How to Connect a GPS/GLONASS NMEA Tracker to AI Agent ASI Biont for Real‑Time Fleet Monitoring and Geofencing

Why Integrate a GPS Tracker with an AI Agent?

Managing a fleet of vehicles, courier bikes, or taxi cabs means tracking locations, setting geofences, and sending alerts when a vehicle enters or leaves a zone. Traditionally, this requires dedicated server software, complex configuration, and manual scripting. With ASI Biont, you simply describe your setup in chat, and the AI agent connects directly to your GPS/GLONASS NMEA tracker via a COM port, parses the data, and automates notifications—all without writing a single line of code yourself.

What Is a GPS/GLONASS NMEA Tracker?

Most GPS/GLONASS trackers output location data as NMEA 0183 sentences over a serial port (RS‑232 or TTL). Common sentences include $GPGGA (fix data), $GPRMC (recommended minimum), and $GPGSA (satellites). The device sends these strings at a fixed interval (e.g., 1 Hz) at baud rates like 4800, 9600, or 115200. To read this data, you need a COM port (physical or USB‑to‑serial adapter) on your computer.

How ASI Biont Connects to the Tracker

ASI Biont does not run on your local machine. Instead, it uses a lightweight Hardware Bridge (bridge.py) that you run on your PC. The bridge opens the COM port and maintains a WebSocket connection to the ASI Biont cloud. When you ask the AI to read the GPS data, it sends an industrial_command with protocol serial:// and the command serial_write_and_read. The bridge writes any hex data you provide (often empty, just to trigger a read) and returns the raw NMEA sentence from the tracker.

Real commands used:

  • industrial_command(protocol='serial://', command='serial_write_and_read', data='', target='COM3') – reads the next available NMEA line.
  • industrial_command(protocol='serial://', command='serial_write_and_read', data='4154540D0A') – sends ATT\r\n (some trackers accept AT commands) and returns the response.

The bridge handles all low‑level serial I/O (including Windows overlapped I/O fixes) and relays the response back to the AI.

Step‑by‑Step: Connect a GPS Tracker and Set Up Geofence Alerts

1. Download and Run the Bridge

From the ASI Biont dashboard (Devices → Create API Key → Download bridge), get bridge.py. Install dependencies:

pip install pyserial requests websockets

Run it with your token and the COM port:

python bridge.py --token=your_api_token --ports=COM3 --baud 9600 --rate=1

Now the bridge is listening on COM3 and connected to ASI Biont.

2. Ask the AI to Read GPS Data

In chat, write:

“Connect to the GPS tracker on COM3, read the NMEA sentence, and show me the current latitude and longitude.”

The AI sends serial_write_and_read (empty data) and receives something like:

$GPGGA,123519,4807.038,N,01131.000,E,1,08,0.9,545.4,M,46.9,M,,*47

3. Parse with AI and Check Geofence

The AI can then use execute_python to run a sandboxed script that parses the NMEA with pynmea2 and checks whether the point lies inside a geofence polygon. Here’s the code the AI would generate and execute (this runs in the cloud, not on your PC):

import pynmea2
import json

# Raw NMEA received from bridge
nmea_str = "$GPGGA,123519,4807.038,N,01131.000,E,1,08,0.9,545.4,M,46.9,M,,*47"

# Parse the sentence
msg = pynmea2.parse(nmea_str)
lat = msg.latitude
lon = msg.longitude
print(f"Latitude: {lat}, Longitude: {lon}")

# Simple rectangular geofence (Berlin center ±0.1°)
if 52.4 <= lat <= 52.6 and 13.3 <= lon <= 13.5:
    print("Inside geofence")
else:
    # Send Telegram alert
    import requests
    token = "YOUR_TELEGRAM_BOT_TOKEN"
    chat_id = "YOUR_CHAT_ID"
    msg_text = f"Vehicle {lat},{lon} is outside geofence!"
    requests.post(f"https://api.telegram.org/bot{token}/sendMessage",
                  data={"chat_id": chat_id, "text": msg_text})
    print("Alert sent!")

The AI runs this script, parses the data, and if the coordinate is outside the zone, it sends a Telegram notification. All in one chat conversation.

4. Automate Periodic Checks

Because the sandbox has a 30‑second timeout, you can’t run an infinite loop. Instead, set up a schedule by asking the AI:

“Every 5 minutes, read COM3, parse the NMEA, and send me a Telegram report with current position and geofence status.”

The AI will instruct you to run a small background script on your local machine (using the bridge’s WebSocket feed) or simply remind you to ask again. For true automation, the AI can generate a .bat or shell script that calls industrial_command repeatedly via the ASI Biont API (using your token). You can then schedule that script with cron or Task Scheduler.

Why This Integration Is a Game Changer

  • No manual coding – AI writes and runs the integration code. You just describe your hardware.
  • Works with any NMEA tracker – baud rate, sentence type, or brand doesn’t matter; the bridge delivers raw data.
  • Immediate results – set up geofencing, history logging, or Telegram alerts in under a minute.
  • Extensible – once the data is in ASI Biont, you can combine it with other devices, generate Excel reports (using openpyxl), or plot routes on a map with matplotlib.

Real‑World Pitfalls to Avoid

  • Baud rate mismatch – Most GPS modules default to 9600, but check your datasheet. Use --baud correctly.
  • Permission issues on Linux – Add your user to the dialout group: sudo usermod -a -G dialout $USER.
  • NMEA sentence filtering – Some trackers output many sentences per second. The bridge reads only one per serial_write_and_read call. Use --rate to throttle or read multiple times.
  • Geofence accuracy – Avoid complex polygons in sandbox; use simple bounding boxes or point‑in‑polygon algorithms (e.g., ray casting).

Conclusion

Integrating a GPS/GLONASS NMEA tracker with ASI Biont turns your fleet into a smart, AI‑managed system. You no longer need a dedicated backend – just a COM port, the Hardware Bridge, and a few sentences in chat. The AI handles parsing, geofencing, and notifications, putting real‑time logistics control in your hands.

Try it yourself today – create an account on asibiont.com, download the bridge, and tell the AI: “Connect to my GPS tracker on COM3 and alert me when a vehicle leaves the delivery zone.” The AI does the rest.

← All posts

Comments