Current Status
RADMAP (AINST) is a real-time AI-powered Radiation Mapping system built on Raspberry Pi 5, integrating LiDAR, Gamma Spectrometer, GPS, Temperature & Humidity sensors, and OpenCV. It generates live 2D radiation heatmaps, 3D SLAM-based Digital Twin visualization, and real-time telemetry dashboards using Google APIs.
Problem
Conventional radiation monitoring systems provide limited spatial awareness and lack real-time visualization, making it difficult to accurately detect, monitor, and analyze radiation distribution in dynamic environments.
The objective was to develop an edge-based, AI-assisted monitoring platform capable of continuous multi-sensor data acquisition, geospatial mapping, and Digital Twin visualization for environmental radiation analysis.
Technical Decisions
1. Raspberry Pi 5 Edge Computing
Built the complete sensor acquisition pipeline on Raspberry Pi 5 for low-latency processing and real-time telemetry.
2. Multi-Sensor Integration
Integrated Gamma Spectrometer, LiDAR, GPS, Temperature, Humidity, and OpenCV camera into a unified embedded monitoring framework for comprehensive environmental monitoring.
3. SLAM + Digital Twin
Implemented SLAM-based localization to generate synchronized 3D Digital Twin visualization from live environmental data.
4. 2D & 3D Radiation Mapping
Developed dynamic radiation heatmaps and geospatial visualization using Google Maps and real-time interpolation algorithms.
5. Google Live Dashboard
Designed a real-time monitoring dashboard displaying radiation intensity, GPS coordinates, environmental parameters, live camera feed, and Digital Twin analytics.
Why?
The project was developed to explore AI, Embedded Systems, Computer Vision, and Geospatial Intelligence for intelligent environmental radiation monitoring, enabling real-time situational awareness through Digital Twin technology.
Objective About This Project
Develop a scalable, AI-powered radiation monitoring platform capable of integrating heterogeneous sensors, generating real-time 2D/3D radiation maps, and visualizing environmental data through an interactive Digital Twin dashboard.
Challenges & Mistakes Encountered
Gamma Spectrometer Communication
What Happened
The detector powered successfully but failed to transmit valid radiation packets.
Initial Assumption
The issue was assumed to be a hardware fault.
Investigation
Serial communication logs revealed protocol incompatibilities between the detector firmware and Raspberry Pi drivers.
Root Cause
Device communication mismatch and unsupported serial protocol configuration.
Raspberry Pi 5 Configuration was used to create and implement the required communication layer.
Solution | Lesson
Implemented custom serial communication handling, optimized USB driver configuration, and validated stable telemetry streaming.
Real-Time SLAM Synchronization
What Happened
Live camera frames, LiDAR scans, and GPS coordinates became unsynchronized during continuous movement.
Initial Assumption
Processing latency was believed to be caused by rendering overhead.
Investigation
Profiling identified asynchronous sensor acquisition and timestamp inconsistencies.
Root Cause
Independent sensor refresh rates introduced coordinate drift and unstable Digital Twin rendering.
SLAM 3D Twinning & Stitching was implemented to address this issue.
Solution | Lesson
Implemented multithreaded data synchronization, timestamp alignment, and buffered sensor fusion for stable real-time 3D mapping.
Gamma Spectrometer Communication Failure
What Happened
The Gamma Spectrometer initialized successfully but failed to stream valid radiation telemetry to the Raspberry Pi, preventing real-time radiation visualization on the dashboard.
A local terminal COM port was used to develop an independent interface for 2D/3D heatmap visualization using radiation peaks and counts.
Investigation
Verified serial communication, analyzed USB-to-Serial logs, tested baud rate configurations, and monitored device initialization sequences to isolate communication failures.
Root Cause
The detector's native communication protocol was incompatible with the default Raspberry Pi serial configuration, causing packet loss and incomplete telemetry.
Fix
Configured custom serial communication parameters, optimized USB driver settings, synchronized packet parsing, and established a stable telemetry pipeline for continuous radiation data acquisition.
Lesson
Reliable embedded systems require protocol validation, hardware compatibility testing, and robust serial communication handling before integrating higher-level visualization components.
SLAM & Multi-Sensor Synchronization Issue
Problem
During real-time operation, LiDAR scans, GPS coordinates, live camera frames, and radiation measurements became unsynchronized, resulting in inaccurate Digital Twin rendering and unstable heatmap overlays.
A custom multi-channel CNSPEC MCA Gamma Spectrometer 2D radiation heatmap workflow was developed because the required workflow was not directly supported by the target operating-system configuration.
First Solution
Implemented buffered sensor queues and timestamp-based synchronization, which reduced frame inconsistencies but introduced latency under continuous data streams.
Errors were encountered during integration of high-intensity Gamma Spectrometer radiation counts into the heatmap visualization.
Real-time CSV generation and dashboard import solved the initial synchronization problem.
New Edge Case
High-frequency LiDAR updates and camera frame rates produced coordinate drift during servo motor rotation, causing temporary misalignment in 3D mapping.
Final Solution
Designed a multithreaded sensor fusion pipeline with synchronized timestamps, optimized SLAM updates, and adaptive buffering to maintain accurate real-time Digital Twin visualization.
The main dashboard now integrates Google APIs to generate 2D and 3D Digital Twin visualizations with stitching and intensity-based radiation spots.
Lesson
Real-time Digital Twin systems depend on precise sensor synchronization, efficient multithreading, and low-latency data fusion to achieve reliable spatial mapping and visualization.
Limitations
1. Prototype-Based Hardware
The system is designed as a research prototype and has not been validated for industrial-scale or hazardous radiation environments.
2. Sensor Calibration Dependency
Radiation mapping accuracy depends on proper calibration of the Gamma Spectrometer, LiDAR, GPS, and environmental sensors.
Integrating combined data from all sensor readings remains challenging.
3. Raspberry Pi 5 Processing Constraints
Real-time SLAM, computer vision, and multi-sensor processing are limited by Raspberry Pi 5 computational resources during high-frequency data acquisition.
4. Limited Radiation Validation
The prototype was tested using controlled materials and simulated datasets; large-scale field validation across diverse radiation sources is still pending.
5. Indoor Localization Challenges
GPS accuracy decreases indoors or in obstructed environments, which may introduce minor localization errors in Digital Twin visualization.
Impact
RADMAP demonstrates the integration of AI, Embedded Systems, Computer Vision, IoT, and Digital Twin technologies into a unified real-time environmental radiation monitoring platform.
It enables live multi-sensor telemetry, 2D/3D radiation visualization, and intelligent geospatial analytics for research and environmental safety applications.
The project significantly strengthened my expertise in embedded systems, sensor fusion, SLAM (Simultaneous Localization and Mapping), edge computing, backend engineering, and real-time data visualization while providing hands-on R&D experience in intelligent monitoring systems.
Future Vision
1. Autonomous Robotic Deployment
Deploy RADMAP on autonomous ground robots or drones for remote radiation monitoring in hazardous environments.
2. AI-Based Radiation Prediction & Dynamic Localization
Integrate Machine Learning models to predict radiation hotspots and detect environmental anomalies from historical sensor data.
3. Cloud-Native 3D Digital Twin
Extend the platform using Docker, Kubernetes, MQTT, and cloud infrastructure for scalable multi-device monitoring.
4. Advanced Sensor Fusion & Integration
Enhance mapping accuracy by integrating IMU, RTK-GPS, and high-resolution LiDAR with advanced SLAM algorithms.
5. Industrial-Scale Monitoring
Transform the research prototype into a production-ready environmental monitoring system with real-time alerts, analytics, mobile dashboards, and GIS-based monitoring for nuclear facilities and smart cities.