CatchSmart Indoor Localization - Enterprise IoT Tracking System
Khaled built a complete end-to-end multi-platform indoor localization system solo, solving PATA Group's multi-million Euro 'blue wood' timber tracking problem. The system includes iOS and Android mobile apps, a Java microservices backend, a React web portal with custom factory maps, and an Unreal Engine 4 3D visualization application. Deployed across factory plants, warehouses, and workshops, the system tracks 5000+ workers and 1000+ expensive spare parts daily in harsh industrial metal environments, with custom signal processing algorithms overcoming BLE signal conduction challenges.
Problem Solved
PATA Group was losing millions of Euros annually due to timber containers being mishandled during export, causing wood to turn blue. Without tracking individual workers and materials through the manufacturing process, they could not identify quality assurance gaps in the complex multi-facility operation.
- •Architected and built complete end-to-end system solo across 5 platforms (iOS, Android, Web, Backend, 3D)
- •Designed custom signal processing algorithms for harsh metal factory environments
- •Deployed and configured 3000 dual RFID/BLE beacons in 10-day marathon
- •Built microservices architecture with zone-based horizontal auto-scaling on AWS
- •Created Unreal Engine 4 3D factory visualization with real-time beacon updates
- •Developed cross-platform code sharing using J2ObjC translation
- ✓Built entire multi-platform system solo: iOS, Android, Java backend, React web portal, Unreal Engine 3D visualization
- ✓Deployed and configured 3000 dual RFID/BLE beacons in 10-day manual configuration marathon
- ✓Solved multi-million Euro 'blue wood' problem through real-time worker and timber tracking
- ✓Designed custom signal processing algorithms for harsh metal factory environments overcoming signal conduction through metal structures
- ✓Architected microservices backend with zone-based horizontal auto-scaling on AWS
- ✓System tracks 5000+ workers and 1000+ expensive spare parts daily across multiple facilities
- ✓System still in production 6+ years later (2018-2025+)
- ✓Received Rolls-Royce Finland CTO validation - pursued licensing for their factory plants
Performance
- • Real-time tracking across 3000 beacons
- • Hybrid edge + cloud architecture for massive data streams
Scale
- • 5000+ workers tracked daily
- • 1000+ spare parts tracked
- • 3000 beacons deployed
- • 6+ years in production
Technology Stack
Challenge
Metal factory structures conducted BLE signals like electrical wires, making beacon at the end of factory detectable at front via metal poles - RSSI distance calculations completely unreliable
Solution
Developed custom signal processing algorithms inspired by circuit theory from university education - software components mimicking electronic filtering circuits for noise reduction, combined with zone clustering and path prediction instead of pure RSSI-based positioning
Impact
Achieved accurate positioning despite hostile RF environment that would defeat standard BLE positioning approaches
Challenge
3000 beacons multiplied by multiple users generating thousands of readings per second, overwhelming backend capacity
Solution
Implemented hybrid edge + cloud architecture with smart filter layers on mobile devices for quick filtering before sending to backend, dramatically reducing network and server load
Impact
Achieved scalable system capable of handling enterprise-level beacon data streams
Challenge
Different Android BLE chip manufacturers produced inconsistent signal readings, making cross-device accuracy impossible
Solution
Standardized single Android device model for all 5000+ workers to ensure uniform BLE hardware across entire workforce
Impact
Achieved consistent positioning accuracy across all Android devices
Challenge
Factory internet dropped frequently, breaking real-time WebSocket connections
Solution
Implemented automatic WebSocket to polling fallback with client-side path prediction based on historical data, self-correcting visualization when reconnected
Impact
Maintained smooth user experience despite unreliable connectivity
Challenge
Physically configuring 3000 beacons across multiple factory facilities with precise zone assignments
Solution
Executed 10-day manual reconfiguration marathon using RFID reader/writer, tracking each beacon's MAC address, UUID, location, and zone assignment in Excel
Impact
Successfully deployed complete beacon infrastructure enabling system launch
Situation
PATA Group, a major European timber manufacturer, was losing millions of Euros annually from timber containers turning blue during export to North Korea due to mistreatment or extreme cold. Without visibility into which workers handled which materials at which stages, they could not identify where quality assurance was breaking down. CatchSmart was tasked with solving this multi-million Euro tracking problem across their sprawling factory plants, warehouses, and workshops.
Task
Khaled was assigned to build a complete indoor localization system that would track 5000+ workers and valuable assets in real-time across multiple industrial facilities. This required not just a mobile app, but an entire ecosystem: iOS and Android apps for workers, a backend to process massive beacon data streams, a web portal for managers, and a 3D visualization tool for executives - all built solo in a 6-month timeline.
Action
Khaled architected and built the entire system single-handedly across five platforms. He deployed 3000 dual RFID/BLE beacons in a grueling 10-day configuration marathon. When standard BLE positioning failed due to metal structures conducting signals unpredictably, he developed custom signal processing algorithms inspired by circuit theory from his university education - treating the software like electronic filtering circuits. He designed a hybrid edge + cloud architecture where mobile devices performed initial filtering to handle the massive data streams. He built an Unreal Engine 4 3D visualization by importing factory CAD models via Datasmith, allowing executives to see beacon positions update in real-time across a virtual factory. Every architectural decision had to work across all five platforms simultaneously - one wrong choice would cascade through the entire system.
Result
The system successfully tracked 5000+ workers and 1000+ expensive spare parts daily, solving the blue wood problem by providing complete visibility into material handling. The system has remained in production for over six years (2018-2025+), demonstrating its reliability and value. The solution was so impressive that Rolls-Royce Finland's CTO pursued licensing it for their own factory plants. The project transformed Khaled from an iOS developer into a full-stack IoT systems architect capable of solo multi-platform delivery at industrial scale.
Technical
- • Full-stack multi-platform development across iOS, Android, Web, Backend, and 3D
- • Industrial IoT signal processing in hostile RF environments
- • Microservices architecture with horizontal auto-scaling
- • Apache Kafka for high-velocity data streams
- • Unreal Engine 4 C++ development and CAD integration
- • Cross-platform code sharing with J2ObjC
Soft Skills
- • End-to-end ownership of complex systems
- • Crisis problem-solving under extreme deadlines
- • Executive stakeholder communication
- • Systematic documentation for large-scale deployments
Key Insights
- 💡 True full-stack means understanding constraints from BLE hardware to cloud infrastructure
- 💡 University circuit theory became foundation for real-world signal filtering algorithms
- 💡 Edge computing on devices is essential for scaling IoT systems
- 💡 Different Android BLE chips produce different readings - hardware standardization may be only solution
- 💡 Testing at scale reveals problems no lab environment can predict