Given the recent global pandemic and domestic labor shortage, there is a pressing demand for digital means that enable construction site managers to obtain information more efficiently in support of their daily tasks. Mobile workers at the site often find traditional software applications, which are structured around forms and require multiple finger actions such as key presses and clicks, to be inconvenient, thereby diminishing their willingness to use such systems. A chatbot, or conversational AI, can make a system more user-friendly and accessible by offering an intuitive way for users to interact with it. This research introduces a clearly demonstrated Natural Language Understanding (NLU) model and prototypes an AI-powered chatbot system that supports site managers in their everyday tasks, specifically for inquiries regarding the dimensions of building components. The chatbot's answering component utilizes Building Information Modeling (BIM) methodologies. The chatbot's preliminary testing revealed its effectiveness in identifying the intents and entities behind the queries made by site managers, achieving a satisfactory level of accuracy for both intent prediction and the provided response. Site managers can now leverage alternative approaches for obtaining the information they need, as indicated by these results.
With Industry 4.0's impact, physical and digital systems have undergone a complete revolution, leading to optimized digitalization strategies for maintenance plans of physical assets. Road network conditions and the prompt implementation of maintenance schedules are fundamental to the success of predictive maintenance (PdM) in road infrastructure. To accurately and swiftly detect and classify road crack types, we devised a PdM approach that relies on pre-trained deep learning models. This study examines how deep neural networks can be used to categorize roads depending on the level of deterioration. By training the network, we enable it to identify a variety of road defects, including cracks, corrugations, upheavals, potholes, and other types. Evaluating the total damage inflicted, considering its severity, we can pinpoint the degradation rate and develop a PdM framework to pinpoint the frequency of damage occurrences, thereby enabling prioritized maintenance actions. By employing our deep learning-based road predictive maintenance framework, inspection authorities and stakeholders can resolve maintenance issues concerning specific damage types. Our proposed framework demonstrated impressive performance, as assessed by precision, recall, F1-score, intersection-over-union, structural similarity index, and mean average precision metrics.
Employing convolutional neural networks (CNNs), this paper details a method for fault detection within the scan-matching algorithm to enhance SLAM precision in dynamic environments. Changes in the environment, as perceived by a LiDAR sensor, occur when dynamic objects are present. Consequently, the process of aligning laser scans using scan matching is prone to failure. Consequently, a more resilient scan-matching algorithm is required for 2D SLAM to address the shortcomings of existing scan-matching methods. Laser scan data from a 2D LiDAR, originating from an environment of unknown characteristics, is processed initially. This is subsequently subjected to ICP (Iterative Closest Point) scan matching. Image conversion of the matched scans is then performed, with these images being used to train a CNN model to identify flaws related to the scan matching. Following training, the trained model determines the faults present in new scan data. Real-world scenarios are incorporated into the diverse dynamic environments utilized for training and evaluation. The proposed method proved highly accurate in identifying scan matching failures within every tested experimental environment.
This paper details a multi-ring disk resonator, featuring elliptic spokes, designed to compensate for the anisotropic elasticity of (100) single-crystal silicon. By using elliptic spokes instead of straight beam spokes, the structural coupling between each ring segment can be manipulated. The degeneration of two n = 2 wineglass modes can be a result of the strategically optimized design parameters of the elliptic spokes. Employing a design parameter of 25/27 for the aspect ratio of the elliptic spokes, a mode-matched resonator was obtained. Agomelatine mouse Evidence for the proposed principle was provided by both numerical simulations and physical experiments. medical risk management Demonstrating an experimentally validated frequency mismatch of just 1330 900 ppm, the current study notably outperforms the 30000 ppm maximum achievable by conventional disk resonators.
In the field of intelligent transportation systems (ITS), the increasing use of computer vision (CV) applications is a direct consequence of technological advancements. The aim of these applications is to increase the intelligence, enhance the efficiency, and improve the safety of traffic within transportation systems. Progress in computer vision systems demonstrably impacts the resolution of problems encountered in traffic surveillance and regulation, event detection and handling, dynamic road pricing methodologies, and ongoing road condition assessments, and numerous other crucial aspects, by means of more effective techniques. This survey investigates the use of CV applications in literature, examining machine learning and deep learning methodologies within Intelligent Transportation Systems (ITS), the practicality of computer vision in ITS, the benefits and challenges posed by these technologies, and future research directions aimed at enhancing ITS effectiveness, efficiency, and safety. The review, which amalgamates research from diverse sources, strives to illustrate how computer vision (CV) techniques facilitate the development of smarter transportation systems. It presents a complete examination of computer vision applications within intelligent transportation systems (ITS).
The past decade has witnessed significant progress in deep learning (DL), which has profoundly benefited robotic perception algorithms. Undeniably, a substantial component of the autonomous system architecture across different commercial and research platforms is contingent on deep learning for situational understanding, particularly from visual sensor input. A study was conducted to assess the applicability of general-purpose deep learning algorithms, focusing on detection and segmentation networks, in processing image-analogous output from cutting-edge lidar. This effort, to the best of our knowledge, is the initial work to focus on low-resolution, 360-degree lidar images, rather than the complex task of processing 3D point clouds. The pixels in these images store depth, reflectivity, or near-infrared information. Sentinel node biopsy Through suitable preprocessing, we demonstrated that universal deep learning models can handle these images, thereby enabling their application in environmental scenarios where visual sensors have inherent limitations. Evaluating a range of neural network architectures, our study employed both a qualitative and quantitative methodology to assess their performance. Using deep learning models for visual camera data yields considerable benefits, particularly due to their greater availability and maturity than counterparts based on point cloud processing.
Thin composite films, comprising poly(vinyl alcohol-graft-methyl acrylate) (PVA-g-PMA) and silver nanoparticles (AgNPs), were deposited using the blending approach, also termed the ex-situ method. A copolymer aqueous dispersion was formed via the redox polymerization of methyl acrylate (MA) on poly(vinyl alcohol) (PVA), with ammonium cerium(IV) nitrate serving as the initiator. The polymer was then blended with AgNPs, which were synthesized through a green approach using water extracts of lavender, a by-product of the essential oil industry. To determine nanoparticle dimensions and assess their stability in suspension over 30 days, dynamic light scattering (DLS) and transmission electron microscopy (TEM) techniques were applied. The spin-coating process was used to deposit PVA-g-PMA copolymer thin films, containing different volume fractions of AgNPs (0.0008% to 0.0260%), onto silicon substrates, allowing for the investigation of their optical properties. Employing the combination of UV-VIS-NIR spectroscopy and non-linear curve fitting, the refractive index, extinction coefficient, and thickness of the films were quantified; furthermore, room-temperature photoluminescence measurements were carried out to investigate the emitted light from the films. Measurements of film thickness dependence on nanoparticle concentration demonstrated a consistent linear increase, ranging from 31 nm to 75 nm as the weight percent of nanoparticles rose from 0.3 wt% to 2.3 wt%. The sensing properties of films toward acetone vapors were determined by measuring reflectance spectra in a controlled environment before and during exposure to the analyte molecules at the same location; the swelling degree of the films was subsequently quantified and compared to the corresponding undoped films. Films containing 12 wt% AgNPs exhibited the best sensing response to acetone, as demonstrated. The properties of the films were evaluated, and the effect of AgNPs was both uncovered and detailed.
Magnetic field sensors, crucial for advanced scientific and industrial equipment, must exhibit both reduced dimensions and enhanced sensitivity across a broad spectrum of magnetic fields and temperatures. Nevertheless, commercial sensors are scarce for gauging high magnetic fields, spanning from 1 Tesla to megagauss. In light of this, the search for advanced materials and the engineering of nanostructures displaying exceptional properties or novel phenomena is critical for applications in high-field magnetic sensing. The central theme of this review revolves around the investigation of thin films, nanostructures, and two-dimensional (2D) materials, which show non-saturating magnetoresistance across a broad range of magnetic fields. Results from the review illustrated how manipulating the nanostructure and chemical composition of thin polycrystalline ferromagnetic oxide films, specifically manganites, led to an outstanding colossal magnetoresistance, exceeding megagauss values.