Case Study: MADPAW — Pet Tracker, GPS + Activity ML, 30-Day Battery Life

# Case Study: MADPAW — Pet Tracker, GPS + Activity ML, 30-Day Battery Life

Problem: A 30+ day battery life target for pet tracker hardware (a collar device). The device performs GPS location tracking plus activity classification (sleep/walk/run) from a 3-axis accelerometer, and syncs with the owner's iPhone.

Result: 32-day battery test passed (lab environment), 95% activity accuracy, 12K devices sold, App Store rating 4.8/5.

1. Power Budget

The pet tracker uses an 800mAh LiPo battery. The processing that must not drain it within 30 days:

Optimization was required.

2. Smart Scheduling

The device stays in native sleep and only wakes when motion is detected:

• Accelerometer interrupt (motion threshold) wake-up
• Motion detected → ML inference + GPS fix
• No motion → sleep mode, GPS off

With this pattern, the daily budget dropped to 3%/day. 30 days = 90%. 32-day lab test passed.

3. On-Device ML Pipeline

3-axis accelerometer at 25Hz → 60-sample window → Core ML model.

Model: Create ML Activity Classifier, 4 classes (sleep/walk/run/play). Training:

• 200 hours of data (15 dogs, various breeds and sizes)
• Train 80% / val 10% / test 10%
• Validation accuracy: 94.2%
• Model size: 18KB (.mlmodelc)

The ESP32-class hardware has no Apple Neural Engine — the Core ML model was converted to TFLite Micro. Inference: 8ms / window.

4. iOS App: Owner Side

The iOS app is a low-power BLE listener:

1@MainActor
2final class TrackerManager: ObservableObject {
3    private var central: CBCentralManager?
4    @Published var lastSyncedAt: Date?
5    @Published var lastActivity: ActivityClass?
6 
7    func startListening() {
8        central = CBCentralManager(delegate: self, queue: .main, options: [
9            CBCentralManagerOptionRestoreIdentifierKey: "MADPAW.tracker"
10        ])
11    }
12}

Background BLE scan runs only when the phone is home (geofence) — for battery savings.

Map UI: GPS history for the last 24 hours, activity heatmap.

5. Cloud Sync

The iOS app acts as a proxy:

• BLE \<-\> Firebase Realtime DB
• Activity events: tracker → BLE → iOS → Firebase
• Owner can share with multiple family members (Firestore family-id collection)

Daily volume per device: ~3MB. Firebase free tier is sufficient (12K devices → ~36GB/month).

6. Hardware Pin-out & Firmware

• ESP32 (S3, ARM Cortex M4)
• u-blox NEO-M9N GPS module
• ADXL355 accelerometer
• Power management IC (TI BQ24295)
• BLE-only (no Wi-Fi during travel — power)

Firmware in C (Arduino IDE → migrating to ESP-IDF in v2):

• FreeRTOS task layout
• 4KB ML model embedded
• OTA via BLE (iOS app uploads firmware)

7. Manufacturing & Production

• 12K devices sold
• China supplier (Shenzhen)
• 4 hardware revisions over 18 months
• App Store-side approval was fairly smooth (no NFC, no health regulatory)
• Returns: 1.8% (industry 3-5%)

8. Lessons Learned

1. The battery target must be committed to upfront. "I'll optimize it later" doesn't work — the power profile needs to be a day-0 target.

2. Always-on listener vs. scheduled yielded a 95% battery saving. Motion-triggered wake-up is best practice for DIY hardware.

3. ML model size matters. The 18KB Core ML model works via TFLite Micro. At 100KB+ it wouldn't have fit on the hardware.

4. BLE-only design. Wi-Fi consumed too much power during travel. Wi-Fi is used only for the home dock and syncing.

5. iOS background BLE requires CBCentralManagerOptionRestoreIdentifierKey — without restoration, sync would drop while the owner was away from home.

6. Activity training data: 200 hours — the first model used 50 hours, with 78% validation accuracy. 4x the data → a jump to 94%.

7. App Store reviews are driven by the battery accuracy claim. A "one charge per 30 days" claim needs proof. Lab data plus a 100-customer beta test supported the App Store description. Trust increases conversion by 35%.

Result: IoT hardware + iOS app is a combined problem. Battery optimization, on-device ML, and the BLE state machine each carried 2 months of technical debt. Combined, that's 6-8 months of R&D. When the output is the best-in-class pet tracker, it's worth it.