Smart Door Intrusion Detection & Notification System with Scheduled Monitoring for Big Stores

A US retail chain needed modern security to prevent after-hours intrusions; we built an AI-powered edge-cloud system delivering accurate real-time alerts.

< 2 Minutes

Incident response time

100%

off-hours intrusion detection

2–4.5 Seconds

Alert delivery latency

Overview

A major US retail chain needed a modern intrusion detection system for repeated after-hours security breaches. Legacy cameras lacked real-time alerts, smart triggers, and contextual evidence. RigRex built a hybrid edge-cloud security platform using sensors, AI cameras, AWS automation, and a React dashboard to detect intrusions, reduce false alerts, and notify teams quickly.

Research

The research focused on store security gaps, alert reliability, response speed, and edge-cloud system requirements.

Research Area	Key Finding
Legacy Security Systems	Existing cameras lacked real-time triggers and alerts.
After-Hours Intrusions	Stores needed automated detection outside business hours.
Manual Processes	Manual arming and disarming caused unreliable enforcement.
Alert Fatigue	Staff received unnecessary alerts during working hours.
Visual Verification	Teams needed image evidence with each intrusion alert.
System Performance	Alerts had to be delivered in under 5 seconds.
Privacy Requirements	Face detection needed local processing to reduce privacy risk.

The research showed the need for a fast, automated, and privacy-aware security system that could reduce noise while improving response time.

Challenges → Solutions

The project addressed security, operational, and technical challenges across multiple retail branches.

Challenge	Solution
Repeated after-hours intrusions	Built automated intrusion detection using sensors and AI cameras.
No real-time alerting	Used AWS IoT Core, Lambda, and Slack alerts for fast notifications.
Manual system activation	Added AWS Scheduler for automatic arming and disarming.
Alert fatigue during working hours	Triggered alerts only during active security windows.
Delayed incident response	Reduced response time from about 30 minutes to under 2 minutes.
Need for visual evidence	Captured images around trigger events and attached them to alerts.
Sensor and camera race conditions	Built coordinated edge logic with circular frame buffering.
Privacy concerns	Processed face detection locally on Raspberry Pi.
Centralized monitoring gap	Built a React admin dashboard for schedules, logs, and store settings.

These solutions created a reliable intrusion detection workflow with faster alerts, fewer false positives, and stronger operational control.

Conclusion

This project shows how an edge-cloud security system can improve retail intrusion response through automation, local AI processing, secure cloud infrastructure, and real-time alerts. The system helped DES move from delayed manual response to proactive detection, allowing store teams to respond within seconds instead of minutes.