Engineering synthetic biology sensors with artificial intelligence: From programmable circuits to next-generation biosensing.

Artificial Intelligence (AI) models are critical for detecting advanced Internet of Things (IoT) botnets. However, these systems are highly vulnerable to Adversarial Machine Learning (AML), where malicious inputs are crafted to cause misclassification (e.g., identifying malicious traffic as benign), posing a systemic threat to IoT security. This systematic literature review (SLR) addresses the persistent "reality gap" between theoretical AML research, often derived from computer vision, and the practical, domain-specific constraints of network security. This paper synthesizes research from 2020–2025, providing comprehensive taxonomies of: (RQ1) targeted AI models, from traditional ML to modern Federated Learning (FL) frameworks; (RQ2) attack methodologies, highlighting the shift from feature-space (e.g., PGD) to realistic problem-space attacks (e.g., binary diversification, XAI-based attacks); (RQ3) proactive (e.g., Adversarial Training) and reactive defense strategies; and (RQ4) evaluation frameworks, critiquing the use of outdated datasets. Finally, (RQ5) we analyze open challenges, focusing on the IoT resource-constraint dilemma—where effective defenses like Adversarial Training are too computationally expensive for edge devices —and performance trade-offs. We conclude by outlining future directions, emphasizing the need for constraint-aware defenses, secure FL, and leveraging Generative AI.

Interactive optimisation (IO) combines the analytical power of optimisation frameworks with human’s contextual expertise. However, prior IO approaches require human users to repeatedly provide the same type of input or directly modify the model to incorporate different information. As a result, IO frameworks elicit a narrow range of human knowledge or require substantial optimisation expertise from users. To address these limitations, an IO framework is proposed that allows human users to respond to multiple types of queries. The framework aims to produce higher-fidelity stochastic multi-objective mixed-integer linear programming models. It employs targeted questions to elicit specific information from users, a Monte Carlo-based framework to transform human responses into input data for a scenario-based optimisation model formulation, and uses a Conditional Value at Risk (CVaR) formulation to balance expected performance with risk tolerances. Computational experiments on a supplier selection problem demonstrate that this framework can narrow the reality gap and converge towards the ground-truth solution. Moreover, it dynamically adapts to user feedback, and when the human expresses insufficient confidence in the solution’s performance, it can recommend solutions with narrower performance confidence intervals.

Digital twins (DTs) enable powerful predictive analytics, but persistent discrepancies between simulations and real systems—known as the reality gap—undermine their reliability. Coined in robotics, the term now applies to DTs, where discrepancies stem from context mismatches, cross domain interactions, and multiscale dynamics. Among these, context mismatch is pressing and underexplored, as DT accuracy depends on capturing operational context, often only partially observable. However, DTs have a key advantage: simulators can systematically vary contextual factors and explore scenarios difficult or impossible to observe empirically, informing inference and model alignment. While sim-to-real transfer like domain adaptation shows promise in robotics, its application to DTs poses two key challenges. First, unlike one-time policy transfers, DTs require continuous calibration across an asset’s lifecycle—demanding structured information flow, timely detection of out-of-sync states, and integration of historical and new data. Second, DTs often perform inverse modeling, inferring latent states or faults from observations that may reflect multiple evolving contexts. These needs strain purely data-driven models and risk violating physical consistency. Though some approaches preserve validity via a reduced-order model, most domain adaptation techniques still lack such constraints. To address this, we propose a reality gap analysis (RGA) module for DTs that continuously integrates new sensor data, detects misalignments, and recalibrates DTs via a query-response framework. Our approach fuses domain-adversarial deep learning with reduced-order simulator guidance to improve context inference and preserve physical consistency. We illustrate the RGA module in a structural health monitoring case study on a steel truss bridge in Pittsburgh, PA, showing faster calibration and better real-world alignment.

During manual operations, the human brain relies on mediated visual stimuli such as color to estimate an object’s weight and adjust muscle force through the central nervous system (CNS). This study examines the neuromuscular “reality gap” induced by the color–weight illusion (CWI) during repetitive lifting tasks in an augmented reality (AR) interface. We analyzed the median frequency (MDF) and Co-Contraction Index (CCI) of the biceps and triceps muscles to quantify physiological strain under varying luminance conditions in both AR and physical environments. The results reveal that AR significantly amplifies the CWI, with black stimuli triggering an aggressive joint-stiffening strategy in the AR group (APG). Compared with the physical reality group, the AR group showed lower overall endurance (91.4 ± 22.8 vs. 100.1 ± 12.5 repetitions) and a stronger physiological response to the black stimulus. In the AR group, the black condition was associated with a terminal CCI of 84.7 ± 25.4% and an MDF decline of approximately 21.7 Hz, whereas the corresponding contrast was attenuated in the physical reality group. These findings demonstrate a critical decoupling between behavioral output and internal physiological strain, indicating that the CNS treats virtual visual cues as high-reliability signals that increase metabolic “bio-cost” despite task completion parity. This research identifies a “masking effect” where behavioral metrics hide severe ergonomic risks, providing novel approaches for managing musculoskeletal health in industrial settings and personalizing coordination training in clinical rehabilitation.

Bridging the sensor reality gap: Adaptive learning from implicit degradation priors for low-light image enhancement

Controlling a flexible wheeled robot for complex tasks such as stair climbing is highly challenging. The nonlinearity inherent in soft materials hinders accurate modeling, creating a trade-off in Reinforcement Learning (RL) between simulation fidelity and learning speed. We propose an RL-friendly, multi-body model that approximates the deformation of the flexible wheel as a Mass-Spring-Damper (MSD) system composed of rigid links and joints. This model enables end-to-end RL within a fast rigid-body simulator, facilitating a blind control policy that relies solely on proprioceptive feedback. To reduce the reality gap and enhance policy robustness, we randomize the main parameters of the MSD system. In real-world experiments, a robot successfully climbed an 18 cm step, corresponding to approximately 51% of the wheel radius—a feat impossible for a rigid-wheeled equivalent. To our knowledge, this is the first successful application of RL-based blind control for stair climbing with a flexible wheeled robot. However, structural limitations in our model and challenges in parameter identification hinder sim-to-real transfer, and improving robustness remains a key issue for future work.

The Newcomer's dilemma: Navigating strategic risks in nuclear fuel cycle development a simulation-based framework for Indonesia

The successful transfer of autonomous driving stacks (ADS) from simulation to the real world faces two main challenges: the Reality Gap (RG)-mismatches between simulated and real behaviors-and the Performance Gap (PG)-differences between expected and achieved performance across domains. We propose a Methodology for Closing Reality and Performance Gaps (MCRPG), a structured and iterative approach that jointly reduces RG and PG through parameter tuning, cross-domain metrics, and staged validation. MCRPG comprises three stages-Digital Twin, Parallel Execution, and Real-World-to progressively align ADS behavior and performance. To ground and validate the method, we present an open-source, cost-effective Development and Validation Platform (DVP) that integrates an ROS-based modular ADS with the CARLA simulator and a custom autonomous electric vehicle. We also introduce a two-level metric suite: (i) Reality Alignment via Maximum Normalized Cross-Correlation (MNCC) over multi-modal signals (e.g., ego kinematics, detections), and (ii) Ego-Vehicle Performance covering safety, comfort, and driving efficiency. Experiments in an urban scenario show convergence between simulated and real behavior and increasingly consistent performance across stages. Overall, MCRPG and DVP provide a replicable framework for robust, scalable, and accessible Sim2Real research in autonomous navigation techniques.

Hand localization in cluttered industrial environments remains challenging due to variations in appearance and the gap between synthetic and real-world data. Domain randomization addresses this “reality gap” by intentionally introducing randomized and unrealistic visual features in simulated scenes, encouraging neural networks to focus on essential domain-invariant cues. In this study, we applied domain randomization to generate a synthetic Red-Green-Blue–Depth (RGB-D) dataset for training multimodal instance segmentation models, with the aim of achieving color-agnostic hand localization in complex industrial settings. We introduce a new synthetic dataset tailored to various hand detection tasks and provide ready-to-use pretrained instance segmentation models. To enhance robustness in unstructured environments, the proposed approach employs multimodal inputs that combine color and depth information. To evaluate the contribution of each modality, we analyzed the individual and combined effects of color and depth on model performance. All evaluated models were trained exclusively on the proposed synthetic dataset. Despite the absence of real-world training data, the results demonstrate that our models outperform corresponding models trained on existing state-of-the-art datasets, achieving higher Average Precision and Probability-Based Detection Quality.

Autonomous driving systems rely on vast and diverse datasets for robust object recognition. However, acquiring real-world data, especially for rare and hazardous scenarios, is prohibitively expensive and risky. While purely synthetic data offers flexibility, it often suffers from a significant reality gap due to discrepancies in visual fidelity and physics. To address these challenges, this paper proposes a novel real–virtual fusion framework for efficiently generating highly realistic augmented image datasets for autonomous driving. Our methodology leverages real-world driving data from South Korea’s K-City, synchronizing it with a digital twin environment in Morai Sim (v24.R2) through a robust look-up table and fine-tuned localization approach. We then seamlessly inject diverse virtual objects (e.g., pedestrians, vehicles, traffic lights) into real image backgrounds. A critical contribution is our focus on inconsistency mitigation, employing advanced techniques such as illumination matching during virtual object injection to minimize visual discrepancies. We evaluate the proposed approach through experiments. Our results show that this real–virtual fusion strategy significantly bridges the reality gap, providing a cost-effective and safe solution for enriching autonomous driving datasets and improving the generalization capabilities of perception models.

To develop and validate an agent-based Large Language Model (LLM) system for extracting structured data from breast cancer synoptic pathology reports and assess the performance gap between synthetic and real-world validation. We developed a modular artificial intelligence (AI) agent-based framework employing sequential specialized LLMs for parsing pathology reports and extracting structured data. We normalized College of American Pathologists (CAP) cancer protocols into 8 sections, 86 subsections, and 229 discrete fields. Seven leading LLMs (gemini-2.5-pro, llama3.3-70b, phi4-14b, deepseek-r1 14B/70B, gemma3-27b, gemini-2.0-flash-lite) were validated using dual evaluation: synthetic validation (864 controlled test cases) and real-world ground truth (6651 annotated fields from 90 pathology reports). Synthetic validation demonstrated strong performance (accuracy: 93.8%-99.0%). Real-world evaluation revealed field extraction recall ranging from 61.8% to 87.7%, demonstrating a substantial "reality gap" with performance drops of 11-32 percentage points. The gemini-2.5-pro model achieved the highest real-world recall (87.7%). Model size did not predict performance: the 14B-parameter deepseek-r1 (77.6%) outperformed its 70B-parameter counterpart (70.4%). The substantial performance degradation from synthetic to real-world data underscores the complexity of authentic clinical documentation. Smaller models can achieve competitive or superior recall, reducing computational costs. With even the best models missing 12%-38% of annotated fields, mandatory human verification is essential for clinical deployment. While LLM-based extraction systems show promise for pathology data extraction, synthetic validation alone provides false confidence. Rigorous real-world ground truth evaluation with expert annotation is essential before clinical deployment. These systems are best positioned as screening tools with mandatory human oversight rather than autonomous decision-making systems.

A digital twin (DT) for a robot can support its development and deployment; it is a valuable resource for simulation and monitoring. Creating a DT for a robot, however, is not an easy task, involving heterogeneous simulation models potentially developed by several stakeholders. This paper proposes a systematic and highly automated approach to develop a DT for a robot based on diagrammatic models and on an industry standard for co-simulation: the Functional Mockup Interface (FMI). Our modelling notation is RoboSim, a tool-independent framework to model, verify, and generate code for control software and for simulations of physical robotic platforms. We take advantage of RoboSim’s facilities for structured modelling and code generation to obtain results that help bridge the reality gap and produce DTs with less engineering effort. We present here our technique, using a manufacturing cell as a case study, and its assessment based on existing criteria for DT frameworks. The evaluation establishes that our technique provides significant coverage (specifically, 60%) of the Digital Twinning spectrum.

Multi-robot systems (MRSs) offer distinct advantages in large-scale exploration but require tight coupling between decentralized decision-making and collaborative estimation. This survey reviews learning-based multi-robot Active Collaborative Simultaneous Localization and Mapping (AC-SLAM), modeling it as a coupled system comprising a Decentralized Partially Observable Markov Decision Process (Dec-POMDP) decision layer and a distributed factor-graph estimation layer. By synthesizing these components into a conceptual framework, recent methods for cooperative perception, mapping, and policy learning are systematically critiqued. The analysis concludes that Hierarchical Reinforcement Learning (HRL) and graph-based spatial abstraction currently offer superior scalability and robustness compared to monolithic end-to-end approaches. Furthermore, a comprehensive analysis of Sim-to-Real transfer strategies is provided, ranging from domain randomization to emerging Real-to-Sim techniques based on NeRF and 3D Gaussian Splatting. Finally, future directions are outlined, moving from geometric mapping toward LLM-driven active semantic understanding and dynamic digital twins to bridge the reality gap.

Research on the implementation of cooperative behavior in actual autonomous mobile robots aimed at eliminating the reality gap

The study aims to explore the intersection of Operations Research (OR), Artificial Intelligence (AI), and simulation by examining how classical OR methodologies strengthen AI models, particularly in machine learning, robotics, natural language processing, and autonomous systems. It further investigates the critical role of simulation in training, testing, and validating AI algorithms, emphasizing its relevance for optimization, intelligent decision-making, and real-world system modelling. The study adopts a comprehensive analytical and literature-based methodology, reviewing foundational OR techniques, simulation principles, and modern AI applications. It synthesizes interdisciplinary research across mathematics, computer science, and engineering, supported by case analyses in robotics, healthcare, transportation, and autonomous systems. Additionally, a demonstration algorithm is developed to simulate automatic gear-control behavior in vehicles, illustrating how simulation models practically support AI-oriented operational decision-making. Findings reveal that OR optimization techniques significantly enhance AI efficiency, particularly in parameter tuning, resource allocation, and adaptive decision-making. Simulation is shown to be indispensable for AI training, offering controlled, safe, scalable, and cost-effective environments. The study identifies persistent challenges—including the reality gap, computational demands, and model bias—yet confirms that simulation and OR jointly accelerate AI development and broaden its practical reliability. The integrated OR–AI–simulation framework is applicable to numerous fields, including autonomous vehicle navigation, robotic motion planning, intelligent healthcare systems, logistics optimization, and smart city management. Industries benefit from improved forecasting, reduced operational costs, enhanced safety, and high-fidelity algorithm testing. Simulated environments also support reinforcement learning, surgical training, autonomous decision-making, and large-scale scenario evaluation, contributing to more efficient and intelligent real-world systems. The study’s novelty lies in its unified perspective that connects classical OR optimization principles with AI advancements through simulation-based experimentation. It uniquely synthesizes concepts from mathematics, computer science, and AI to highlight simulation as a bridge enabling intelligent automation. The inclusion of a practical simulation algorithm for automatic vehicle gear control further demonstrates how OR-driven simulation can concretely operationalize AI-based decision systems. Operations Research (OR) and Artificial Intelligence (AI) have both independently evolved as transformative fields which have shown impact on decision-making and problem-solving across diverse domains. While Operations Research makes available a foundation of mathematical modeling and optimization techniques, Artificial Intelligence sets up intelligence through learning, reasoning, and data-driven methods. This seminar paper presented by us explores how Operations Research gets involved in the development and enhancement of Artificial Intelligence (AI) systems. We, in this paper, have tried to discuss Key applications and case studies to highlight the synergies between these fields, remarkably in optimization, logistics, resource allocation, and automated decision-making. Simulation, an approach of Operations Research, has become a cornerstone in the field of Artificial Intelligence (AI), suggesting an experimental platform for testing hypotheses, training algorithms, and evaluating systems in controlled, cost-effective, and scalable environments. Our paper explores the central role simulations play in advancing Artificial Intelligence (AI) research and applications, with a focus on their integration in machine learning, robotics, and decision-making systems.

This article examines Indonesia’s autogate policy not only as a digital service innovation, but as a component of immigration system resilience. Drawing on policy implementation theory by Edwards III, Sabatier & Mazmanian, e-government literature on the design reality gap, system resilience thinking, and the information security CIA triad, the study analyses how autogate is designed, regulated, and operated at major international airports. Using a qualitative descriptive approach based on documentary and literature review, it maps legal foundations, organizational arrangements, and technical narratives surrounding automated border control. The findings show that autogate has improved processing times, standardized procedures, and strengthened traceability through digital logs, thereby supporting efficiency, accountability, and internal control. However, recurrent system downtime, biometric errors, incomplete system integration, and heavy dependence on infrastructure reveal significant vulnerabilities in confidentiality, integrity, and availability of data. Fragmented governance across immigration, airport operators, cybersecurity, and communications authorities reinforces a design–reality gap, limiting coordination, local ownership, and adaptive capacity. The article concludes that autogate currently functions as a double-edged innovation, symbolizing digital modernization while only partially reinforcing immigration system resilience. It recommends clearer mandates, stronger cross-sector coordination, improved data governance, and investment in human resources and risk-based oversight to align automated border control with a more robust, rights-sensitive model of digital immigration governance.

This study analyzes the essential correlation between the deployment of charging infrastructure and the adoption of electric vehicles (EVs) in the United States during 2024-2025, a timeframe characterized by unparalleled market volatility. By examining more than 204,000 public charging stations and 1.56 million EV sales in 2024, we measure the link between infrastructure and adoption and how it affects market stability. We found that EV adoption grew 114% in the same time that charging infrastructure grew 25% per year, which caused big supply-demand imbalances. The end of federal tax credits in September 2025 caused sales to drop by 46% in the fourth quarter of 2025. This showed how fragile growth based on policy can be. We show that the best EV-to-charger ratios (4 to 10 EVs per port) are linked to higher adoption rates in different regions. However, gaps in infrastructure in rural areas still make it hard for everyone to get to the chargers. This research offers essential insights for policymakers and industry stakeholders managing the shift from rapid growth to market development.

Growing consumer demand for sustainable food products has amplified the use of “green” marketing claims, yet transnational agri-food supply chains face a critical “perception–reality gap” due to data fragmentation and the absence of independent verification, fostering significant greenwashing risks. This study explores how the synergistic integration of Blockchain, Internet of Things (IoT), and Artificial Intelligence (AI) can bridge this gap. Utilizing a PRISMA-inspired qualitative systemic analysis and scenario modeling, we propose the “Converging Technologies for Sustainable Agri-Food” (CTSAF) model, formalized through a mathematical Green Claim Veracity Index (Vi) and AI-driven anomaly detection algorithms. The analysis evaluates three maturity-level scenarios against expert-calibrated Key Performance Indicators (KPIs). Results demonstrate that while traditional and blockchain-only systems remain vulnerable to the “Oracle Problem”, the integrated CTSAF model (Scenario III) achieves “Very High” performance in data accuracy and audit efficiency. By transforming passive record-keeping into an autonomous governance layer, this framework provides a strategic roadmap for substantiating environmental claims in alignment with the EU Green Claims Directive and the Digital Product Passport framework.

The focus of this study is to examine the lived experiences of third- and fourth-year computer engineering students at Bulacan State University-Main Campus with regards to their academic and laboratory needs in the face of the aftermath of calamities. The study examines how students deal with the adverse effects of resource constraints by examining how the "Bahala Na" sociocultural narrative manifests itself as a coping mechanism for stress on students. The qualitative phenomenological method that the researchers employed was based on the Socio-Cultural Resource Conservation Model, which combines Gripaldo's Theory of Bahala Na as Theistic Circumstantialism with Hobfoll's Conservation of Resources (COR) Theory. Researchers utilized purposive sampling, choosing cases with valuable and meaningful data from just third- and fourth-year Computer Engineering students. Using a semi-structured interview approach, which was validated by professionals in the fields of psychology and disaster risk reduction, the researchers gathered data and then utilized Braun and Clarke's six-phase thematic analysis framework to analyze the participant narratives. Due to the lack of access to physical labs, the students experienced "Logistical Paralysis," as evidenced by the "No Power, No Code" situation, as well as a "Simulation vs. Reality Gap." The students' learning experiences also showed "Communal Resourcefulness," while their "Bahala Na" coping strategy ranged from "Passive Bahala Na" (fatalistic resignation) to "Active Bahala Na" (courage in uncertainty). This study shows that "Bahala Na" is a resource-driven reaction brought on by the significant lack of material and environmental resources during post-calamity, rather than an underlying laziness. The research suggests creating institutionalized "Compassionate Flexibility" with deadlines that are specific to the situation, creating disaster-resilient lesson plans using asynchronous learning, and establishing community centers or student-friendly evacuation shelters with dependable power and internet connectivity.