux

# Install as system service
./velocity-deploy install --binary ./velocity-report-linux-arm64

The installer will:

1. Install the binary to /usr/local/bin/velocity-report
2. Create a dedicated velocity service user
3. Create the data directory at /var/lib/velocity-report/
4. Install and start the systemd service

Success criteria:

# Verify service is running
sudo systemctl status velocity-report
# Should show "active (running)" in green

Where files are stored:

• Database: /var/lib/velocity-report/sensor_data.db (SQLite database with all vehicle detections)
• PDF Reports: Generated at tools/pdf-generator/output/ when requested
• Application logs: View with sudo journalctl -u velocity-report.service -f

If something goes wrong: The installation steps use the deployment tool commands shown later in this guide. Check logs with sudo journalctl -u velocity-report -f (press Ctrl+C to exit)

Step 5: Access the Web Dashboard #

Estimated time: 5 minutes

Open your browser and visit:

http://raspberrypi.local:8080

(Or use your Pi’s IP address: http://192.168.1.XXX:8080)

What you’ll see:

• Real-time vehicle detections with speeds and timestamps
• Speed distribution histograms
• Time-of-day traffic patterns
• Speed heatmaps

[PLACEHOLDER: Screenshot of web dashboard showing real-time vehicle detections, speed histogram, and time-of-day traffic patterns]

Success criteria: Dashboard loads and shows “No data yet” or live vehicle detections

If dashboard won’t load:

1. Check service is running: sudo systemctl status velocity-report
2. Find your Pi’s IP address: hostname -I
3. Try connecting from the Pi itself: curl http://localhost:8080/

If all else fails, check the logs: sudo journalctl -u velocity-report -f

Step 6: Mount the Radar Sensor #

Estimated time: 1-2 hours for complete weatherproof installation

1. Prepare weatherproof enclosure:

Mounting preparation:

• Drill mounting holes in back plate for hose clamps
• Install cable glands for power and (optional) Ethernet

Sensor positioning:

• Mount sensor inside with clear view through front panel
• Use acrylic or polycarbonate window if sensor doesn’t face forward

2. Install sensor inside enclosure:

• Aim radar sensor through front of enclosure
• Critical: Avoid metal obstructions in front of sensor (Doppler radar uses RF energy)
• Use plastic/nylon standoffs to mount sensor board

3. Mount enclosure to pole:

Positioning:

• Mount 4-8 feet off ground (reduces false positives from small objects)
• Use two stainless steel hose clamps (top and bottom)

Aiming:

• Angle: 20-45° off-axis from traffic flow
• Orientation: Face oncoming OR receding traffic (not perpendicular)
• Tighten clamps securely but avoid over-tightening (can crack enclosure)
• Record your mounting angle: Measure and note the angle off-axis for cosine correction in site configuration

[PLACEHOLDER: Diagram showing proper radar sensor mounting angle (20-45° off-axis from traffic flow) with top-down view of street and sensor position]

[PLACEHOLDER: Photo showing weatherproof enclosure mounted on utility pole with proper angle and positioning, including close-up of hose clamp mounting]

4. Weatherproofing checklist:

• ✅ All cable glands properly sealed
• ✅ Desiccant pack inside enclosure
• ✅ Enclosure gasket intact and clean
• ✅ Test enclosure seal before final mounting

Success criteria: Enclosure is weatherproof, sensor aims correctly, mounting is secure

Why mount higher? Mounting 4-8 feet off ground reduces false positives from animals, balls, or blowing debris. It provides cleaner line of sight to vehicle traffic.

Pole mounting best practices:

• Choose location with clear view (no trees/signs blocking)
• Ensure pole is stable (utility poles preferred over signposts)
• Check local regulations about attaching equipment to public infrastructure
• Consider solar panel if no AC power available nearby

Step 7: Generate PDF Reports #

Estimated time: Varies — requires data collection period

After collecting data for a few days or weeks, generate professional reports.

Via Web Dashboard:

1. Navigate to the Reports tab
2. Select your site from the dropdown
3. Configure report settings:
   • Date range: Select start and end dates for the report period
   • Cosine angle: Correction factor for sensor mounting angle (see below)
4. Click Generate Report
5. Download the PDF when ready

Cosine angle correction: If your sensor isn’t mounted parallel to traffic flow, measured speeds will be lower than actual speeds. The cosine angle setting compensates for this. For a sensor mounted at 30° off-axis, set cosine angle to 30°—the system applies the correction factor automatically. Leave at 0° if mounted parallel to traffic.

What’s in the report:

• p50 (median): Half of vehicles go faster than this
• p85 (traffic engineering standard): Speed at which 85% of traffic travels at or below
• p98 (top 2%): Threshold where the fastest regular drivers operate
• Histograms, time-of-day charts, and crash physics analysis

Comparison Reports: Measuring Intervention Effectiveness #

Comparison reports let you analyse the impact of traffic calming measures by comparing two time periods side-by-side. This is invaluable for advocacy—showing city officials that a speed hump reduced p85 speeds by 12 mph is far more compelling than anecdotal observations.

When to use comparison reports:

• Before/after interventions: Measure the effect of speed humps, signage, or enforcement campaigns
• Seasonal comparisons: Compare summer vs winter traffic patterns
• Week-over-week analysis: Track whether speeding issues are consistent or sporadic

To generate a comparison report via Web Dashboard:

1. Navigate to the Reports tab
2. Select your site from the dropdown
3. Set the Primary period dates (e.g., after intervention: 1-7 December 2025)
4. Enable Compare with previous period
5. Set the Comparison period dates (e.g., before intervention: 1-7 November 2025)
6. Click Generate Report

The report includes:

• Side-by-side metrics: p50, p85, p98 for both periods with percentage changes
• Dual-period histogram: Overlaid speed distributions with clear legend
• Comparison distribution table: Detailed breakdown of speed buckets across periods

[PLACEHOLDER: Sample page from PDF report showing speed distribution histogram, p50/p85/p98 statistics, and time-of-day traffic patterns]

Making your case: Print the report and bring it to city council. Instead of “cars go too fast,” say “85% of drivers exceed the posted 25 mph limit, with p85 at 38 mph.” With comparison reports, you can add: “After the speed hump installation, p85 dropped from 42 mph to 31 mph—a 26% reduction.”

Site Configuration Periods: Time-Based Sensor Settings #

When you reposition your sensor or adjust its mounting angle, historical data needs to be corrected using the angle that was in effect at the time of collection. The site configuration periods feature (based on a Type 6 Slowly Changing Dimension pattern) tracks these changes automatically.

Why configuration periods matter:

• Accurate historical data: If you moved the sensor from 15° to 30° on 1st December, data before that date uses the 15° correction, and data after uses 30°
• Retroactive corrections: Realised your angle measurement was wrong? Update the configuration period and all reports automatically apply the correct correction
• Comparison report accuracy: When comparing two time periods, each period uses the appropriate cosine correction for that date range

Managing configuration periods via Web Dashboard:

1. Navigate to Sites → select your site
2. Click Configuration Periods
3. Add a new period with:
   • Start date: When this configuration became active
   • End date: When this configuration ended (leave blank for current)
   • Cosine error angle: The sensor mounting angle for this period
   • Notes: Optional description (e.g., “Moved sensor to east side of pole”)

Example scenario:

When generating a report for April 2025, the system automatically applies the 35° correction. A comparison report spanning February (21°) vs April (35°) applies each correction independently.

Network Access & Security #

Local Network Deployment (Recommended) #

The web dashboard runs on port 8080 and is accessible on your local network by default:

http://raspberrypi.local:8080
# or
http://192.168.1.XXX:8080

Security considerations for LAN-only deployment:

• ✅ No authentication required if your network is trusted (home/office)
• ✅ Router firewall blocks external access by default
• ✅ Data never leaves your network - no cloud services
• ⚠️ Anyone on your WiFi can access the dashboard

Best practices:

• Use strong WiFi password (WPA3 if supported)
• Change default Pi password immediately
• Keep Pi OS updated: sudo apt update && sudo apt upgrade
• Consider network segmentation for additional security

Remote Access with Tailscale (Optional) #

Tailscale provides secure remote access from anywhere without exposing your Pi to the public internet.

Why Tailscale?

• Zero-configuration VPN
• End-to-end encrypted
• NAT traversal (works behind routers)
• Free for personal use (up to 20 devices)
• No port forwarding or dynamic DNS needed

Setup (5 minutes):

1. Install Tailscale on your Pi:

# Download and install Tailscale
curl -fsSL https://tailscale.com/install.sh | sh

# Start Tailscale and authenticate
sudo tailscale up

2. Authenticate via the URL shown (opens browser)

3. Install Tailscale on your phone/laptop from app store

4. Access dashboard from anywhere:

# Use the Tailscale IP shown in admin console
http://100.x.y.z:8080

Benefits:

• Access dashboard while away from home
• Share access with trusted colleagues (invite to Tailscale network)
• Monitor multiple deployments from single dashboard
• No exposure to public internet scanners

See also: Tailscale documentation

Public Internet Deployment (Not Recommended) #

Please, do not expose this service directly to the public internet. The dashboard has no authentication, no HTTPS, and no rate limiting.

If you need remote access, please use Tailscale.

Using Your Data for Advocacy #

Presenting to City Council #

Do:

• Print professional PDF reports
• Compare your data to posted speed limits
• Propose specific solutions (speed humps, signage, enforcement)
• Bring photos showing context (residential area, school zone)

Don’t:

• Share raw database dumps
• Attack specific drivers
• Make emotional appeals without data backup
• Demand immediate action without acknowledging budget constraints

Building Community Support #

1. Share with neighbours - Show them the data
2. Partner with local groups - PTA, neighbourhood associations
3. File public records requests - Compare to city traffic studies
4. Document over time - Show patterns, not one-off incidents

Example Talking Points #

• ❌ “Cars go way too fast on our street!”

• ✅ “85% of drivers exceed the 25 mph limit, with p85 at 39 mph, well above the engineering standard for residential safety.”

• ❌ “Someone’s going to get hurt!”

• ✅ “At 39 mph, crash energy is 143% higher than at the posted 25 mph limit. Our data shows consistent speeding during school hours.”

[PLACEHOLDER: Photo of community member presenting PDF report at city council meeting with speed data displayed on screen]

Troubleshooting #

Most issues happen because of:

1. Wrong baud rate (must be 19200)
2. Sensor still in CSV mode (run OJ command to switch to JSON)
3. Wrong device port (check ls /dev/tty* before and after plugging in sensor)
4. Insufficient power (use quality 2.5A+ power supply)

Quick fixes:

• No sensor data? → Check the device exists: ls /dev/ttyUSB0
• Service won’t start? → Check logs: sudo journalctl -u velocity-report -f
• Dashboard won’t load? → Verify service running: sudo systemctl status velocity-report
• Need more help? → See TROUBLESHOOTING.md or ask on Discord

Uninstalling #

To completely remove velocity.report:

# Stop and disable service
sudo systemctl stop velocity-report
sudo systemctl disable velocity-report

# Remove files
sudo rm /usr/local/bin/velocity-report
sudo rm /etc/systemd/system/velocity-report.service
sudo rm -rf /var/lib/velocity-report/

# Remove service user
sudo userdel velocity

Warning: This deletes all collected data. Export PDFs first if you want to keep them.

Wrap-Up & Next Steps #

You’ve successfully built a professional, weatherproof traffic radar system.

What you’ve accomplished:

• Built hardware for permanent neighbourhood traffic monitoring
• Configured a Doppler radar sensor with RS232 serial interface
• Deployed a complete web-based monitoring system
• Set up local data storage with no cloud dependencies
• Created a weatherproof infrastructure installation suitable for all-weather deployment

Keep it running: A week of data shows patterns. A month is compelling. Three months across different seasons is irrefutable evidence.

Maintenance recommendations:

• Check enclosure monthly for condensation
• Clean sensor lens seasonally
• Document installation with photos
• Monitor system logs for any issues
• Test weatherproofing seals periodically

Make it count:

Traffic safety advocacy shouldn’t require a six-figure budget or an engineering degree. With ~$350-450 in parts and a weekend of work, you’ve built something that produces the same metrics cities pay consultants thousands for.

Show your neighbours. File public records requests to compare your data to official counts. Bring your PDF report to city council meetings. Advocate for traffic calming with evidence nobody can dismiss.

Resources & Links #

• Project Overview: See the main README for project background and philosophy
• GitHub Repository: github.com/banshee-data/velocity.report
• OmniPreSense Support: omnipresense.com/support
• Community Discord: discord.gg/XXh6jXVFkt

Related Documentation:

• Troubleshooting: See TROUBLESHOOTING.md for common issues
• System Design: Read ARCHITECTURE.md for technical details
• Report Customisation: Check PDF Generator README
• Contributing: See CONTRIBUTING.md for conventions and workflow

Traffic Safety Resources:

• Vision Zero Network: visionzeronetwork.org
• NACTO Urban Street Design Guide: nacto.org
• FHWA Speed Management: safety.fhwa.dot.gov/speedmgt

Appendix: Sensor Selection Guide #

Understanding OmniPreSense Product Codes #

OmniPreSense offers radar sensors in multiple configurations. The product code format is:

203-OPS[model]-[data_type]-[modulation]-[interface]

Product Code Breakdown:

Examples:

• 203-OPS7243-A-CW-R2 = Sensor in IP67 housing, speed-only, continuous wave, RS232 interface
• 203-OPS7243-C-FC-R2 = Sensor in IP67 housing, speed+distance, FMCW, RS232 interface

Available Models Comparison #

Key specifications:

• A-type: Speed only, ≈100m range (recommended for traffic monitoring)
• C-type: Speed + distance, ≈60m range (FMCW)
• CW modulation: Doppler (speed measurement)
• FC modulation: FMCW (adds range/frequency-modulated distance)
• R2 interface: RS232 industrial (requires serial HAT)
• 7243 series: IP67 weatherproof enclosure for outdoor deployment

Power Requirements #

All models operate on 5V DC:

RS232 models (R2 interface):

• Require separate 5-24V power supply
• RS232 provides data lines only, no power
• Typical draw: 300-440mA at 5V (~2.2W)

Important: RS232 models (R2) require external 5V power in addition to the RS232 data connection.

Back to top

Let’s build safer streets together. If the speeds don’t drop, the work can’t stop.