1GREEN ENERGY 5 PS
GREEN ENERGY Problem Statements
GE-2: Solar Generation Forecasting + Load Scheduling for Small Facilities
Problem: Small businesses with solar panels waste energy due to poor planning and lack of forecasting.
Build a solution that forecasts solar generation and suggests load scheduling (AC, pumps, machines) to maximize self-consumption.
Expected Output: Forecast model + scheduling recommendations + savings estimate.
Real-world angle: Makes solar adoption financially stronger.
GE-3: Carbon Footprint Estimator for SMEs With Actionable Reduction Plan
Problem: SMEs want to reduce carbon footprint but lack a simple tool that converts operations into measurable actions.
Build a platform that estimates emissions from common activities (electricity, travel, logistics) and generates reduction steps with ROI.
Expected Output: Footprint report + reduction roadmap + tracking.
Real-world angle: Practical sustainability planning.
GE-5: Community EV Charging Load Balancer (Fairness + Peak Control)
Problem: EV charging clusters create peak load issues and unfair access during rush hours.
Build a system that schedules charging slots fairly, reduces peak stress, and prioritizes emergency needs.
Expected Output: Slot allocator + peak load limiter + fairness rules.
Real-world angle: Grid-friendly EV scaling.
GE-1: Smart Energy Theft & Loss Detection for Community Grids
Problem: Many local distribution networks suffer from high energy losses due to theft, faulty meters, and untracked load spikes.
Build a system that detects anomalies in energy consumption patterns and flags suspicious activity using limited or incomplete data.
Expected Output: Anomaly detection + alerts + explanation dashboard.
Real-world angle: Reduces losses, improves grid efficiency.
GE-4: Predictive Maintenance for Wind/Solar Assets Using Low-Cost Signals
Problem: Renewable assets fail due to lack of monitoring and delayed maintenance.
Build a system that predicts faults using minimal signals (temperature, vibration proxy, power output trends).
Expected Output: Failure prediction + maintenance alerts + confidence score.
Real-world angle: Reduces downtime, improves reliability.
2AGRITECH 5 PS
AGRITECH Problem Statements
AG-3: Farm-to-Market Price Intelligence With Trustworthy Signals
Problem: Farmers get exploited due to lack of transparent pricing and market trend awareness.
Build a tool that aggregates pricing signals, identifies fair ranges, and warns about abnormal price drops.
Expected Output: Price tracker + fair range + anomaly alerts.
Real-world angle: Helps farmers negotiate better
AG-4: Supply Chain Spoilage Risk Predictor for Perishables
Problem: Fruits/vegetables spoil due to delays, poor storage, and lack of early risk detection.
Build a system that predicts spoilage risk based on time, temperature proxy, distance, and handling stages.
Expected Output: Risk score + preventive actions + escalation triggers.
Real-world angle: Reduces waste, increases profit.
AG-5: Pest/Disease Early Warning System With Human Verification
Problem: Early symptoms are ignored until crop damage becomes irreversible.
Build a platform that allows farmers to report symptoms and get risk prediction with a human verification workflow.
Expected Output: Report flow + risk prediction + verified advisory.
Real-world angle: Prevents large-scale losses.
AG-1: Crop Advisory for Low-Internet Farmers (Offline-First)
Problem: Farmers need timely advice but face low connectivity and limited digital literacy.
Build a solution that provides crop guidance offline and syncs when connectivity returns.
Expected Output: Offline advisory + sync engine + local language support (optional).
Real-world angle: Designed for rural conditions.
AG-2: Smart Irrigation Misuse Detector (Water + Power Waste)
Problem: Over-irrigation wastes water and electricity, reducing yield quality.
Build a system that detects irrigation misuse patterns and suggests corrective actions using simple inputs (time, weather, crop stage).
Expected Output: Misuse detection + recommendations + alerts.
Real-world angle: Saves water, improves yield.
3 HEALTHTECH 5 PS
HEALTHTECH Problem Statement
HT-1: Patient Queue & Triage Optimizer for High-Load Clinics
Problem: Clinics face overcrowding and delayed triage, impacting patient outcomes.
Build a system that optimizes queue flow and prioritizes cases based on symptoms and urgency.
Expected Output: Triage scoring + queue dashboard + alerts.
Real-world angle: Better care delivery with limited staff.
HT-2: Medication Adherence Tracker With Behavioral Nudges
Problem: Patients skip medicines due to forgetfulness, side effects, or poor understanding.
Build a tool that tracks adherence and provides nudges, reminders, and escalation to caregivers.
Expected Output: Reminder system + adherence log + risk flags.
Real-world angle: Improves treatment success.
HT-3: Privacy-Safe Health Record Summary Generator
Problem: Doctors waste time reading long case histories; patients don’t understand their own records.
Build a solution that summarizes health records into doctor-ready and patient-friendly views while avoiding sensitive data misuse.
Expected Output: Summary engine + redaction rules + audit log.
Real-world angle: Faster and safer consultations.
HT-4: Public Health Outbreak Signal Detection From Local Reports
Problem: Early outbreak signals are missed because data is fragmented and unstructured.
Build a system that detects outbreak patterns from community reports and generates actionable alerts.
Expected Output: Trend detection + heatmap + alert thresholds.
Real-world angle: Early intervention support.
HT-5: Mental Health Support Navigator for Students/Young Professionals
Problem: Many people don’t know where to start for mental health support, and stigma delays action.
Build a tool that guides users to appropriate support resources with safe escalation workflows.
Expected Output: Assessment + resource mapping + safety escalation.
Real-world angle: Responsible support, not diagnosis.
4DIGITAL PUBLIC INFRASTRUCTURE (DPI) 5 PS
DIGITAL PUBLIC INFRASTRUCTURE (DPI) Problem Statements
DPI-1: Citizen Grievance Routing System With SLA Tracking
Problem: Citizen complaints get lost due to poor routing and no SLA accountability.
Build a platform that classifies complaints, routes them to the right department, and tracks resolution timelines.
Expected Output: Routing engine + SLA dashboard + escalation rules.
Real-world angle: Governance accountability.
DPI-2: Fraud & Duplicate Detection in Public Benefit Applications
Problem: Benefit schemes face leakage due to duplicate applications and fraud patterns.
Build a system that flags suspicious entries while ensuring fairness and minimizing false rejections.
Expected Output: Risk scoring + explainability + manual review workflow.
Real-world angle: Protects public funds.
DPI-3: Digital Document Verification & Trust Layer for Institutions
Problem: Fake certificates and forged documents create serious trust issues.
Build a system that verifies documents using digital signatures/QR validation and tamper detection logic.
Expected Output: Verification flow + audit log + trust score.
Real-world angle: Faster and safer verification.
DPI-4: Emergency Resource Allocation Dashboard (Disaster Response)
Problem: During emergencies, resource allocation fails due to lack of real-time visibility and coordination.
Build a dashboard that tracks requests, supply availability, and dispatch prioritization.
Expected Output: Allocation logic + priority rules + tracking.
Real-world angle: Faster response, fewer gaps.
DPI-5: Consent-Based Data Sharing for Citizen Services
Problem: Citizens repeatedly submit the same documents to multiple departments, increasing friction and leakage risk.
Build a consent-based data-sharing system that allows controlled sharing with clear access logs.
Expected Output: Consent flow + access logs + expiry controls.
Real-world angle: Trust-first DPI design.
5SECURITY4HER 5 PS
SECURITY4HER Problem Statement
S4H-1: Ride-Share "Intent" Detection (Anti-Kidnapping & Deviation Monitoring)
Problem: Women in ride-shares often realize they are being kidnapped only after the driver has already deviated significantly or reached a secluded area, making it too late to call for help.
Build a solution Design a real-time background service that monitors GPS data against expected routes and driver behavior (stops/speed) to autonomously detect kidnapping intent without requiring manual user intervention.
Expected Output: A mobile application module that triggers a "Silent Dispatch" of live coordinates to authorities if the vehicle deviates from the safe corridor or stops in a high-risk zone for too long.
Real world angle: This shifts safety tech from "panic buttons" (which require the victim to act) to "passive monitoring" (which works even if the victim is frozen in fear or threatened).
S4H-2: Smart Emergency Dispatch (Voice Stress Analysis)
Problem: Emergency helplines (like 911/100/112) often receive thousands of calls. Operators struggle to differentiate between a "prank call," a "low-priority dispute," and a "life-threatening assault" instantly, leading to delayed police response for critical cases.
Build a solution Design an AI-layer for emergency dispatch centers that analyzes the audio acoustics (pitch, jitter, background noise) of the incoming call in real-time to detect "high-stress markers" (screaming, trembling voice, struggle sounds).
Expected Output: A dashboard for police dispatchers that auto-prioritizes calls, flagging the ones with "High-Threat Audio Signatures" to the top of the queue so immediate help is sent, even if the caller hangs up or cannot speak clearly.
Real world angle: This optimizes the system rather than just the user app, ensuring that limited police resources are sent to the women who are in the most immediate physical danger.
S4H-3: Workplace Harassment Reporting With Anonymous Escalation
Problem: Many harassment cases go unreported due to retaliation fear.
Build a platform that enables safe reporting, anonymous escalation, and structured resolution tracking.
Expected Output: Reporting flow + case tracking + role-based access.
Real-world angle: Institutional accountability.
S4H-4: Behavioral Distress Recognition System
Problem: Public surveillance cameras are currently passive recording devices that only provide evidence after a crime occurs, failing to identify or stop harassment and stalking in real-time.
Build a solution Develop a computer vision system that analyzes live camera feeds to detect specific "pre-crime" body language patterns like persistent following (stalking), surrounding a victim, or sudden flight responses.
Expected Output: A security dashboard that alerts monitoring personnel with a highlighted video clip immediately when a specific distress behavior is detected, enabling intervention before physical contact is made.
Real world angle: This transforms city cameras from "silent witnesses" into "active guardians" that can spot trouble in dense crowds where human security guards might miss the signs.
S4H-5: Digital Safety Companion for Online Harassment & Scam Defense
Problem: Women face increasing online harassment, impersonation, and scams.
Build a tool that detects risky patterns, provides guidance, and enables reporting workflows.
Expected Output: Risk classification + guidance + report templates.
Real-world angle: Realistic and scalable digital safety.
