[{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/agentic-ai/","section":"Tags","summary":"","title":"Agentic AI","type":"tags"},{"content":" Demo Code This prototype presents an end-to-end agentic pipeline that converts raw LiDAR point clouds into explainable robot navigation decisions. The system ingests .pcd/.ply point cloud files alongside YAML scene metadata, and processes them through four orchestrated stages:\nPerception: An Open3D-based module segments the ground plane, clusters obstacle geometry, and produces a structured semantic scene summary including object positions, regions, and free-space estimates.\nRAG (Retrieval-Augmented Generation): A LangChain retrieval layer backed by a FAISS vector index queries an in-repo safety and planner knowledge base to surface relevant operational context for the current scene.\nState Extraction — MCP-style tool functions parse the YAML metadata to extract robot pose, ego speed, and the positions of nearby actors, feeding structured state into the decision layer.\nCoordinator: A rule-based coordinator fuses perception output, retrieved context, and robot state into a final navigation decision. When an OpenAI key is available, a LangGraph-orchestrated LLM-backed path is attempted first, with automatic fallback to rule-based logic.\nThe system outputs a risk level (Low / Medium / High), a recommended action, a scene assessment, and supporting evidence. A Streamlit front-end exposes an interactive 3D Plotly visualization of the scene alongside all intermediate outputs, making the pipeline fully inspectable at every stage.\nStack: Python · Open3D · LangGraph · LangChain · FAISS · OpenAI API · Streamlit · Plotly\n","date":"April 2026","externalUrl":null,"permalink":"/prototypes/scene-analyzer/","section":"Prototypes","summary":"An end-to-end agentic robotics pipeline that fuses 3D point-cloud perception (Open3D), RAG-based safety knowledge retrieval (LangChain + FAISS), and robot state extraction to produce explainable navigation decisions (Low/Medium/High risk + recommended action) from raw LiDAR data.","title":"Agentic Robot Scene Risk Analyzer","type":"prototypes"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/causality/","section":"Tags","summary":"","title":"Causality","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/pubtypes/conference/","section":"Pubtypes","summary":"","title":"Conference","type":"pubtypes"},{"content":" Paper Code Abstract # This paper presents CRAFTER, an automated framework for designing and deploying self-adaptive IoT systems using Causal Reinforcement Learning (CRL). As IoT devices increasingly populate pervasive computing spaces, smart environments are enabled with advanced monitoring and interactive services. The dynamic nature of these environments, such as fluctuating workloads and evolving application demands, poses significant challenges in maintaining consistent Quality of Service (QoS) levels of IoT applications. While existing self-adaptation techniques offer adaptive capabilities, they are often designed to deal with specific application domains, hindering the design of self-adaptive solutions that can be re-used across multiple IoT verticals. In addition, there is a lack of automated pipelines that act on identifying key performance drivers to take effective adaptation decisions. CRAFTER addresses these issues by using Causality as a formal framework for performance analysis of IoT systems. CRAFTER generates causal graphs to uncover dependencies among system components and guide adaptation decisions based on cause–effect relationships. Then, adaptation agents can leverage this knowledge to take more effective adaptation decisions in dynamic situations. Our experimental evaluation demonstrates how CRAFTER enables deriving causal graphs spanning diverse IoT use cases. Furthermore, we showcase how CRAFTER improves self-adaptation performance by 25% compared to state-of-the-art Reinforcement Learning-based approaches.\n","date":"April 2026","externalUrl":null,"permalink":"/publications/crafter-seams-2026/","section":"Publications","summary":"CRAFTER relies on Causal Reinforcement Learning for designing autonomous IoT systems.","title":"CRAFTER: Causality-based Self-Adaptation for Autonomous IoT Systems","type":"publications"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/distributed-systems/","section":"Tags","summary":"","title":"Distributed Systems","type":"tags"},{"content":" Professional Experience # Postdoc Researcher Nov. 2025–Present Orange Innovation Châtillon, France Developing adaptive data fusion techniques combining LLMs and ML models to maximize perception precision across heterogeneous robotic swarm sensors. Designing LLM-assisted agentic optimization workflows for 6G industrial IoT communication, translating perception outputs into real-time network decisions via tool-integrated AI agents. Leading research for the CANCUN ANR project, enabling efficient and sustainable IIoT communication. Postdoc Researcher Jan. 2025–Oct. 2025 Télécom SudParis Évry, France Technical lead and main contributor for the PANDORA EU project for developing scalable, trustworthy, and autonomous AIoT operation. Built a causal-GAN synthetic data generation pipeline producing trustworthy, realistic IoT datasets to improve ML model robustness in data-scarce industrial environments. Developed an autonomous MLOps framework for adaptive ML model distribution and lifecycle management across heterogeneous Edge-Cloud infrastructures. Mentored PhD and MSc researchers in autonomous system design and experimental methodology. PhD Researcher Nov. 2021–Dec. 2024 Télécom SudParis Évry, France Developed CRAFTER: a Python/PyTorch causal reinforcement learning system for self-adaptive IoT, reducing response latency by 25% - deployed across a 5-node containerized testbed simulating 100 IoT devices. Developed PlanEMQX: an AI planning-enabled MQTT broker for adaptive, priority-aware data flow management, cutting latency by 20% for time-critical flows - Distinguished Artifact Award, IEEE ICSA 2024. Designed hybrid AI architecture integrating causal inference, RL, and symbolic planning into a unified self-adaptation middleware - published across 3 A-ranked CORE venues. Mentored MSc researchers in autonomous systems and experimental methodology. R\u0026amp;D Intern Apr. 2021–Oct. 2021 Télécom SudParis Évry, France Developed a Java-based simulation tool for performance evaluation of IoT data exchange. Optimized IoT data flow performance using AI planning methodologies. Telecom Engineering Intern May. 2019–Jul. 2019 Dar Al-Handasah Beirut, Lebanon Conducted network analysis and optimization for telecom infrastructures. Education # PhD in Computer Science 2024 Institut Polytechnique de Paris (IP Paris), France Thesis: Enabling autonomous IoT systems : A middleware-based hybrid AI approach to self-adaptation Master\u0026#39;s in Computer Science 2021 Lebanese University, Lebanon Thesis: Designing an Edge-based Data Exchange Infrastructure for Smart Buildings Bachelor of Engineering - Computer Engineering 2020 American University of Beirut, Lebanon Final Year Project: Instructor\u0026rsquo;s Problem Set Recycling and Evolving ","date":"April 2026","externalUrl":null,"permalink":"/","section":"Houssam Hajj Hassan","summary":"Professional Experience # Postdoc Researcher Nov. 2025–Present Orange Innovation Châtillon, France Developing adaptive data fusion techniques combining LLMs and ML models to maximize perception precision across heterogeneous robotic swarm sensors. Designing LLM-assisted agentic optimization workflows for 6G industrial IoT communication, translating perception outputs into real-time network decisions via tool-integrated AI agents. Leading research for the CANCUN ANR project, enabling efficient and sustainable IIoT communication. Postdoc Researcher Jan. 2025–Oct. 2025 Télécom SudParis Évry, France Technical lead and main contributor for the PANDORA EU project for developing scalable, trustworthy, and autonomous AIoT operation. Built a causal-GAN synthetic data generation pipeline producing trustworthy, realistic IoT datasets to improve ML model robustness in data-scarce industrial environments. Developed an autonomous MLOps framework for adaptive ML model distribution and lifecycle management across heterogeneous Edge-Cloud infrastructures. Mentored PhD and MSc researchers in autonomous system design and experimental methodology. PhD Researcher Nov. 2021–Dec. 2024 Télécom SudParis Évry, France Developed CRAFTER: a Python/PyTorch causal reinforcement learning system for self-adaptive IoT, reducing response latency by 25% - deployed across a 5-node containerized testbed simulating 100 IoT devices. Developed PlanEMQX: an AI planning-enabled MQTT broker for adaptive, priority-aware data flow management, cutting latency by 20% for time-critical flows - Distinguished Artifact Award, IEEE ICSA 2024. Designed hybrid AI architecture integrating causal inference, RL, and symbolic planning into a unified self-adaptation middleware - published across 3 A-ranked CORE venues. Mentored MSc researchers in autonomous systems and experimental methodology. R\u0026D Intern Apr. 2021–Oct. 2021 Télécom SudParis Évry, France Developed a Java-based simulation tool for performance evaluation of IoT data exchange. Optimized IoT data flow performance using AI planning methodologies. Telecom Engineering Intern May. 2019–Jul. 2019 Dar Al-Handasah Beirut, Lebanon Conducted network analysis and optimization for telecom infrastructures. Education # PhD in Computer Science 2024 Institut Polytechnique de Paris (IP Paris), France Thesis: Enabling autonomous IoT systems : A middleware-based hybrid AI approach to self-adaptation Master's in Computer Science 2021 Lebanese University, Lebanon Thesis: Designing an Edge-based Data Exchange Infrastructure for Smart Buildings Bachelor of Engineering - Computer Engineering 2020 American University of Beirut, Lebanon Final Year Project: Instructor’s Problem Set Recycling and Evolving ","title":"Houssam Hajj Hassan","type":"page"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/iot/","section":"Tags","summary":"","title":"IoT","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/llm/","section":"Tags","summary":"","title":"LLM","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/prototypes/","section":"Prototypes","summary":"","title":"Prototypes","type":"prototypes"},{"content":" Type All Conference Journal Workshop Artifact Demo Thesis ","date":"April 2026","externalUrl":null,"permalink":"/publications/","section":"Publications","summary":" Type All Conference Journal Workshop Artifact Demo Thesis ","title":"Publications","type":"publications"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/pubtypes/","section":"Pubtypes","summary":"","title":"Pubtypes","type":"pubtypes"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/rag/","section":"Tags","summary":"","title":"RAG","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/reinforcement-learning/","section":"Tags","summary":"","title":"Reinforcement Learning","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/robotics/","section":"Tags","summary":"","title":"Robotics","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/self-adaptation/","section":"Tags","summary":"","title":"Self-Adaptation","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/","section":"Tags","summary":"","title":"Tags","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/pubtypes/artifact/","section":"Pubtypes","summary":"","title":"Artifact","type":"pubtypes"},{"content":" Code Abstract # This artifact paper presents a guide for PSMark, a distributed benchmarking framework to evaluate Publish/Subscribe (pub/sub) systems against real-world representative IoT workloads. PSMark addresses limitations in existing pub/sub benchmarks by supporting: (i) heterogeneous device behaviors (e.g, varying payload sizes, publication rates, and connection stability); (ii) distributed multi-node deployments; and (iii) cross-protocol evaluation across MQTT and DDS.\n","date":"April 2026","externalUrl":null,"permalink":"/publications/psmark-artifact-percom-2026/","section":"Publications","summary":"PSMark is a distributed, multi-protocol benchmark for evaluating topic-filtered pub/sub systems under workloads representative of real-world IoT environments.","title":"Artifact: PSMark: A Distributed IoT Benchmark for Publish/Subscribe Under Domain-Based Workloads","type":"publications"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/benchmark/","section":"Tags","summary":"","title":"Benchmark","type":"tags"},{"content":" Demo Paper Code CRAFTER relies on Causal Reinforcement Learning (CRL) for enabling self-adaptive IoT systems. This is achieved by (i) using causal discovery to generate causal graphs showing cause-effect relationships in IoT systems, and (ii) using Reinforcement Learning for taking adaptation actions at runtime. CRAFTER is evaluted on a metaverse scenario involving the Louvre museum.\nProperties of smart spaces, devices, virtual sensors, and applications and their QoS requirements in JSON format. CRAFTER is built using Java to emulate IoT sensors (112 in total, based on real sensor workloads), processing nodes, and IoT applications. These components are packaged as Docker containers. The networking infrastructure is configured using Mininet and consists of 11 nodes:\n2 Edge nodes for data processing 1 node hosting the data exchange system 1 node for hosting Metaverse applications 7 nodes for deploying IoT sensors. Communication among the different components is achieved via MQTT and using the EMQX message broker. CRAFTER\u0026rsquo;s RL agent is designed and trained using the Stable-baselines3 library and the Gymnasium environment. The causal-learn library is used to extract causakl graphs from IoT datasets.\nCRAFTER\u0026rsquo;s high-level architecture is shown below:\n","date":"April 2026","externalUrl":null,"permalink":"/prototypes/crafter/","section":"Prototypes","summary":"CRAFTER uses Causal Reinforcement Learning for autonomous IoT systems, reducing latency by 25%.","title":"CRAFTER: Causal Reinforcement Learning for Self-adaptive IoT","type":"prototypes"},{"content":" Paper Code Abstract # The Publish/Subscribe (pub/sub) paradigm is widely used in the Internet of Things (IoT). Standalone sensors, wearables, and other devices act as producers that publish messages to consumers such as edge servers or even other IoT devices. Selecting and configuring a pub/sub protocol for an IoT system requires considering network requirements, device reliability, and required Quality-of-Service guarantees. Pub/sub benchmarking suites can help compare expected behavior of various protocols, implementations, and network configurations. However, current pub/sub benchmarks focus primarily on stress testing systems assuming mostly static configurations of homogeneous publishers which are not representative of real-world IoT deployments. To address this, we present PSMark, a distributed, multi-protocol benchmark for evaluating topic-filtered pub/sub systems under workloads representative of real-world IoT environments. PSMark supports (i) workloads representative of heterogeneous IoT device deployments including variations in device communication parameters, (ii) evaluation of distributed IoT deployments with multiple data aggregation servers, (iii) cross-protocol measurements across MQTT and DDS, with extensibility to additional protocols, and (iv) a modular design for adding additional metrics and interfaces. We further construct twelve IoT-focused workloads derived from seven real-world datasets in the domains of manufacturing, healthcare, smart homes, and smart cities. Finally, we benchmark five popular MQTT brokers and one DDS implementation using PSMark and analyze their performance across multiple testbeds and Quality-of-Service settings.\n","date":"April 2026","externalUrl":null,"permalink":"/publications/psmark-percom-2026/","section":"Publications","summary":"PSMark is a distributed, multi-protocol benchmark for evaluating topic-filtered pub/sub systems under workloads representative of real-world IoT environments.","title":"PSMark: A Distributed IoT Benchmark for Publish/Subscribe Under Domain-Based Workloads","type":"publications"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/publish/subscribe/","section":"Tags","summary":"","title":"Publish/Subscribe","type":"tags"},{"content":"","date":"April 2026","externalUrl":null,"permalink":"/tags/qos/","section":"Tags","summary":"","title":"QoS","type":"tags"},{"content":" Paper Abstract # Smart spaces that have deployed machine learning (ML) models for intelligent decision typically evolve over time. Model selection is a key step in deploying (ML) models, particularly in dynamic environments where data distribution shifts can impact model accuracy over time. Existing approaches often rely on evaluating many candidate models against predefined metrics, which is computationally expensive and neither suitable for real-time applications nor dynamic environments. This paper presents an adaptive model selection technique that combines meta-modeling with drift detection to improve efficiency and robustness. Meta-models are used to evaluate the suitability of candidate models under different constraints, such as predictive accuracy and computational cost, without requiring full evaluation on the target data. This reduces the overhead of model selection while preserving deployment quality. In addition, a drift detection mechanism monitors changes in the data distribution and updates the selection strategy accordingly. The proposed approach contributes to automated machine learning (AutoML) by enabling adaptive, efficient, and reliable model selection in real-time environments.\n","date":"January 2026","externalUrl":null,"permalink":"/publications/modelselection-comsnets-2026/","section":"Publications","summary":"This paper presents a comprehensive framework for network-aware path planning that incorporates wireless network quality metrics as path constraints","title":"Adaptive Model Selection using Meta Models and Drift Adaptation","type":"publications"},{"content":"","date":"January 2026","externalUrl":null,"permalink":"/tags/drift-adaptation/","section":"Tags","summary":"","title":"Drift Adaptation","type":"tags"},{"content":"","date":"January 2026","externalUrl":null,"permalink":"/tags/machine-learning/","section":"Tags","summary":"","title":"Machine Learning","type":"tags"},{"content":"","date":"January 2026","externalUrl":null,"permalink":"/tags/model-selection/","section":"Tags","summary":"","title":"Model Selection","type":"tags"},{"content":"","date":"January 2026","externalUrl":null,"permalink":"/pubtypes/workshop/","section":"Pubtypes","summary":"","title":"Workshop","type":"pubtypes"},{"content":" Paper Abstract # Autonomous Mobile Robots (AMRs) in industrial environments require reliable wireless connectivity for coordination, control, and safety operations. Traditional path planning algorithms focus solely on geometric constraints, often leading robots through areas with poor network coverage that can compromise mission-critical operations. This paper presents a comprehensive framework for network-aware path planning that incorporates wireless network quality metrics as path constraints. We validate our Sionna-based ray-tracing simulations against real-world measurements from the Hernangómez et al. iV2I+ dataset, achieving strong correlation between simulated and real-world measurements (R² = 0.87 for SNR, 0.82 for throughput). Using this validated simulation framework, we implement three novel path planning algorithms: A* with network constraints, conditional variational autoencoder (CVAE)-based neural path planning, and Graph neural network-based multi-path (GraphMP) planning. Our evaluation demonstrates trade-offs between network quality requirements and path efficiency, with CVAE achieving 95.2% constraint satisfaction and GraphMP showing 23% shorter planning times. We provide practical guidelines for selecting appropriate algorithms and network quality thresholds based on application requirements, enabling more reliable AMR operations in industrial settings.\n","date":"December 2025","externalUrl":null,"permalink":"/publications/amr-ants-2025/","section":"Publications","summary":"This paper presents a comprehensive framework for network-aware path planning that incorporates wireless network quality metrics as path constraints","title":"Network-Aware Path Planning for Autonomous Mobile Robots in Industrial Environments","type":"publications"},{"content":"","date":"December 2025","externalUrl":null,"permalink":"/tags/path-planning/","section":"Tags","summary":"","title":"Path Planning","type":"tags"},{"content":"","date":"April 2025","externalUrl":null,"permalink":"/tags/ai-planning/","section":"Tags","summary":"","title":"AI Planning","type":"tags"},{"content":"","date":"April 2025","externalUrl":null,"permalink":"/tags/security/","section":"Tags","summary":"","title":"Security","type":"tags"},{"content":"","date":"April 2025","externalUrl":null,"permalink":"/tags/self-protection/","section":"Tags","summary":"","title":"Self-Protection","type":"tags"},{"content":" Paper Code Abstract # Today\u0026rsquo;s smart spaces deploy various IoT devices to offer services for occupants. Such devices are exposed to security risks that may pose serious threats to network services and users\u0026rsquo; privacy. To avoid the disruption of normal operations, self-protecting solutions have been developed to allow IoT networks to autonomously respond to cyber threats in real-time. However, existing self-protecting systems focus solely on architectural adaptations to respond to cyber threats, overlooking the mitigation actions described in cybersecurity standards \u0026ndash;which represent the correct cybersecurity posture\u0026ndash; as well as the impact of the adaptation strategies on the Quality-of-Service (QoS) performance. To overcome these existing limitations, this paper presents SPARQ, a novel framework for designing self-protecting IoT systems that considers both the security exposure to cyber attacks and the QoS performance. We leverage Attack Graph as a threat model for analyzing the cyber exposure of the system and Queuing Network Models to analyze QoS in IoT systems. Based on the analysis outcomes, SPARQ provides mitigation plans to reduce the cyber risk while also minimizing the impact on QoS. We evaluate the proposed approach on two use cases from real-world scenarios including a critical infrastructure and a smart home. The experimental evaluation shows that SPARQ is capable of reducing the cyber risk significantly while also improving the QoS performance by 35% compared to existing approaches.\n","date":"April 2025","externalUrl":null,"permalink":"/publications/sparq-seams-2025/","section":"Publications","summary":"SPARQ is a framework for designing self-protecting IoT systems that considers both the security exposure to cyber attacks and the QoS performance.","title":"SPARQ: A QoS-Aware Framework for Mitigating Cyber Risk in Self-Protecting IoT Systems","type":"publications"},{"content":" Paper Code SPARQ is a novel framework for designing self-protecting IoT systems that considers both the security exposure to cyber attacks and the QoS performance. We leverage Attack Graph as a threat model for analyzing the cyber exposure of the system and Queuing Network Models to analyze QoS in IoT systems. Based on the analysis outcomes, SPARQ provides mitigation plans to reduce the cyber risk while also minimizing the impact on QoS.\nSPARQ relies on Pythonn to generate attack graphs, on the Java Modelling Tools simulator to evaluate IoT QoS performance, and on PDDL for generating mitigation plans.\n","date":"April 2025","externalUrl":null,"permalink":"/prototypes/sparq/","section":"Prototypes","summary":"SPARQ is a novel framework for designing self-protecting IoT systems that considers both the security exposure to cyber attacks and the QoS performance.","title":"SPARQ: QoS-aware Self-protection for IoT","type":"prototypes"},{"content":"","date":"February 2025","externalUrl":null,"permalink":"/tags/dataset-generation/","section":"Tags","summary":"","title":"Dataset Generation","type":"tags"},{"content":" Paper Abstract # The strategic placement of nodes in Wireless IoT Networks (WIoTs) is crucial for ensuring optimal coverage, connectivity, and energy efficiency. Traditionally, node placement has relied on heuristic and manual methods, which often result in inefficiencies and suboptimal network performance. In this paper, we focus on optimizing the coverage performance of WIoTs, which play a pivotal role in environmental monitoring and event detection. In particular, we first develop a tool that allows IoT designers to simulate and generate datasets for multiple sensor deployment options. Then, we empirically generate a dataset that can contribute to the growing field of optimized sensor placement strategies by bridging algorithmic simulations with predictive modeling. Finally, we use the generated dataset to train a decision tree model for sensor node placement predictions. The prototype implementation of our tool and the generated datasets are publicly available for exploitation from the research community.\n","date":"February 2025","externalUrl":null,"permalink":"/publications/empirical-menacomm-2025/","section":"Publications","summary":"This paper focuses on optimizing the coverage performance of WIoT networks, which play a pivotal role in environmental monitoring and event detection.","title":"Empirical Dataset Generation for AI-optimized IoT Infrastructure Placement","type":"publications"},{"content":"","date":"February 2025","externalUrl":null,"permalink":"/tags/sensor-placement/","section":"Tags","summary":"","title":"Sensor Placement","type":"tags"},{"content":"","date":"December 2024","externalUrl":null,"permalink":"/tags/causal-discovery/","section":"Tags","summary":"","title":"Causal Discovery","type":"tags"},{"content":" Thesis Abstract # The proliferation of Internet of Things (IoT) devices has transformed traditional spaces into smarter, more interconnected environments. These advanced IoT systems consist of devices that sense physical phenomena and generate data, which are then processed by computing nodes before being used by applications. Such applications typically define specific Quality-of-Service (QoS) requirements, such as availability, accuracy, and latency, that must be met. To achieve this, IoT systems are normally configured to ensure that the QoS requirements of the deployed applications are met. This involves adjusting multiple parameters such as network settings, processing resources, and tuning data exchange systems. However, modern smart spaces are inherently dynamic and unpredictable. Changes in the number of IoT devices, network conditions, and available computational resources create a continuously evolving environment. Thus, to ensure that IoT systems operate autonomously, it is essential to design advanced self-adaptive mechanisms for maintaining QoS requirements of applications across dynamic smart spaces.This thesis proposes a middleware-based, hybrid Artificial Intelligence (AI) self-adaptation approach for enabling autonomous IoT operations across dynamic smart spaces. By combining logic-based approaches with data-driven AI techniques, we design effective and explainable self-adaptation solutions for IoT systems operating over dynamic environments. This is achieved through three main research contributions. First, queueing network modeling techniques are leveraged to compose QoS models that represent IoT systems under different situations and/or configurations of smart spaces. By simulating QoS models, we generate performance metrics datasets that can be used as input in self-adaptive approaches. These approaches dynamically adjust to changing conditions by selecting the configuration that best meets the QoS requirements specified by the applications. The second contribution enables AI-driven adaptation of IoT systems in smart spaces. By combining AI techniques such as Automated Planning and Reinforcement Learning, we design a framework involving intelligent agents capable of taking adaptation decisions at runtime. Possible adaptation actions include data flow configurations (e.g., priorities, drop rates) and resource control (e.g., network resources, computing resources). Finally, the third contribution enables proactive and explainable autonomous systems by relying on Causal Reinforcement Learning. To achieve this, Causality methodologies are employed to provide a formal analysis of the performance of IoT systems. Subsequently, causal models enable agents to take efficient adaptation actions in dynamic environments. We validate our proposed approach by developing a prototype implementation of an IoT system and experimenting with case studies considering different types of IoT environments. Our QoS models are evaluated and compared with a prototype implementation for validating the accuracy of the generated performance metrics datasets. We then evaluate the effectiveness of our solution by leveraging data from real deployments to ensure that our approach is valid in real-life settings. The self-adaptation approach presented in this thesis can be exploited to design resilient and modern IoT systems where interactive and real-time services are required, enabling smart spaces to autonomously adapt to changes in their environment, and ensuring optimal performance even under dynamic conditions.\n","date":"December 2024","externalUrl":null,"permalink":"/publications/thesis-ipparis-2024/","section":"Publications","summary":"This PhD thesis proposes a middleware-based, hybrid Artificial Intelligence (AI) self-adaptation approach for enabling autonomous IoT operations across dynamic smart spaces.","title":"Enabling Autonomous IoT Systems: A Middleware-based Hybrid AI Approach to Self-adaptation","type":"publications"},{"content":"","date":"December 2024","externalUrl":null,"permalink":"/pubtypes/thesis/","section":"Pubtypes","summary":"","title":"Thesis","type":"pubtypes"},{"content":" Paper Abstract # Buildings account for a significant share of global energy consumption, with Heating, Ventilation, and Air Conditioning (HVAC) systems being responsible for up to 60% of a building\u0026rsquo;s energy usage. For this purpose, existing scheduling and control solutions can be used to design more sustainable energy systems and limit their environmental impact. However, these approaches mainly consider HVAC, ignoring other energy systems in buildings such as lighting control and plug loads. In addition, these solutions have to be customized for a specific building instance, hindering portability across different application domains. This paper presents a holistic approach for efficiently scheduling smart building energy systems through AI planning methodologies. AI planning enables decoupling domain knowledge from problem representations, enhancing portability and allowing for straightforward runtime adaptation when needed. We evaluate our approach in a smart office setting and show how AI planning enables reducing energy consumption by up to 30%.\n","date":"June 2024","externalUrl":null,"permalink":"/publications/scheduling-ict4s-2024/","section":"Publications","summary":"This paper presents a holistic approach for efficiently scheduling smart building energy systems through AI planning methodologies.","title":"Efficient Scheduling of Smart Building Energy Systems using AI Planning","type":"publications"},{"content":"","date":"June 2024","externalUrl":null,"permalink":"/tags/energy-efficiency/","section":"Tags","summary":"","title":"Energy Efficiency","type":"tags"},{"content":" Paper Code Abstract # In today’s IoT environments, message brokers play a pivotal role in facilitating data exchange between IoT devices and applications. Existing message broker implementations offer different configuration options for IoT systems designers for performance tuning. However, designers still have to manually configure the message broker to find the best parameters combination that satisfies the requirements of the deployed applications. In addition, runtime changes might lead to performance degradation and require reconfiguration. This paper presents a publish/subscribe message broker architecture for enabling adaptive data exchange in IoT environments. Core software components are proposed for (i) refining per-subscription data flows based on the applications deployed in the environment, (ii) dynamically assigning drop rates or priorities to data flows according to the requirements of these applications, and (iii) enabling the adaptation of data flows based on dynamic changes in the environment or evolving applications’ requirements. The proposed architecture is enriched with automated planning capabilities for providing such adaptation of data flows. To demonstrate the applicability of our architecture, we implement the PlanEMQX prototype. Our experimental evaluation shows improvements of 20% of response time for time-sensitive data flows.\n","date":"June 2024","externalUrl":null,"permalink":"/publications/planemqx-icsa-2024/","section":"Publications","summary":"This paper introduces an automated synthesis of QoS-aware mediating artifacts.","title":"A Message Broker Architecture for Adaptive Data Exchange in the IoT","type":"publications"},{"content":" Paper Code Abstract # Internet of Things (IoT) applications consist of diverse resource-constrained/rich devices with a considerable portion being mobile. Such devices demand lightweight, loosely coupled interactions in terms of time, space, and synchronization. IoT protocols at the middleware layer support several interaction types (e.g., asynchronous messaging, streaming, etc.) ensuring successful interactions between devices that use the same protocol. Additionally, they introduce different Quality of Service (QoS) delivery modes for data exchange with respect to available device and network resources. On the other hand, interconnecting heterogeneous IoT devices requires mapping both their functional and QoS properties. This calls for advanced interoperability solutions integrated with QoS modeling and analysis techniques. This paper introduces an automated synthesis of QoS-aware mediating artifacts. Such mediators enable the interconnection between IoT devices employing heterogeneous middleware protocols. Additionally, representative QoS models are synthesized. Leveraging these models, system designers can evaluate the effectiveness of the interconnection in terms of end-to-end QoS. We evaluate the usefulness of our approach through experimentation with a case study employing heterogeneous middleware protocols. In particular, we statistically analyze through simulations the effect of varying system parameters on the end-to-end QoS.\n","date":"June 2024","externalUrl":null,"permalink":"/publications/interop-icsa-2024/","section":"Publications","summary":"This paper introduces an automated synthesis of QoS-aware mediating artifacts.","title":"Automating the Evaluation of Interoperability Effectiveness in Heterogeneous IoT Systems","type":"publications"},{"content":"","date":"June 2024","externalUrl":null,"permalink":"/tags/interoperability/","section":"Tags","summary":"","title":"Interoperability","type":"tags"},{"content":"","date":"June 2024","externalUrl":null,"permalink":"/tags/software-architecture/","section":"Tags","summary":"","title":"Software Architecture","type":"tags"},{"content":" Paper Code PlanEMQX is a prototype built on top of the EMQX message broker, enabling adaptive data exchange in IoT environments. This is achieved by (i) refining per-subscription data flows based on the applications deployed in the environment, (ii) dynamically assigning drop rates or priorities to data flows according to the requirements of the subscribing applications, and (iii) enabling the adaptation of data flows based on dynamic changes in the environment or evolving applications\u0026rsquo; requirements.\nPlanEMQX is deployed using the Mininet network emulator. The PDDL modelling language is then used to define domain and problem files used by AI Planning. PlanEMQX relies on the Metric-FF planner to generate configuration plans for IoT infrastructures.\nTo create data flows per-subscription, PlanEMQX leverages EMQX\u0026rsquo;s Authentication API and applies the following process:\nTo define priorites per data flow, PlanEMQX follows the process below:\n","date":"June 2024","externalUrl":null,"permalink":"/prototypes/planemqx/","section":"Prototypes","summary":"PlanEMQX is an AI Planning-enabled message broker that can reduce latency by \u003e30%.","title":"PlanEMQX: AI-enabled Message Broker","type":"prototypes"},{"content":"","date":"June 2023","externalUrl":null,"permalink":"/tags/data-generation/","section":"Tags","summary":"","title":"Data Generation","type":"tags"},{"content":" Paper Code EDICT is a simulation tool for evaluating the performance of Edge interactions in IoT-enhanced environments. The standard NGSI-LD (Next Generation Service Interfaces-Linked Data) protocol specification is used to represent systems deployed in IoT-enhanced environments. EDICT then generates a performance metrics dataset as a CSV file by relying on queueing network models.\nEDICT relies on Pon the Java Modelling Tools simulator to evaluate IoT QoS performance.\n","date":"June 2023","externalUrl":null,"permalink":"/prototypes/edict/","section":"Prototypes","summary":"EDICT uses queueing networks for generation IoT performance datasets.","title":"EDICT: A Simulation Tool for Generating IoT Performance Datasets","type":"prototypes"},{"content":"","date":"June 2023","externalUrl":null,"permalink":"/tags/queueing-networks/","section":"Tags","summary":"","title":"Queueing Networks","type":"tags"},{"content":"","date":"June 2023","externalUrl":null,"permalink":"/tags/simulation/","section":"Tags","summary":"","title":"Simulation","type":"tags"},{"content":" Paper Code Abstract # This paper demonstrates EDICT, a simulation tool for performance metrics datasets in IoT environments. Such environments are represented using the NGSI-LD data model. EDICT uses NGSI-LD instances and Open Queueing Networks for the composition of a concrete QoS model that is then simulated to generate a performance metrics dataset. This dataset captures performance metrics of Edge interactions such as response time and throughput. We demonstrate the utility of EDICT by showing how a perfomance metrics dataset can be generated for a specific IoT environment.\n","date":"June 2023","externalUrl":null,"permalink":"/publications/edict-demo-dcoss-2023/","section":"Publications","summary":"EDICT is a tool for simulating Edge interactions in IoT-enhanced environments.","title":"(DEMO) EDICT: Simulation of Edge Interactions across IoT-enhanced Environments","type":"publications"},{"content":"","date":"June 2023","externalUrl":null,"permalink":"/pubtypes/demo/","section":"Pubtypes","summary":"","title":"Demo","type":"pubtypes"},{"content":" Paper Code Abstract # This paper presents EDICT, a tool for simulating Edge interactions in IoT-enhanced environments. Recently, ML and AI-based techniques have gained prominence to solve IoT related challenges. However, such models require large and diverse datasets to perform well. Finding real-world datasets that capture the performance of IoT systems is a challenging task due to the cost of deploying devices and instrumenting environments, as well as privacy/security concerns. This task becomes more challenging when datasets for specific situations (e.g., overloaded system, emergency scenarios) are needed. EDICT enables IoT systems designers to evaluate the performance of their IoT systems at design time. EDICT is capable of generating performance metrics datasets for specific instances of IoT-enhanced environments under different configuration parameters. To support runtime adaptation of smart environments, EDICT enables rapid performance prediction using ML techniques.\n","date":"June 2023","externalUrl":null,"permalink":"/publications/edict-dcoss-2023/","section":"Publications","summary":"EDICT is a tool for simulating Edge interactions in IoT-enhanced environments.","title":"EDICT: Simulation of Edge Interactions across IoT-enhanced Environments","type":"publications"},{"content":" Paper Code Abstract # This paper presents the implementation and guideline of PlanIoT, an adaptive flow management framework for IoT-enhanced spaces. Such spaces are composed of applications deployed at the Edge with varying QoS requirements in terms of response time, timely delivery, throughput, etc. Configuring the Edge infrastructure requires tuning multiple parameters for optimal QoS satisfaction of applications. This is a complex task especially when the system has to be re-adapted (e.g., emergency situations). The PlanIoT framework manages application data flows in an adaptive manner. This is achieved via the following core software components: (i) a queueing network composer; (ii) an automated planning modeler; and (iii) an AI planner. This artifact presents implementation details of these components as well as guidelines for using the PlanIoT framework.\n","date":"May 2023","externalUrl":null,"permalink":"/publications/planiot-artifact-seams-2023/","section":"Publications","summary":"PlanIoT is a middleware approach for enabling adaptive data flow management in IoT-enhanced spaces using AI Planning.","title":"Artifact: Implementation of an Adaptive Flow Management Framework for IoT Spaces","type":"publications"},{"content":" Paper Code Abstract # This paper presents PlanIoT, a middleware approach for enabling adaptive data flow management in IoT-enhanced spaces (e.g., buildings) using automated planning methodologies. Today\u0026rsquo;s sensorized spaces deploy applications falling to diverse categories such as analytics, real-time, transactional, video streaming and emergency response. Depending on the category, applications have different QoS requirements related to timely delivery, networking resources, accuracy, etc. Typically, state-of-the-art data exchange systems introduce policies for bandwidth allocation or prioritization for specific data types and applications (e.g., camera data). PlanIoT introduces a generic QoS model to evaluate the performance of data flowing in Edge infrastructures and generates their performance metrics dataset. Such a dataset is used as input to automated planning representations to intelligently satisfy QoS requirements of deployed applications. The experimental results show that PlanIoT improves the end-to-end response time of time-sensitive flows by more than 50%, especially with an overloaded Edge infrastructure. We also show the adaptivity of our approach by considering emergency cases that require Edge infrastructure reconfiguration.\n","date":"May 2023","externalUrl":null,"permalink":"/publications/planiot-seams-2023/","section":"Publications","summary":"PlanIoT is a middleware approach for enabling adaptive data flow management in IoT-enhanced spaces using AI Planning.","title":"PlanIoT: A Framework for Adaptive Data Flow Management in IoT-enhanced Spaces","type":"publications"},{"content":" Paper Code PlanIoT is a framework-based solution that enables adaptive data flow management at the middleware-layer using AI Planning. This is achieved via the following core software components: (i) a queueing network composer; (ii) automated planning modeler; and (iii) an AI planner.\nPlanIoT uses the Java Modelling Tools simulator to compose and simulate queueing networks that evaluate MQTT-based publish/subscribe interactions in IoT environments. The PDDL modelling language is then used to define domain and problem files used by AI Planning. PlanIoT relies on the Metric-FF planner to generate configuration plans for IoT infrastructures.\nPlanIoT\u0026rsquo;s process and workflow for generating configuration and adaptation plans for IoT infrastructures are shown below:\n","date":"April 2023","externalUrl":null,"permalink":"/prototypes/planiot/","section":"Prototypes","summary":"PlanIoT uses AI Planning for adaptive IoT data flow management, reducing latency by \u003e50%.","title":"PlanIoT: Adaptive IoT Data Flow Management","type":"prototypes"},{"content":"","date":"September 2021","externalUrl":null,"permalink":"/tags/covid-19/","section":"Tags","summary":"","title":"COVID-19","type":"tags"},{"content":"","date":"September 2021","externalUrl":null,"permalink":"/tags/d2d/","section":"Tags","summary":"","title":"D2D","type":"tags"},{"content":" Paper Abstract # In this paper we present an application for COVID-19 contact tracing based on device-to-device technology. This paradigm allows communication between two devices without the need for network infrastructure. Recently, various technologies such as Bluetooth and Wi-Fi Direct have been used in this field. Our approach is based on Wi-Fi Direct. We first explore similar approaches that have been developed. We then present the architecture of the proposed application and show a prototype implementation.\n","date":"September 2021","externalUrl":null,"permalink":"/publications/d2d-smartnets-2021/","section":"Publications","summary":"This paper presents an application for COVID-19 contact tracing based on device-to-device technology.","title":"D2D Communication: COVID-19 Contact Tracing Application Using Wi-Fi Direct","type":"publications"},{"content":"I\u0026rsquo;m a postdoctoral researcher at Orange Innovation, where I design and deplopy robotic swarm architectures combining multimodal perception, LLM-based reasoning, and AI planning to deliver robust autonomous operation under real-world constraints. Previously, at Télécom SudParis, I led core technical contributions for the PANDORA EU project, where I developed scalable AIoT pipelines including ML-based synthetic data generation, adaptive model selection, and network-aware planning for robotic systems. I\u0026rsquo;ve obtained my PhD from Institut Polytechnique de Paris (IP Paris) in 2024 with honors.\nI build scalable AI systems that enable autonomous decision-making in industrial and distributed environments. My expertise spans self-adaptive systems, AIoT, edge–cloud deployments, and distributed systems. I specialize in turning advanced AI research into deployable, resilient systems that operate at scale.\nProfessional Experience # Postdoc Researcher Nov. 2025–Present Orange Innovation Châtillon, France Developing adaptive data fusion techniques combining LLMs and ML models to maximize perception precision across heterogeneous robotic swarm sensors. Designing LLM-assisted agentic optimization workflows for 6G industrial IoT communication, translating perception outputs into real-time network decisions via tool-integrated AI agents. Leading research for the CANCUN ANR project, enabling efficient and sustainable IIoT communication. Postdoc Researcher Jan. 2025–Oct. 2025 Télécom SudParis Évry, France Technical lead and main contributor for the PANDORA EU project for developing scalable, trustworthy, and autonomous AIoT operation. Built a causal-GAN synthetic data generation pipeline producing trustworthy, realistic IoT datasets to improve ML model robustness in data-scarce industrial environments. Developed an autonomous MLOps framework for adaptive ML model distribution and lifecycle management across heterogeneous Edge-Cloud infrastructures. Mentored PhD and MSc researchers in autonomous system design and experimental methodology. PhD Researcher Nov. 2021–Dec. 2024 Télécom SudParis Évry, France Developed CRAFTER: a Python/PyTorch causal reinforcement learning system for self-adaptive IoT, reducing response latency by 25% - deployed across a 5-node containerized testbed simulating 100 IoT devices. Developed PlanEMQX: an AI planning-enabled MQTT broker for adaptive, priority-aware data flow management, cutting latency by 20% for time-critical flows - Distinguished Artifact Award, IEEE ICSA 2024. Designed hybrid AI architecture integrating causal inference, RL, and symbolic planning into a unified self-adaptation middleware - published across 3 A-ranked CORE venues. Mentored MSc researchers in autonomous systems and experimental methodology. R\u0026amp;D Intern Apr. 2021–Oct. 2021 Télécom SudParis Évry, France Developed a Java-based simulation tool for performance evaluation of IoT data exchange. Optimized IoT data flow performance using AI planning methodologies. Telecom Engineering Intern May. 2019–Jul. 2019 Dar Al-Handasah Beirut, Lebanon Conducted network analysis and optimization for telecom infrastructures. Education # PhD in Computer Science 2024 Institut Polytechnique de Paris (IP Paris), France Thesis: Enabling autonomous IoT systems : A middleware-based hybrid AI approach to self-adaptation Master\u0026#39;s in Computer Science 2021 Lebanese University, Lebanon Thesis: Designing an Edge-based Data Exchange Infrastructure for Smart Buildings Bachelor of Engineering - Computer Engineering 2020 American University of Beirut, Lebanon Final Year Project: Instructor\u0026rsquo;s Problem Set Recycling and Evolving ","externalUrl":null,"permalink":"/about/","section":"Houssam Hajj Hassan","summary":"I’m a postdoctoral researcher at Orange Innovation, where I design and deplopy robotic swarm architectures combining multimodal perception, LLM-based reasoning, and AI planning to deliver robust autonomous operation under real-world constraints. Previously, at Télécom SudParis, I led core technical contributions for the PANDORA EU project, where I developed scalable AIoT pipelines including ML-based synthetic data generation, adaptive model selection, and network-aware planning for robotic systems. I’ve obtained my PhD from Institut Polytechnique de Paris (IP Paris) in 2024 with honors.\n","title":"About","type":"page"},{"content":"","externalUrl":null,"permalink":"/categories/","section":"Categories","summary":"","title":"Categories","type":"categories"},{"content":" houssam.hajjhassan@orange.com Orange Gardens, Châtillon, France ","externalUrl":null,"permalink":"/contact/","section":"Houssam Hajj Hassan","summary":" houssam.hajjhassan@orange.com Orange Gardens, Châtillon, France ","title":"Contact","type":"page"},{"content":"","externalUrl":null,"permalink":"/venues/","section":"Venues","summary":"","title":"Venues","type":"venues"},{"content":"","externalUrl":null,"permalink":"/years/","section":"Years","summary":"","title":"Years","type":"years"}]