SATAP( Secure, AI-Driven Threat Assessment Platfrom)

Cover Page
Project Name
Theme
Proposed Solution & Architecture
Innovation & Social/Market Impact
Architecture (Flowchart, Techstack and Prototype)
Thank You (Team Info)
Raw Data Input and Map Visualization
Secure Blockchain Ledger

Project Story: SATAP

Inspiration

SATAP was born from a simple but serious question: What happens when critical military decisions depend on slow, manual analysis of overwhelming data?

In modern warfare, especially information warfare, commanders are flooded with unstructured intelligence—field reports, intercepted communications, situation logs—often under extreme time pressure. Manual analysis not only slows response time but also creates room for uncertainty, bias, and lack of accountability.

Our motivation was not just to make intelligence analysis faster, but more trustworthy. We wanted to build a system that ensures every decision can be **explained, verified, and audited—what we define as true accountability (A²: Accountability + Auditability).

What SATAP Does

SATAP (Secure, AI-Driven Threat Assessment Platform) is a unified intelligence-support platform with two core capabilities:

Rapid Threat Assessment (AI/ML)

SATAP ingests unstructured text reports such as field logs or intercepted communications and processes them using NLP techniques. The system generates a numeric Threat Score (Sₜₕᵣₑₐₜ) that categorizes reports into LOW, MEDIUM, or HIGH risk, allowing commanders to instantly prioritize attention where it matters most.

Verifiable Intelligence Audit (Blockchain)

Every final decision made by a human analyst—whether to Escalate or Dismiss a report—is permanently recorded on an immutable ledger. This prevents retrospective manipulation and guarantees a transparent decision trail for future audits or investigations.

How We Built It

We designed SATAP using a modular three-pillar architecture:

Pillar 1: AI Threat Engine

Built using Python with Scikit-learn
Applied TF-IDF vectorization to convert text reports into numerical features
Trained a Logistic Regression classifier on carefully crafted synthetic intelligence data
Outputs a final Threat Score (Sₜₕᵣₑₐₜ) for each report

Pillar 2: Secure Ledger

Implemented a custom lightweight blockchain in JavaScript / React
Each block is cryptographically linked using SHA-256 hashing:

[ \text{Hash}{n} = H(\text{Data}{n} + \text{Hash}_{n-1}) ]

Block data includes Analyst ID, Report ID, and Decision
Ensures tamper-proof and verifiable records

Pillar 3: Geo-Visualization Dashboard

Built using React.js, styled with **Tailwind CSS
Displays AI-generated threat scores in real time
Extracted location entities are plotted on a Leaflet.js interactive map for clear situational awareness

Challenges We Faced

Domain-Specific Data

Real military intelligence data is classified, so creating realistic synthetic training data was our biggest challenge. We manually crafted hundreds of intelligence-style reports using realistic jargon and operational scenarios to ensure meaningful AI training.

Cryptography Dependency Issues

We initially faced build and compatibility issues with an external cryptography library (crypto-js). This was resolved by switching to the native Web Crypto API (crypto.subtle.digest), improving reliability and reducing external dependencies.

Real-Time Synchronization

Coordinating seamless updates between report input, AI scoring, decision logging, and the dashboard required careful orchestration using Firebase Firestore real-time listeners.

Accomplishments We’re Proud Of

The “Trust Wrapper”: Successfully combining AI and blockchain to create a compact yet fully verifiable intelligence system Performance: Achieved end-to-end threat assessment in ≤ 5 seconds, meeting real-world operational constraints Cross-Domain Integration: Delivered a single solution spanning Defence Systems, AI/ML, Web3, and Cybersecurity

What We Learned

This project taught us that technology choices matter early—native, stable dependencies significantly improve long-term reliability.

More importantly, we learned that cryptography isn’t just for finance. When applied correctly, it becomes a powerful tool to build trust, accountability, and transparency in AI systems—transforming AI from a black-box assistant into a verifiable teammate.