The effect of autonomous mobile robots on improving endoscope reprocessing efficiency: a prospective comparative trial (with video).

Marine microplastic pollution presents a critical environmental challenge, affecting ecosystems, wildlife, and human health as millions of tons of plastic waste enter oceans each year. Microplastics, due to their small size, are difficult to detect and accumulate widely in marine environments, where they integrate into the food web. Artificial Intelligence (AI) offers promising advancements for modeling, detecting, and mitigating the effects of microplastics in marine ecosystems. This narrative review examines recent developments in AI applications for addressing microplastic pollution. The review focuses on AI-driven modeling for predicting microplastic flows, intelligent waste detection systems that utilize remote sensing and autonomous robotics, and AI-based interventions aimed at reducing microplastic release. AI-driven models enhance the accuracy of predicting microplastic accumulation zones, supporting targeted clean-up efforts and informed policy-making. Advanced detection systems provide real-time monitoring over extensive areas, while AI-based filtration and material innovation technologies help reduce microplastic pollution at the source. AI holds significant potential to mitigate marine microplastic pollution, yet challenges such as data availability, model refinement, and global collaboration remain. Future research should focus on enhancing AI models, refining detection systems, and encouraging international data-sharing and cooperation. Collaboration across sectors is essential to fully leverage AI's potential in safeguarding marine ecosystems from microplastic pollution.

Multi-session perception-aware coverage path planning for active semantic SLAM and automatic change detection

Efficient path planning for autonomous mobile robots in e-commerce warehouses: An improved particle swarm optimization approach with dynamic replanning

Causality-enhanced decision-making for autonomous mobile robots in dynamic environments

Soft robots enable compliant interaction with humans and the environment. However, their widespread deployment is constrained by significant challenges, including limited mobility and the inherent complexity of incorporating power and control systems into their bodies. In this paper, we present a fully integrated electrohydraulic soft robotic fish that achieves three-dimensional maneuverability and autonomous control. To achieve effective underwater locomotion, the robot is actuated by two flapping wings, each independently controlled and powered by electrohydraulic actuators. Theoretical modeling and finite element simulation of the electrohydraulic actuator are conducted and validated through experimental results. Furthermore, the robot's electronic system integrates a range of miniaturized components—including a control board, voltage amplifiers, and sensors. This autonomous soft robot enables depth control and directional navigation. The design and control methods of this work can inspire the design of fully autonomous and intelligent soft robots for complex tasks and environments.

• Provides a unified taxonomy that organizes more than 100 VLA architectures. • Maps 26 major VLA datasets using a framework based on task difficulty and modality richness. • Presents a large-scale quantitative analysis linking model design choices to normalized performance. • Demonstrates that diffusion-based decoders and hierarchical fusion significantly improve manipulation success. • Introduces the VLA-FEB benchmark with new metrics for measuring multimodal fusion quality and alignment. • Proposes an agentic VLA framework where LLM planners verify and re-plan actions using uncertainty-driven feedback for self-improving robotic autonomy. Vision Language Action (VLA) models represent a new frontier in robotics by unifying perception, reasoning, and control within a single multimodal learning framework. By integrating visual, linguistic, and action modalities, they enable multimodal fusion systems designed for instruction-driven manipulation and generalist autonomy. This systematic review synthesizes the state of the art in VLA research with an emphasis on architectures, algorithms, and applications relevant to robotic manipulation. We examine 102 models, 26 foundational datasets, and 12 simulation platforms, categorizing them according to their fusion strategies and integration mechanisms. Foundational datasets are evaluated using a novel criterion based on task complexity, modality richness, and dataset scale, allowing a comparative analysis of their suitability for generalist policy learning. We further introduce a structured taxonomy of fusion hierarchies and encoder-decoder families, together with a two-dimensional dataset characterization framework and a meta-analytic benchmarking protocol that quantitatively links design variables to empirical performance across benchmarks. Our analysis shows that hierarchical and late fusion architectures achieve the highest manipulation success and generalization, confirming the benefit of multi-level cross-modal integration. Diffusion-based decoders demonstrate superior cross-domain transfer and robustness compared to autoregressive heads. Dataset analysis highlights a persistent lack of benchmarks that combine high-complexity, multimodal, and long-horizon tasks, while existing simulators offer limited multimodal synchronization and real-to-sim consistency. To address these gaps, we propose the VLA Fusion Evaluation Benchmark to quantify fusion efficiency and alignment. Drawing on both academic and industrial advances, the review outlines future research directions in adaptive and modular fusion architectures, computational resource optimization, and the deployment of interpretable, resource-efficient robotic systems. We further propose a forward-looking agentic VLA paradigm where LLM planners integrate VLA skills as verifiable tools within a closed feedback loop for adaptive and self-improving robotic control. This work provides both a conceptual foundation and a quantitative roadmap for advancing embodied intelligence through multimodal information fusion across robotic domains. A public repository summarizing models, datasets, and simulators is available at: https://muhayyuddin.github.io/VLAs/ .

The Autonomous Olfactory Canine Robot (AOCR) is an advanced electronic nose (e-nose) engineered for trace gas monitoring and environmental odor classification. Its primary purpose is to overcome the severe electrical noise and baseline drift inherent in Metal-Oxide Semiconductor (MOS) gas sensors by utilizing a highly isolated dual-controller architecture. An Arduino Mega 2560 safely manages 5V high-current sensor heaters and precise 16-bit analog-to-digital conversions, while an ESP32 handles 3.3V logic, WiFi synchronization, and environmental compensation. The AOCR successfully mitigates hardware interference, calculates normalized chemical "Signatures" from its eight-sensor array, and securely uploads a comprehensive 66-column dataset to a Google Sheets cloud database to facilitate real-time monitoring and future machine learning analysis.

The problem of optimizing the deployment of modern computer vision models on small-space embedded systems using specialized neuroprocessors (NPU) is being solved. The target platform is the Orange Pi 5 single-board computer based on the Rockchip RK3588S system-on-chip, which integrates a 6-TOPS NPU. The study encompasses the complete pipeline for adapting the YOLOv11 architecture to embedded execution from operator compatibility analysis and structural model modifications to meet hardware constraints, to the implementation of a real-time, high-throughput video processing pipeline. We present a detailed methodology for converting models from PyTorch to the vendor-specific RKNN format using post-training quantization to INT8 precision, which delivers substantial inference acceleration and memory footprint reduction with minimal accuracy loss. To overcome the inherently blocking nature of NPU inference, we propose a multiprocessing video processing architecture that employs parallel worker processes. Through extensive experimentation, we identify the optimal number of concurrent processes for different YOLOv11 variants (n, s, m). Our implementation achieves 54 FPS for YOLOv11-n, 48 FPS for YOLOv11-s, and 27 FPS for YOLOv11-m at 640 × 640 input resolution. Crucially, we demonstrate that exceeding the optimal process count saturates memory bandwidth, increases SoC temperature, and reduces energy efficiency without improving throughput. These findings validate the feasibility of building cost-effective, energy-efficient, and high-performance computer vision systems using widely available single-board computers. The results are directly applicable to real-time use cases such as autonomous drones, robotics, smart surveillance, and edge AI applications where low latency and hardware accessibility are critical factors.

The progression of mobile robotic systems designed for operation in challenging and unstructured environments has become increasingly important in areas such as space exploration, military missions, and disaster response. Early robotic platforms relied primarily on simple wheel or trackbased locomotion, which, although effective on flat surfaces, often struggle to maintain stability and traction on uneven terrain. As a result, advanced mobility mechanisms like the rocker-bogie suspension system have gained attention due to their ability to distribute load efficiently and adapt to rough surfaces, as demonstrated in planetary rover missions. Despite these improvements, many conventional systems still face limitations in handling unpredictable environments, often experiencing instability, reduced grip, and difficulty in overcoming obstacles, which ultimately affects their operational efficiency and autonomy. To overcome these challenges, this work proposes an autonomous mobile robot built upon a rocker-bogie suspension framework, designed to enhance terrain adaptability and movement stability. The system employs a six-wheel configuration with independent motor control, allowing better traction and smoother navigation across irregular surfaces. Additionally, integrated sensors enable real-time obstacle detection and environmental awareness, supporting intelligent navigation without continuous human intervention. By combining mechanical robustness with sensor-driven autonomy, the proposed system aims to improve performance, reliability, and efficiency, making it highly suitable for deployment in hazardous or inaccessible environments where traditional robotic systems are less effective.

Reliable indoor navigation for mobile robots in industrial environments remains a challenging task due to the limitations of conventional control methods. Many existing systems depend on open-loop motor control or basic proximity sensors, which often lead to cumulative positioning errors, wheel slippage on smooth surfaces, and the inability to verify actual position against planned coordinates. To overcome these issues, this work presents a color-guided autonomous rover that integrates sensor fusion with an intelligent navigation approach to achieve accurate localization within a structured grid environment. The system is built around an ESP32 controller and incorporates an MPU6050 inertial measurement unit along with wheel encoders to enable closed-loop PID control, allowing continuous correction of motion deviations for precise linear and rotational movements. A dual I2C communication architecture is implemented to separate high-frequency motion sensing from color detection operations carried out by the TCS34725 sensor, thereby improving processing efficiency. The navigation mechanism is driven by a custom algorithm that converts grid-based target positions into directional commands and distance measures. At each designated node, the rover validates its position using predefined color markers embedded in the grid layout. Additionally, a WebSocketenabled interface supports both manual adjustment and autonomous operation, enhancing system flexibility. Experimental results indicate that the integration of inertial feedback significantly reduces drift compared to traditional approaches, demonstrating a reliable and intelligent solution for precision-oriented indoor robotic navigation.

The rapid growth of Industry 4.0 technologies has had a significant impact on warehouse operations. Because of this, intelligent, adaptable, and autonomous systems must now be able to work well in places where things are always changing. Autonomous Mobile Robots (AMRs) are now a key way to automate tasks like moving things, keeping track of stock, and running the logistics process. But most traditional AMR systems use static navigation methods and can't change to deal with moving individuals or objects in real time. This rule makes stores less safe and less efficient. This study presents an AI-driven Autonomous Mobile Robot integrated with vision-based Simultaneous Localization and Mapping (SLAM) for intelligent navigation in warehouses. The proposed system employs both LiDAR and camera-based perception techniques to generate real-time occupancy grid maps and ascertain the robot's precise location in intricate environments. A* is used in a path planning algorithm to find the best routes for navigation. A new risk assessment module that uses artificial intelligence checks the environment all the time and changes how the robot acts when it needs to.

The article presents an analysis of the two-year experience of preparing the MKA team from Yakutsk for the All-Russian Robofest Robotics Festival in the AutoNet14+ category. The relevance of the research is determined by the need to identify effective pedagogical mechanisms for the formation of engineering thinking among schoolchildren. Based on the analysis of regulations, engineering documentation and a retrospective analysis of the decisions made, it is shown how key competencies of students are formed during the design and programming of an autonomous robot: analysis of technical requirements, development of alternative solutions, work in conditions of incomplete information and time constraints. Special attention is paid to the concept of "speed of engineering thinking", implemented in the team's ability to quickly refine the design and make the necessary changes to the program code. The result of the formed competencies was the victory at the Robofest robotics festival. Many years of experience in participating in robotics competitions allows us to build an individual preparation trajectory for the All-Russian Olympiad of Schoolchildren in Informatics of the Robotics profile.

We propose In Roboto Experimentation as a fourth fundamental modality for biological inquiry, succeeding and complementing in vivo, in vitro, and in silico methods. It is defined as the systematic execution of biological experiments by AI-directed, closed-loop robotic platforms that autonomously perform the core experimental cycle of observation, classification, decision-making, and physical manipulation of biological samples without human intervention over extended timescales. This modality emerges from the necessity to interrogate biological regimes that are intrinsically inaccessible to human experimenters due to limitations of temporal scale, timing precision, and operational consistency. This perspective paper introduces the conceptual framework for In Roboto Experimentation and presents the design rationale for the first fully realized platform of this kind. The platform is a closed-loop microscope-AI-laser system engineered to execute a six-month experiment to test the Centriolar Damage Accumulation Theory of Aging (CDATA). It integrates Recombination-Induced Tag Exchange (RITE) fluorescent centriole tagging, real-time AI vision for segmentation and tracking (CellPose 3.0), autonomous 405 nm laser microablation, and a bounded Claude Code supervisory agent governed by a pre-registered declarative protocol (PROMPT.md). Sections §5 through §10 of the full paper provide complete technical specifications covering hardware integration, software architecture, the operational autonomous loop, and safety interlocks. The primary application is the definitive validation or falsification of the hypothesis that centriolar damage serves as a cell-intrinsic aging counter. This work establishes the epistemological and technical foundations for a new class of experiments that leverage robotic autonomy to explore biological questions at previously unreachable scales and durations.

Abstract Smart farming is constitutes a means of facilitating the European Green Deal and the Farm to Fork strategy. However, credible sustainability results require measurable and validatable indicators, verifiable data and automation that is reliable even under field conditions. This overview study presents developments in IoT-based sensing, artificial intelligence-based analysis and autonomous robotics, and links them to EU target areas. The peer-reviewed studies (2000–2025) were retrieved from the Scopus, Web of Science, and Google Scholar databases and supplemented with key EU legal and strategic documents. The article proposes a policy-driven digital agroecological management (PDAM) framework that establishes adaptable indicators based on past trends in EU targets (pesticides, nutrients, soil, biodiversity and climate) and then develops a system of perception-analysis-implementation based on these indicators. The most effective tools support GNSS-based machine control, variable rate application, and remote sensing, while AI-based decision support tools, autonomous weed control, and digital twin field validation are still weaker. Interoperability, data governance, cybersecurity and safety regulation emerge as critical scaling constraints for auditable smart farming systems.

The development of mobile robots for rough terrain exploration has gained significant importance in recent years, especially in applications such as planetary exploration, military operations, and disaster management. Historically, robotic mobility systems have evolved from simple wheeled platforms to more advanced suspension mechanisms, notably the rocker-bogie system, which was popularized through its successful implementation in Mars rovers. Despite these advancements, conventional mobile robots using standard wheel or track mechanisms face challenges in navigating uneven and unpredictable terrains. The primary problem addressed in this project is the limited adaptability and stability of traditional robotic systems when operating on rugged surfaces. Conventional systems often suffer from poor traction, instability, and inability to overcome obstacles, which restricts their usability in critical environments. Additionally, these systems may require constant human intervention, reducing their efficiency in autonomous operations. To overcome these limitations, there is a need for a robust and intelligent robotic system capable of traversing complex terrains with minimal human control. The proposed system focuses on designing an autonomous mobile robot based on the rocker-bogie suspension mechanism. This system integrates a six-wheel configuration with independent motor control and sensor-based obstacle detection to ensure stability, flexibility, and adaptability. The significance of this project lies in its potential to enhance robotic mobility in challenging environments. By combining mechanical innovation with autonomous control, the proposed robot offers improved performance, reliability, and efficiency, making it suitable for real-world applications where human access is limited or hazardous.

Emergency departments (EDs) are fast-paced, dynamic, safety-critical spaces where clinicians are overworked and underpaid. To support clinicians, researchers are exploring the contextualization and development of clinically assistive robots (CARs) that can assume non-critical tasks to reduce clinician overload. In this article, we introduce Clinically Assistive Robot System for Emergency Medicine (CAR-EM), collaboratively developed with ED clinicians. CAR-EM includes an autonomous robot and a task specification interface. It completes tasks by leveraging control synthesis, a framework that automatically transforms high-level tasks into control while providing guarantees and feedback. We conducted a feasibility study across two different hospital EDs, where interprofessional clinicians tasked the robot to perform patient assessments and item deliveries. Clinicians found the system easy to use, and particularly helpful to offload busywork. This work demonstrates control synthesis as a feasible tool to develop autonomy for robots in safety-critical spaces, and identifies considerations for failure interventions. We also discuss ethical considerations for deploying robots in hospitals, including healthcare worker displacement and work disruption. Thus, our work: (1) highlights the unique requirements of situating robots in real world hospital EDs, and (2) demonstrates a novel approach leveraging guarantees and feedback from control synthesis methods to successfully implement context-specific CAR behaviors. Through this work, we aim to further research for safer and more reliable robots in real world, uncertain environments.

Purpose The purpose of this paper is to provide details of the role of autonomous robots in the transportation and delivery sector of the logistics industry. Design/methodology/approach Following an introduction, this first discusses autonomous cargo shipping. It then considers autonomous road transport which is followed by details of short-distance and last-mile delivery robots. Finally, conclusions are drawn. Findings Transportation and delivery is a central element of the logistics industry but the sector faces operational, environmental and economic challenges. The industry is increasingly adopting automation and robotics, and this is now being extended to the transportation and delivery sector. Autonomous cargo ships have been demonstrated and have the potential to reduce costs and improve safety. The main barriers to widespread adoption are regulations and legislation. Autonomous road transport is gaining momentum and autonomous trucks are increasingly being deployed, particularly in the USA, where they operate on hub-to-hub routes. Short-distance and last-mile delivery is a rapidly growing sector and drones, autonomous vehicles and sidewalk robots are used widely to deliver foods, groceries and mail. Autonomous vehicles confer benefits which include improved safety, reduced environmental impact, lower fuel and labour costs and more rapid delivery and represent the most significant development in the drive to achieve full automation of the logistics industry. Originality/value This provides details of the rapidly growing role of robotics in the transportation and delivery sector of the logistics industry.

Abstract The rapid growth of educational institutions and the exponential increase in library resources have exposed inherent limitations in traditional library management systems, which rely heavily on manual processes to manage books, user authentication, and inventory tracking. This research paper presents a comprehensive study of a Smart Library Management System (SLMS). The proposed system integrates Internet of Things (IoT) technologies with Radio Frequency Identification (RFID), biometric fingerprint authentication, and autonomous robotics, to automate and optimize library operations. The implementation combines a web-based management interface with an Arduino-powered autonomous robot, capable of navigating to proper location and identifying preferred book. This paper details the system’s design, methodology, results, and broader implications, situating the work within the context of contemporary research on digital automation and information management infrastructures. The work demonstrates significant improvements in operational efficiency, accuracy, and user experience, while also discussing the challenges and limitations inherent to such technological transformations. Keywords: Smart Library Management, Internet of Things (IoT), RFID Technology, Biometric Authentication, Arduino, Library Automation

Abstract Autonomous remote‐sensing technologies are increasingly contributing to biodiversity monitoring by enabling scalable, repeatable, and minimally invasive data collection. We present a ground‐based robotic remote‐sensing framework that integrates artificial intelligence and standardized quality assurance to support the derivation of decision‐ready ecological indicators. Using European coastal dunes as a case study, we deployed an AI‐enabled quadruped robot equipped with near‐ground imaging sensors to monitor the host–herbivore interaction between Pancratium maritimum and Brithys crini . In this citizen‐to‐robot pipeline, expert‐verified citizen‐science imagery was used to train lightweight detection models for on‐board inference and higher‐capacity models for offline auditing, ensuring reproducibility and transparency across missions. Field trials demonstrated that the system achieved consistent image quality, accurate detections, and low‐disturbance operation under natural conditions, capturing spatially explicit evidence of herbivory and host condition. By coupling standardized protocols with robotic autonomy, this approach implements a proximal remote‐sensing layer that complements aerial and satellite observations. The workflow is designed to support transferable quantification of species interactions and habitat condition across sites and seasons, contributing to the integration of robotics and ecological remote sensing for biodiversity assessment and conservation management.