Livestock Tracking with GPS Collars

System Overview

GPS livestock tracking reduces labor costs, prevents theft, and enables early disease detection. Smart collars provide real-time location data, activity monitoring, and behavioral analytics for precision livestock farming.

System Architecture

[GPS Collar] → [LoRaWAN/Cellular] → [Gateway] → [Cloud Platform]
     ↓                                              ↓
  GPS Module                                    Web Dashboard
  Accelerometer                                 Mobile App
  Battery                                       Alerts (SMS/Email)
  Temperature Sensor                            Analytics

Hardware Components

Per Collar Unit

• ESP32 Microcontroller: Low-power WiFi/BT controller
• NEO-6M/NEO-M8N GPS: GPS receiver with antenna
• LoRa Module (SX1276): Long-range communication (optional)
• MPU6050: Accelerometer for activity tracking
• DS18B20: Waterproof temperature sensor
• 18650 Li-ion Battery: 3.7V 3000mAh (or solar + battery)
• TP4056: Battery charging module
• Waterproof Enclosure: IP67 rated box

Power Budget

Component	Active Current	Sleep Current	Duty Cycle
ESP32 Deep Sleep	-	10 μA	95%
GPS (Active)	50 mA	-	3%
LoRa TX	120 mA	-	2%
Average Current	~3.5 mA
Battery Life	~35 days (3000mAh)

Collar Assembly

Enclosure Design

• Use UV-resistant plastic enclosure (IP67 minimum)
• Mount GPS antenna on top (clear sky view)
• Pot electronics with silicone for waterproofing
• Attach to standard cattle collar with cable ties
• Weight: Keep under 300g for animal comfort

Wiring Diagram

ESP32
│
├── NEO-6M GPS
│   ├── VCC → 5V (or 3.3V)
│   ├── GND → GND
│   ├── Tx → GPIO17
│   └── Rx → GPIO16
│
├── MPU6050 Accelerometer
│   ├── VCC → 3.3V
│   ├── GND → GND
│   ├── SDA → GPIO21
│   └── SCL → GPIO22
│
├── LoRa SX1276
│   ├── 3.3V → 3.3V
│   ├── GND → GND
│   ├── NSS → GPIO5
│   ├── MOSI → GPIO23
│   ├── MISO → GPIO19
│   ├── SCK → GPIO18
│   ├── RST → GPIO14
│   └── DIO0 → GPIO26
│
└── DS18B20 Temperature
    ├── VCC → 3.3V
    ├── GND → GND
    └── DATA → GPIO4 (with 4.7k pullup)

GPS Tracker Firmware

Low-power firmware for GPS tracking collar:

#include <TinyGPS++.h>
#include <SPI.h>
#include <LoRa.h>
#include <ArduinoJson.h>
#include <esp_sleep.h>

// GPS on Hardware Serial
#define GPS_RX 17
#define GPS_TX 16
HardwareSerial gpsSerial(1);
TinyGPSPlus gps;

// LoRa pins
#define LORA_SS 5
#define LORA_RST 14
#define LORA_DIO0 26

// MPU6050
#include <Wire.h>
#include <MPU6050.h>
MPU6050 mpu;

// Power management
#define SLEEP_INTERVAL 300000 // 5 minutes
#define GPS_FIX_TIMEOUT 30000 // 30 seconds max for GPS fix

void setup() {
  Serial.begin(115200);
  
  // Initialize GPS
  gpsSerial.begin(9600, SERIAL_8N1, GPS_RX, GPS_TX);
  
  // Initialize LoRa
  LoRa.setPins(LORA_SS, LORA_RST, LORA_DIO0);
  if (!LoRa.begin(915E6)) { // 915 MHz for US
    Serial.println("LoRa init failed!");
  }
  
  // Initialize MPU6050
  Wire.begin();
  mpu.initialize();
  if (!mpu.testConnection()) {
    Serial.println("MPU6050 connection failed");
  }
  
  // Configure deep sleep wake-up
  esp_sleep_enable_timer_wakeup(SLEEP_INTERVAL);
}

void loop() {
  // Get GPS fix
  gpsData gpsData = getGPSPosition();
  
  // Get accelerometer data
  int16_t ax, ay, az;
  mpu.getAcceleration(&ax, &ay, &az);
  float activity = calculateActivity(ax, ay, az);
  
  // Get temperature (if sensor connected)
  float temp = readTemperature();
  
  // Create payload
  StaticJsonDocument<256> doc;
  doc["lat"] = gpsData.lat;
  doc["lng"] = gpsData.lng;
  doc["activity"] = activity;
  doc["temp"] = temp;
  doc["battery"] = readBattery();
  doc["id"] = "COW_001";
  
  // Send via LoRa
  sendLoRa(doc);
  
  // Go to sleep
  Serial.println("Going to sleep...");
  esp_deep_sleep_start();
}

struct gpsData {
  float lat;
  float lng;
};

gpsData getGPSPosition() {
  unsigned long start = millis();
  gpsData data = {0, 0};
  
  while (millis() - start < GPS_FIX_TIMEOUT) {
    while (gpsSerial.available()) {
      gps.encode(gpsSerial.read());
    }
    
    if (gps.location.isUpdated()) {
      data.lat = gps.location.lat();
      data.lng = gps.location.lng();
      Serial.printf("GPS: %.6f, %.6f\n", data.lat, data.lng);
      return data;
    }
  }
  
  Serial.println("GPS timeout");
  return data;
}

float calculateActivity(int16_t ax, int16_t ay, int16_t az) {
  // Calculate vector magnitude
  float magnitude = sqrt(ax*ax + ay*ay + az*az);
  
  // Classify activity
  if (magnitude > 1500) return 3.0; // Running
  if (magnitude > 1200) return 2.0; // Walking
  if (magnitude > 900) return 1.0;  // Grazing
  return 0.5; // Resting
}

void sendLoRa(StaticJsonDocument<256>& doc) {
  char buffer[256];
  serializeJson(doc, buffer);
  
  LoRa.beginPacket();
  LoRa.print(buffer);
  LoRa.endPacket();
  
  Serial.println("LoRa packet sent");
  Serial.println(buffer);
}

float readBattery() {
  int adc = analogRead(35); // Battery voltage pin
  float voltage = adc * (3.3 / 4095.0) * 2; // Voltage divider
  return voltage;
}

float readTemperature() {
  // Implement DS18B20 reading
  return 25.0; // Placeholder
}

Connectivity Options

Comparison Table

Technology	Range	Power	Cost	Best For
LoRaWAN	5-15 km	Very Low	$	Large ranches, rural areas
Cellular (NB-IoT)	Unlimited	Low	$$	Areas with cell coverage
Satellite	Global	High	$$$$	Remote rangeland
WiFi	< 100m	Medium	$	Feedlots, barns

LoRaWAN Gateway Setup

# Install ChirpStack Gateway Bridge
docker run -d \
  --name chirpstack-gateway \
  -p 1700:1700/udp \
  -e GATEWAY_ID=YOUR_GATEWAY_ID \
  -e SERVER=eu1.cloud.thethings.network:1700 \
  chirpstack/chirpstack-gateway-bridge

Tracking Platform

Set up real-time tracking dashboard:

Node-RED Flow for GPS Tracking

[
  {
    "id": "lorawan-input",
    "type": "udp in",
    "port": 1700,
    "datatype": "utf8"
  },
  {
    "id": "parse-gps",
    "type": "json",
    "wires": [["geofence-check"]]
  },
  {
    "id": "geofence-check",
    "type": "function",
    "func": "var lat = msg.payload.lat;\nvar lng = msg.payload.lng;\nvar inside = isInsideGeofence(lat, lng);\nmsg.inside = inside;\nif (!inside) {\n  msg.alert = 'Animal outside geofence!';\n}\nreturn msg;"
  },
  {
    "id": "store-location",
    "type": "postgres",
    "query": "INSERT INTO locations (animal_id, lat, lng, timestamp) VALUES ($1, $2, $3, NOW())"
  },
  {
    "id": "map-display",
    "type": "ui_map",
    "name": "Livestock Map",
    "lat": "payload.lat",
    "lon": "payload.lng"
  }
]

Geofencing Algorithm

function isInsideGeofence(lat, lng) {
  // Define pasture boundaries (polygon)
  const pasture = [
    {lat: 35.123, lng: -106.456},
    {lat: 35.125, lng: -106.456},
    {lat: 35.125, lng: -106.450},
    {lat: 35.123, lng: -106.450}
  ];
  
  // Ray casting algorithm
  let inside = false;
  for (let i = 0, j = pasture.length - 1; i < pasture.length; j = i++) {
    if (((pasture[i].lat > lat) !== (pasture[j].lat > lat)) &&
        (lng < (pasture[j].lng - pasture[i].lng) * (lat - pasture[i].lat) / 
         (pasture[j].lat - pasture[i].lat) + pasture[i].lng)) {
      inside = !inside;
    }
  }
  return inside;
}

Health Monitoring

Behavioral Indicators

Condition	Activity Pattern	Alert Threshold
Healthy	Regular grazing, resting	-
Sick/Injured	Low activity, isolated	< 2 hours movement/day
In Heat	Increased movement, mounting	> 3x normal activity
In Labor	Restless, frequent position changes	Isolated + high activity
Predator Attack	Sudden sprint, scattered herd	All animals running

Alert Configuration

# Alert rules for livestock monitoring
alerts:
  - name: "Geofence Breach"
    condition: "location outside pasture"
    action: "send_sms + send_email"
    priority: "high"
  
  - name: "Low Activity"
    condition: "activity < 0.5 for 24h"
    action: "send_email"
    priority: "medium"
  
  - name: "Heat Detection"
    condition: "activity > 3x baseline"
    action: "log_event"
    priority: "low"
  
  - name: "Low Battery"
    condition: "battery < 3.5V"
    action: "send_email"
    priority: "medium"
  
  - name: "No Data"
    condition: "no update for 2 hours"
    action: "send_sms"
    priority: "high"