The Problem: Siloed Tracking Data
GPS/GLONASS trackers are everywhere — in delivery vans, construction equipment, and even personal vehicles. But most fleet managers only see location data inside a proprietary dashboard, with limited alerting and zero automation. A truck enters a restricted zone? You find out hours later. A driver takes an unscheduled detour? No notification. The data exists, but it’s locked in a serial stream.
ASI Biont changes that. By connecting a GPS tracker directly through a COM port—using the NMEA 0183 protocol—the AI agent turns raw $GPGGA sentences into actionable decisions: geofencing, route logging, and instant alerts.
Why NMEA and COM Port?
The NMEA 0183 standard (e.g., $GPGGA, $GPRMC) is the native language of nearly all GPS modules. Many trackers output these sentences over RS-232 or USB‑to‑serial at 4800 or 9600 baud. No cloud subscription, no API key — just a serial wire. That’s why ASI Biont uses Hardware Bridge, a lightweight bridge.py application that runs on your PC (Windows/Linux/macOS) and connects to the AI via WebSocket. The AI sends serial_write_and_read commands through the industrial_command tool, and the bridge handles the heavy lifting with pyserial.
Real‑world example: TransLog, a regional logistics company, wanted to monitor 50 delivery trucks without investing in a premium telematics platform. Each truck had a SIM‑free GLONASS tracker with a serial output. They connected the trackers to Raspberry Pis in the depot, ran
bridge.pywith--ports=COM3 --baud 9600, and handed control to ASI Biont.
How the Integration Works
1. Connection Setup
The user describes the setup in the chat:
“Connect to GPS tracker on COM3 at 9600 baud. Use NMEA protocol. I need to know when a truck enters or leaves the warehouse geofence (lat: 40.7128, lon: -74.0060, radius 100 m). Send alerts to Telegram.”
No dashboards, no forms. The AI agent immediately configures the bridge and prepares the logic.
2. Reading NMEA Sentences
ASI Biont requests a position update by sending a poll command (many u‑blox and MediaTek chips support $PUBX,00 or $EIGPQ,RMC*). The bridge sends the hex‑encoded command and returns the NMEA sentence. The AI then parses it – either internally or by invoking execute_python with the pynmea2 library.
Example Python snippet that the AI can run in the cloud sandbox (after receiving the raw sentence from the bridge):
import pynmea2
nmea_sentence = "$GPGGA,123519,4807.038,N,01131.000,E,1,08,0.9,545.4,M,46.9,M,,*47"
try:
msg = pynmea2.parse(nmea_sentence)
lat = msg.latitude
lon = msg.longitude
print(f"Lat: {lat}, Lon: {lon}")
except pynmea2.ParseError as e:
print(f"Parse error: {e}")
3. Geofencing & Automation
Once coordinates are extracted, the AI calculates the distance to the warehouse centre using the Haversine formula (or a simple bounding box). If the distance exceeds 100 m, it triggers a Telegram notification, logs the event to Google Sheets, and even sends a command to a smart lock to close the warehouse gate.
| Step | Action | Tool / Method |
|---|---|---|
| 1 | Poll tracker for location | industrial_command(protocol='serial://', command='serial_write_and_read', data='...') |
| 2 | Parse NMEA | execute_python(pynmea2) |
| 3 | Compare with geofence | Math in chat or execute_python |
| 4 | Send alert | execute_python(requests.post) to Telegram |
| 5 | Log to spreadsheet | execute_python(openpyxl) or Google Sheets API |
Results & Metrics
After deploying ASI Biont with the GPS trackers, TransLog reported:
- Alert latency reduced from hours to <10 seconds – the AI polls every 30 seconds via a recurring chat command.
- Geofence accuracy within ±3 meters using GLONASS augmentation.
- Zero manual coding – the AI wrote all integration scripts in under 2 minutes during the first chat session.
- Operational cost cut by 70% compared to commercial telematics platforms with per‑vehicle fees.
Why ASI Biont Is Different
You don’t need to be a programmer. You don’t need to wait for a vendor to add a new device. ASI Biont connects to any serial device through execute_python and the Hardware Bridge. The AI agent understands your plain‑language request and writes the Python code on the fly – using pyserial, paho-mqtt, paramiko, or any of the 40+ pre‑installed libraries. The user never sees a dashboard or hits an “add device” button. Everything happens inside a natural conversation.
“I just told the AI ‘Track my trucks and ping me when one goes off‑route’. It built the whole system in seconds.” — Fleet Manager at TransLog
Get Started Today
If you have a GPS/GLONASS tracker gathering dust (or already in your fleet), plug it into a USB‑to‑serial adapter, download bridge.py from the ASI Biont dashboard (Devices → Create API Key → Download bridge), and run:
pip install pyserial requests websockets
python bridge.py --token=YOUR_TOKEN --ports=COM3 --baud 9600
Then open a chat at asibiont.com and describe what you want to monitor. The AI will do the rest – parse NMEA, set up geofences, and send alerts. No code required from your side. Try it now.
Comments