Cattle Tracking

Mass-Market Scale for LatAm

Partnering for Infrastructure-Zero Traceability

layout: image-right image: /ganaderia_extensiva.png

The LatAm Reality

Extensive Grazing & Traceability

The Context: Millions of hectares, zero infrastructure. No Wi-Fi, no power lines, and spotty cellular coverage.
The Opportunity: A Total Addressable Market (TAM) of 250M+ cattle across the Mercosur region.
The Target: Our business goal is to capture 1% of this market (2.5M connected animals).
The Testbed (Uruguay): 12M head of cattle. Mandatory traceability via the government (SNIG) using RFID and visual tags.
The Challenge: We must build a solution that assumes the pasture provides absolutely zero technological support.

The Functional Mandate

Five Use Cases from One Telemetry Stream

### Biological & Operational Value * **Mortality & Theft Prevention:** Real-time detection of sustained inactivity or non-biological acceleration (truck transport). * **Lameness Detection:** Gait signature anomalies detected days before visual limping appears. * **Estrus Detection:** Precise insemination windows based on mounting-behavior signatures.

### Institutional Value * **EUDR & Traceability Provenance:** Continuous, cryptographically-signed geolocation logs for international compliance. * **Fintech-Ready Audit Trail:** Immutable behavioral logs turn the herd into a bankable asset for collateralized finance and insurance.

The Architectural Constraint: All five bands are derived from the same on-animal accelerometer + position telemetry. We do not change hardware to enable a new use case; we deploy new Edge AI models.

layout: section

The Economic Ceiling

$50 USD per Animal (3-Year Cycle)

### The Restraint **Mass-Market Scale Requires Extreme BOM Reduction** * To deploy across millions of animals, the Total Cost of Ownership (TCO) cannot exceed **an average blended cost of $50 per animal** for a 3-year lifespan. * This is the final price the rancher pays, meaning this $50 ceiling must cover the blended hardware cost (mixing 97 cheap tags with 3 expensive collars) **+** Connectivity **+** Platform **+ Revenue/Margin**. * *Note: The hardware BOM is only one fraction of this ceiling. Every dollar saved on the Collar BOM is a dollar gained in viable subscription margin.* * **Implication:** We cannot afford per-animal satellite subscriptions, nor can we put heavy GNSS/GPS tracking on every single cow.

### The 97/3 Asymmetric Solution **The Herd Acts as its Own Infrastructure** * **97% Members (Ear Tags):** Ultra-low cost, ultra-low power. Simple telemetry (ID, step count). * **3% Leaders (Cowbells):** The "Gateways." Carried by a few cows. They collect data from Members, process it, and transmit it to the cloud.

layout: section

The Hardware Foundation

Tags, Collars, and the Modular Carrier

The Ear-Tag (Member Node)

### Hardware Constraints | Feature | Target / Constraint | |---------|---------------------| | **Weight** | < 20g (Single-pin). If > 20g, dual-pin required. | | **Power** | 3-5 year lifespan via primary battery. Solar harvesting is an optional evaluation but carries high mud-obscuration risks. | | **Durability** | IP68+ for extreme outdoor | | **RF Protocol** | Sub-GHz | | **RF Coexistence** | Must pack a 134.2 kHz passive RFID coil and active Sub-GHz radio in the same 20g housing without signal interference. |

### Behavior & Data Collection | Function | Expected Behavior | |----------|-------------------| | **Positioning** | **Open Evaluation.** As precise as possible within power/cost limits (On-Device GNSS vs Mesh Inference). | | **Sensors** | Accelerometer (behaviour analysis), Battery Health. | | **Telemetry** | 15-minute "heartbeat" | | **Alerts** | Immediate event-based triggers | | **Firmware** | Burn-once (no FOTA for Phase 1) |

The Collar Tracker (Leader Node)

### Hardware Sizing Rationale | Component | Requirement & Rationale | |-----------|-------------------------| | **MCU** | **Cortex-M4/M33 class or equivalent.** Minimum compute required to run local Edge AI to filter telemetry and avoid expensive NTN satellite transmissions. | | **RAM** | **Min 512KB.** Required to run Zephyr RTOS, local Sub-GHz stack, and TinyML state-inference models. | | **Storage** | **8-16MB Non-Volatile Memory.** Caches herd telemetry during satellite outages; stages dual-bank FOTA updates. | | **Zero-Port I/O**| **Magnetic wake + NFC/BLE.** Zero physical ports. | | **Power** | **6m (Pilot) / 1yr (Production).** Achieved via aggressive GNSS duty cycling. **Serviceable design:** batteries can be swapped by ranchers in the field. | | **Weight** | **< 1 Kg maximum.** Maximum acceptable weight for the entire collar assembly to ensure animal welfare. |

### Behavior & Network Flow | Function | Expected Behavior | |----------|-------------------| | **HA Cluster Architecture** | **Min 3 Collars per Herd.** The herd is managed by a cluster, not a single collar. If one battery dies, tags automatically re-associate to surviving collars. Max 500 tags per collar. | | **Positioning** | **4+ GNSS coordinates/day.** High-resolution local tracking. | | **Uplinks** | **NTN & Cellular.** 4x daily NTN batched status. Cellular fallback used for FOTA and real-time tracks. | | **Edge AI** | **Behavior & Mortality.** TinyML classifies state. Triggers instant alerts for theft or mortality. |

The Modular Cowbell Concept

Decoupling the Carrier from the Comms

### The Core Hardware Strategy * **The Problem:** The mechanical collar housing must survive **5+ years**, but satellite/cellular modems become obsolete in **1-3 years**. * **The Solution:** The Cowbell is designed as a **Carrier Board** with **Hot-Swappable Comms Modules** (similar to a standard M.2 socket). * **The Carrier:** Contains the MCU, GNSS, power-path, and ruggedized housing. * **The Module:** Contains the specific modem (NB-IoT, NTN, LEO L-Band).

### The Open Antenna Challenge * Different comms modules operate at vastly different frequencies (sub-GHz, L-band, S-band). * **The Design Question:** Do the antennas live on the hot-swap module itself (simpler integration, higher module BOM) or does the carrier provide multi-feed antenna real estate? * *This is a primary co-design focus area for the ODM partner.*

layout: section

The Network Fabric

Topology, Sub-GHz, and Connectivity

layout: image-right image: /hereford_two_tier_network.png

The "Local Cow Network"

The Spatial & Physics Challenge

* **Sparse Distribution:** Cattle graze across vast, open terrain and can be highly dispersed. * **Adjacent Link Budget:** We require a minimum **100m** reach between *adjacent* cows, with an ideal target of **1km** cow-to-cow. * **Network Span:** A single "Local Cow Network" cluster must support a maximum diameter of **10km**. * **Obstructions:** Operating at an average height of **1.2 meters from the ground**—instead of on a tall pole like normal LoRa gateways—means the earth (terrain) and the 500kg water-mass of the animals (body shadowing) severely absorb RF energy.

layout: center

The Topology Debate

Star vs. Mesh (Under Evaluation)

### Star Topology (Direct to Leader) * **Pros:** Best battery life for the ear tag (it just broadcasts and sleeps). * **Cons:** Requires strong, penetrating sub-GHz radios to reach the Leader directly through the herd and terrain.

### Mesh Topology (Multi-Hop) * **Pros:** Relays data cow-to-cow, bypassing Line-of-Sight issues. * **Cons:** Constant neighbor discovery while moving destroys the ear tag's 3-year battery budget. * **Routing Cost:** A 10km network at 1km links requires **up to 10 hops**, compounding packet collision risks and latency.

Radio Stack Evaluation

Open Evaluation for Sub-GHz Protocols

* **The Goal:** An ultra-low power radio stack. A royalty-free protocol is highly desired to protect mass-scale unit economics, but we are fully open to any technically superior proprietary stack. * **Candidates under evaluation:** * Sub-GHz IEEE 802.15.4g mesh (Zephyr-native) * LoRa P2P / Mesh * MeshCore / ZephCore (Open-source MIT-licensed) * Mioty / DECT NR+ * *Selection will be driven purely by field-measured battery budget and body-shadowing attenuation.*

Member Positioning Strategy

Two Paths Under Evaluation

### Path A: On-Device GNSS * Independent geolocation per ear tag. * **Pros:** Highest absolute spatial resolution. * **Cons:** Highest power draw; GNSS cold-fix energy dominates the battery budget.

### Path B: Mesh-Relative Inference * ToF / RSSI triangulation against the high-precision GNSS anchor on the Leader. * **Pros:** Massive power savings on the Member node. * **Cons:** Lower absolute spatial resolution between Leader anchors.

Firmware-Level Decision: The MCU/Radio hardware must support both paths. The final choice is made in software.

layout: section

Edge Intelligence & Fleet Ops

Algorithms, Storage, and Device Lifecycle

Algorithmic Basis & Edge Intelligence

Sandboxed Plug-in Architecture

### The Intelligence Split * **Members (Tags):** Deterministic classification only (Grazing, Walking, Panic) via rule-based logic. **No AI.** * **Leaders (Cowbells):** Runs the Edge AI analytics workload (Lameness, Estrus, Theft) via TinyML/Zephyr. * *JAAB provides the intelligence; the ODM provides the compute substrate.*

### Aspirational: Wasm/WASI Plug-ins * To let analytics evolve over the 5-year Cowbell lifespan, we are evaluating packaging Edge AI models as **sandboxed WebAssembly plug-ins**. * Safely update behavioral models over the cellular link *without* risking a full Zephyr OS firmware re-flash.

Storage Capacity Rationale

Why 16MB NVRAM is the Hard Minimum for the Collar

### 1. Data Buffering (Offline Caching) To survive NTN satellite outages, the Collar must buffer **3 days** of offline herd telemetry. * **Load:** 500 Tags × 15-min heartbeats = 48,000 payloads/day. * **Size:** ~50 bytes per payload = ~2.4 MB/day. * **Cache Need:** 3 days × 2.4 MB = **~7.2 MB**

### 2. Dual-Bank FOTA Staging Safe Over-The-Air updates require dedicated staging space for Zephyr OS and Edge AI models. * **Staging Need:** **~4 MB**

7.2 MB (Data) + 4.0 MB (OTA) = 11.2 MB Minimum
A standard 16MB (128Mbit) NVRAM is strictly required.

Dual-Path Telemetry Protocol

Edge AI Compresses Data for the Fallback Link

### [Primary] Cellular Path (NB-IoT) **High Bandwidth • Low Cost** * **Active:** When cellular coverage is available. * **Protocol:** NATS C-Client (or MQTT) over TCP. *(Note: We maintain the NATS-based open-source project fluxrig.org and expect to use it, but are open to evaluating alternatives).* * **Security:** TLS 1.3 encryption. * **Telemetry:** High-res behavior analytics & real-time events. * **Management:** Full LwM2M and FOTA enabled. * **Payload:** Standard JSON or CBOR.

### [Fallback] Satellite Path (NTN) **Low Bandwidth • High Cost** * **Active:** Out of cellular range (targeting **3GPP Rel-17 standard NTN** networks like Skylo). * **Protocol:** Raw UDP or CoAP (strictly connectionless). * **Security:** DTLS 1.2 or OSCORE to minimize overhead. * **Telemetry:** Immediate Edge AI anomaly alerts (e.g., theft/panic) + 4x daily batched status. * **Management:** Disabled (no FOTA). * **Payload:** Custom bitmapped binary (**< 50 bytes**).

Device Lifecycle: Security & Logistics

Zero-Port Provisioning & Secure Deployment Flow

### 1. Logistics & Onboarding * **Shipping State:** Devices must ship in a **deep sleep state** (zero power draw, < 5µA) to preserve battery. * **Zero Physical Ports & SIMs:** No USB or charging ports. **No mechanical SIM slots** (must use eSIM/iSIM) to prevent IP69K vibration/moisture failures. Activation via **magnetic reed switch** or **NFC**. * **Traceability:** Every Ear-Tag and Collar must have a **unique, persistent hardware ID** flashed at the factory.

### 2. Device Security * **Network Encryption:** * **Cellular Link:** TLS 1.3 encryption. * **NTN Link:** DTLS 1.2 or OSCORE. * **Keys Provisioning:** Flash-injected security certificates and LwM2M bootstrap credentials at factory. * **Target:** PSA Certified Level 2 hardware root-of-trust.

### 3. Fleet Management * **Node Management:** Collar acts as the local LwM2M proxy for the herd. * **FOTA Updates:** Allowed **only over the Cellular link**. * **Ear-Tag Updates:** Phase 0/1 tags are "burn-once" (no FOTA) to simplify the initial RF stack. Phase 2 mass-market tags must support Sub-GHz multicast FOTA distributed by the Collar.

Standards Compliance Matrix

Certification Targets for Mass Market

| Standard | Scope | Member tag | Leader Cowbell | |---|---|---|---| | **ISO 11784 / 11785** | Animal RFID identification (HDX/FDX) | **Required** | n/a | | **ICAR certification** | EID conformity | **Required** | n/a | | **FCC Part 15 / Regional** | 902–928 MHz unlicensed Sub-GHz | **Required** | **Required** | | **IP67 / IP69K** | Extreme ruggedization & washdown | **Required (IP67)** | **Required (IP69K)** | | **3GPP Rel-17 NTN** | Satellite NB-IoT/LTE-M | n/a | **Required** | | **SGP.32 v1.2** | iSIM zero-touch provisioning | Optional | **Required** | | **PSA Certified Level 2** | Edge cryptography root of trust | Strategic Target | Strategic Target |

layout: section

IP Boundaries & Partnership Roles

JAAB Provides Intelligence • Partner Provides Scale

### JAAB Scope **Application Logic & Cloud Analytics** * We retain full IP ownership of all application-level algorithms. * Edge AI behavioral models (lameness, estrus, theft). * **Cloud Platform:** We prefer to implement our own data & processing platform in the public cloud, but we are open to evaluating intermediary ODM IoT platforms if they offer significant time-to-market advantages.

### ODM Partner Scope **Compute Substrate & RF Expertise** * You provide the hardware muscle, RF/Antenna design, and high-volume ODM manufacturing. * We require flexibility in NRE, prototyping costs, and mass-production scaling.

The Phased Roadmap & Alignment Milestones

Sequential Validation and Co-Design

### Phase 0: Viability Pilot **RF & Power Validation** * Custom Electronic tags (Members). * **Portable Gateway:** Carried by a worker to rapidly validate Sub-GHz propagation through the herd. * **Success Metric:** Empirical validation of the 3-year battery budget and RF link limits to establish a realistic baseline for mass-market design.

### Phase 1: Custom Foundation **Software & System Validation** * Custom Member tags. * Off-The-Shelf (COTS) Leaders (e.g., modifying existing collars or building based on an existing product design). * Validate the Zephyr Edge AI, Dual-Path routing, and Cloud integration at herd scale (100 tags, 5 collars).

### Phase 2: Mass Market **Extreme BOM Reduction & Scale** * Custom Cowbell Carrier mechanicals. * Extreme cost-down on Member tags (Standard high-volume silicon or conditional ASIC). * Mass-market production and commercial rollout based on the validated Phase 1 architecture.

layout: section

The Co-Design Phase

Moving from Strategy to Execution

### 1. Architectural Baseline The physical constraints, RF challenges, and $50 TCO economic limits of the 97/3 asymmetric architecture serve as the foundational baseline for all hardware engineering decisions.

### 2. Joint Solution Design The objective of this phase is to leverage existing ODM products, component partnerships, and mass-manufacturing expertise to collaboratively adapt and accelerate the Phase 0 hardware stack.

Building the future of infrastructure-zero traceability.
andres@jaab.tech