Rizwan U. Farooqui
Cyber Infrastructure in Construction
Assistant Professor Rizwan U. Farooqui, Ph.D. (Co-PI) ,
Department of Building Construction Science
ConstructionCI: Integrating Cyber Infrastructure into Construction Management and Safety
Project Overview
The construction industry is rapidly evolving through the integration of cyberinfrastructure (CI) technologies such as Internet of Things (IoT), machine learning, robotics, edge computing, and cybersecurity systems. However, construction remains among the least digitized industries, largely due to limited workforce readiness rather than lack of technology.
The ConstructionCI initiative addresses this gap by embedding advanced, domain-specific CI training directly into Construction Management and Safety (CM&S) education. The project reframes computing technologies not as auxiliary tools, but as core components of modern construction practice. Its goal is to prepare students to lead innovation in smart, safe, and resilient built environments.
Objectives and Significance
ConstructionCI aims to develop and implement an interdisciplinary training framework that integrates four key CI domains into CM&S curricula: IoT-enabled sensing, machine learning and data analytics, robotics and automation, and cybersecurity in construction systems.
This effort is particularly important for Mississippi and surrounding states, where industry feedback indicates that technology adoption is constrained by a shortage of professionals trained to deploy and manage advanced systems. By strengthening CI competencies within construction education, the project supports regional workforce development while aligning CM&S programs with emerging national industry demands.
Approach
The program is structured around four integrated pillars:
- IoT in Construction: Training on sensing systems, networking protocols, and edge devices for applications such as structural health monitoring, smart buildings, and material tracking.
- Machine Learning and Edge Computing: Hands-on application of ML/DL models for safety monitoring, equipment recognition, and data-driven decision-making, including cloud and edge deployment.
- Robotics and Human鈥揜obot Interaction: Exposure to robotic platforms and programming environments to explore autonomous documentation, inspection, and site analysis.
- Cybersecurity in Smart Construction: Instruction on vulnerabilities in connected construction systems and practical strategies for securing IoT devices, ML models, and digital infrastructure.
Outcomes and Impact
ConstructionCI delivers structured workshops, curriculum integration, and open-access training materials to extend impact beyond a single cohort. The initiative strengthens collaboration between Building Construction Science and computing disciplines, fostering a multidisciplinary environment reflective of Construction 4.0 practice.
Expected impacts
- A future-ready construction workforce
- Increased regional adoption of advanced and secure technologies
- Enhanced competitiveness of CM&S graduates
- Expanded participation of underrepresented students in construction technology fields
Future Direction
ConstructionCI serves as a foundation for long-term curricular transformation and expanded interdisciplinary research. As construction becomes increasingly data-driven and automated, education must evolve accordingly. This initiative represents a strategic step toward aligning construction management education with the technological realities shaping the next generation of the built environment.
Real-Time Proximity Alert System Using Monocular Depth Estimation and YOLO
Project Overview
鈥淪truck-by-equipment鈥 incidents remain a leading cause of fatalities in construction. While advances in computer vision have improved object detection and depth estimation independently, integrating these capabilities into a deployable, real-time safety system remains a challenge.
This project presents a unified, vision-based framework that combines YOLOv11 object detection with RT-MonoDepth monocular depth estimation to monitor proximity hazards using a single standard camera. The system converts ordinary 2D video into actionable 3D spatial intelligence, enabling real-time metric distance measurement between workers and machinery without requiring LiDAR or stereo camera systems.
Objective and Significance
The primary objective is to develop a cost-effective, scalable safety monitoring tool capable of:
-
Detecting workers, vehicles, and heavy equipment in real time
-
Estimating true Euclidean distances between objects
-
Automatically identifying and logging proximity zone breaches
A key technical barrier in monocular depth estimation is scale ambiguity鈥攄epth predictions lack real-world units. This research addresses that limitation through an automatic geometric calibration method that uses detected personnel as anthropometric references. By resolving scale ambiguity dynamically, the system produces object-specific distance measurements in meters from an uncalibrated camera.
The significance of this work lies in lowering the cost and complexity of advanced safety monitoring. It offers a practical alternative to expensive hardware-based sensing systems while maintaining real-time performance.
Methodology
The framework operates through a parallel, multi-threaded processing architecture. YOLOv11 performs semantic detection of personnel and equipment, while RT-MonoDepth generates dense depth maps from the same video stream. A calibration module converts relative depth predictions into metric distances using geometric relationships derived from the pinhole camera model. The system reconstructs 3D coordinates and computes true Euclidean distances between detected objects. The framework achieved real-time performance at approximately 25 frames per second with sub-40 millisecond latency, confirming its suitability for edge deployment.
Future Direction
Future work will focus on deployment on embedded edge platforms, improving robustness under challenging site conditions such as low lighting and dust, and integrating automated alert mechanisms for real-time on-site notification. The long-term goal is to develop scalable, AI-enabled safety systems that can be widely adopted across construction environments.
BuildTrust: AI-Powered Intelligent Construction Dispute Resolution Using LLMs and Blockchain
Project Overview
Construction disputes are a major source of cost escalation and schedule delays. Claims frequently arise from inconsistencies across contracts, revisions, change orders, and project communications. Although documentation is increasingly digital, dispute resolution remains largely manual, time-consuming, and reactive. BuildTrust is an AI-enabled framework that integrates Large Language Models (LLMs) with blockchain-based traceability to support structured construction claim workflows. The system combines semantic contract analysis with tamper-resistant document registration, providing a transparent and technology-driven approach to dispute documentation and assessment.
Objectives and Significance
The primary objective of BuildTrust is to demonstrate the feasibility of integrating AI-driven document intelligence with blockchain-backed auditability in construction claim management.
Specifically, the project aims to:
- Automate extraction of key contractual entities (parties, dates, clauses)
- Detect meaningful modifications between contract versions
- Classify claims into Delay, Cost, or Scope categories
- Provide structured summarization of dispute narratives
- Ensure tamper-evident documentation through blockchain hashing
The significance of this work lies in bridging two emerging domains鈥擜I-based semantic reasoning and decentralized audit systems鈥攖o enhance transparency, efficiency, and trust in construction dispute processes.
Methodology
BuildTrust follows a three-layer architecture:
- Document Ingestion: Uploaded contracts are converted to text and assigned a SHA-256 cryptographic hash.
- AI-Based Semantic Analysis: A locally deployed LLM performs entity extraction, contract comparison, claim classification, and summarization using prompt-based reasoning.
- Blockchain Registration: A Solidity smart contract records document hashes and claim state transitions on a permissioned blockchain, ensuring immutable traceability without storing full documents on-chain.
This integrated pipeline connects semantic interpretation with verifiable workflow logging in a unified system.
Key Outcomes
Experimental validation using synthetic AIA-style contracts demonstrated:
- 75% accuracy in entity extraction under zero-shot prompting
- Up to 100% detection of introduced contractual modifications
- Correct classification of representative delay and cost claims
- Reliable hash-based tamper detection
- Successful blockchain-based document registration
The results confirm the technical feasibility of combining AI-based document intelligence with blockchain-backed traceability for construction dispute workflows.
BuildTrust provides a foundation for AI-enabled, transparent, and structured dispute management systems that can reduce administrative burden and improve accountability.
Future Directions
Future work will focus on testing the framework using real project documentation, improving AI accuracy through domain adaptation or retrieval augmentation, and deploying the system within distributed consortium blockchain environments. The long-term goal is scalable, AI-driven dispute intelligence systems that enhance trust and efficiency in construction project administration.
Secure Carbon: Blockchain-Based Integrity Framework for Construction Carbon Tracking
Project Overview
Carbon emission reporting is increasingly embedded in regulatory, procurement, and ESG disclosure frameworks that directly affect the construction industry. Governments and public infrastructure clients are requiring transparent documentation of embodied and construction-phase emissions. While established standards exist for carbon calculation, the management and verification of supporting carbon documentation remain largely centralized and document-based, creating risks related to data modification, version control, and trust across stakeholders.
This project presents a blockchain-based framework designed to strengthen the integrity and auditability of carbon data in construction. Rather than altering how emissions are calculated, the framework introduces a secure layer for recording and verifying carbon documentation through cryptographic hashing and smart contracts. The work was developed as a prototype and submitted to the Canadian Society for Civil Engineering (CSCE) 2026 Conference, Qu茅bec City
Objectives and Significance
The primary objective of this research is to demonstrate how blockchain technology can support immutable, time-stamped carbon documentation in multi-stakeholder construction environments. Specifically, the project aims to:
- Anchor off-chain carbon documentation to an immutable blockchain ledger
- Enable role-based submission and approval of carbon records
- Improve traceability and reduce risk of post-submission modification
- Enhance transparency in sustainability reporting workflows
The significance of this work lies in addressing governance challenges in carbon reporting. In construction projects involving contractors, owners, and auditors, trust is often procedural rather than technically enforced. This framework introduces a verifiable digital mechanism that supports confidence in reported carbon data without replacing existing accounting standards.
Methods
The proposed framework separates carbon data generation from blockchain-based record keeping.
Carbon documentation is generated off-chain using conventional practices (e.g., material quantities, EPD references, emission factors). A SHA-256 cryptographic hash of the documentation is then computed and submitted to a Solidity-based smart contract deployed in a local Ethereum environment.
The smart contract stores:
- Activity metadata
- Submission timestamp
- Submitter identity
- Cryptographic hash of the carbon document
- Approval status
Only the document hash鈥攏ot the full carbon file鈥攊s stored on-chain. This design preserves confidentiality while enabling tamper-evident verification. Role-based functions allow a contractor to submit records and an auditor to approve them, ensuring traceable governance.
Key Outcomes and Impact
Prototype testing demonstrated that:
- Carbon documentation can be securely anchored to an immutable blockchain record
- Any modification to off-chain documentation results in hash mismatch
- Role-based approval mechanisms function as intended
- Records cannot be altered or deleted after submission
This approach enhances data integrity and auditability while maintaining compatibility with existing carbon accounting methods. The framework offers a practical pathway toward more transparent and verifiable sustainability reporting in construction.
Future Directions
Future research will focus on automation of data transfer between off-chain documentation and blockchain smart contracts, integration with BIM and sensing technologies, and evaluation under multi-project deployment conditions. The long-term goal is to develop scalable, governance-focused digital infrastructure that strengthens trust and accountability in construction carbon reporting.
Real-Time Low-Power Wearable System for Physiological and Motion Monitoring in Construction Safety
Project Overview
The construction industry remains one of the most hazardous occupational sectors, with elevated injury and fatality rates driven by falls, fatigue, heat stress, and unsafe movements. Although wearable technologies have gained attention as safety tools, many existing systems rely on local data storage and post-event analysis, limiting their ability to support immediate intervention.
This project presents the design and validation of a custom, low-power wearable device capable of continuously streaming physiological and motion data in real time. The prototype integrates heart rate and blood oxygen sensing, skin temperature monitoring, and six-axis motion tracking within a compact printed circuit board (PCB) platform. By enabling live access to multi-sensor data through Bluetooth Low Energy (BLE), the system establishes a foundation for proactive, data-driven safety monitoring in construction environments.
Objectives and Significance
The primary objective of this research is to develop and evaluate a compact, energy-efficient wearable platform that enables synchronized real-time monitoring of both physiological and motion signals.
Specifically, the project aims to:
- Integrate multi-sensor physiological and motion monitoring on a single custom PCB
- Enable stable, low-latency BLE-based real-time data transmission
- Demonstrate feasibility for extended work-shift operation
The significance of this work lies in bridging physiological stress indicators and movement characteristics within a unified system. Fatigue, heat strain, and prolonged physical workload can compromise balance and increase the risk of slips and falls. A device capable of continuously monitoring both internal physiological responses and external motion patterns creates opportunities for early detection and timely intervention before injuries occur.
Methods
The wearable system integrates:
- Optical photoplethysmography (PPG) sensing for heart rate and SpO鈧
- Infrared temperature sensing for skin temperature
- Six-axis inertial measurement for motion tracking
- A Bluetooth Low Energy鈥揺nabled microcontroller for wireless streaming
The two-layer PCB measures approximately 30 脳 38 mm and is powered by a 3.7 V, 200 mAh lithium-ion battery. Firmware was developed to synchronize sensor sampling and structure BLE packets for stable transmission. A custom receiver application was created to decode, visualize, and log incoming data in real time. Preliminary validation compared physiological readings with a commercially available wearable device under controlled indoor conditions.
Key Outcomes and Impact
Testing demonstrated:
- Reliable real-time BLE streaming within a practical site range
- Stable heart rate and SpO鈧 readings under static conditions
- Successful multi-axis motion data transmission without buffering loss
- Average current draw of approximately 16 mA, supporting over 10 hours of continuous operation
These results confirm the feasibility of a compact, low-power wearable suitable for full-shift physiological and motion monitoring. The integrated sensing platform provides a technical foundation for future predictive analytics linking fatigue, stress, and unsafe movement patterns to proactive safety management strategies.
Future Directions
Future work will focus on secure wireless transmission, enclosure refinement for durability and comfort, optimized sensor placement, and multi-device fleet deployment capability. In collaboration with interdisciplinary partners, subsequent phases will develop predictive models that use real-time physiological and motion signals to anticipate fatigue and unsafe conditions, advancing toward proactive, data-driven construction safety systems.
