Code4Cure
SHIELD — Smart Health Initiative for Ecosystem & Livestock Defence
SHIELD combines a lightweight accelerometer collar, NFC identity, and AI analytics to detect estrus early, predict calving, and flag health anomalies — empowering farmers and veterinarians with timely action.
India’s livestock sector faces silent but significant losses due to missed reproductive events, fragmented records, and delayed field reporting. These gaps affect productivity, livelihoods, and national food security.
60%+ of disease cases are reported late or not at all because data is siloed in registers, calls, and local logs.
40–50% of heat cycles go unnoticed; missed heats delay conception by $\sim 21$ days, extending calving intervals and reducing yield.
Untreated metabolic issues and reproductive inefficiencies cost farmers thousands per animal and contribute to an estimated national loss of $\mathbf{\text{₹}30,000 \text{ crore}}$ annually.
A unified hardware + software ecosystem that captures behaviour, creates digital identity, and provides predictive insights — designed specifically for India’s dairy & livestock landscape.
Simulated data showing the target impact of SHIELD deployment.
The pipeline is intentionally simple, robust, and auditable: from sensor sampling to cloud analytics and farmer action.
Capture
Accelerometer samples at $\mathbf{50 \text{ Hz}}$Edge Process
ESP32 aggregates & timestamps dataTransmit
WiFi to cloud / gatewayStore
Timeseries Database ($\mathbf{\text{MySQL/Influx}}$)Analyze
AI computes estrus & anomaly scoresPer-animal baselines are computed from $\mathbf{7–14}$ days of data. Features extracted include: **Activity RMS** ($\mathbf{\text{Root Mean Square}}$ of motion), posture transitions, lying duration, entropy, and feeding-motion proxy. A scoring model combines these into interpretable outputs: Estrus Likelihood, Stress Index, and Calving Probability.
Robust, low-power, and farmer-friendly.
Handles sampling, local filtering, sleep management, and secure WiFi transmission.
High-precision 3-axis motion capture for activity, posture and feeding proxies.
Instant animal identity and access to vaccination and medical history on tap.
Designed for farmers, vets, and program managers with actionable cards & one-click reports.
Activity RMS values for Animal ID: 0045
Current estrus score relative to baseline.
Immediate status of the monitored herd.
From prototype to field pilots to scaled deployment.
Build $\mathbf{20}$ collars, test in controlled farm environment, validate estrus detection with vet logs (Target $\mathbf{>90\%}$ accuracy).
Deploy $\mathbf{200}$ collars across $\mathbf{10}$ farms, refine ML models with diverse data, iterate hardware design and power efficiency.
Partner with cooperatives & government, integrate with national surveillance systems, and scale manufacturing pipeline.
A multidisciplinary team of engineers, veterinarians, and data scientists.
Explore technical specifications and learn more about the team behind the project.