Floor Classification Research Articles

HyTra: Hyperclass Transformer for WiFi Fingerprinting-based Indoor Localization

The emerging demand for a variety of novel Location-based Services (LBS) by consumers and industrial users is driven by the rapid and extensive proliferation of mobile smart devices. Sensors embedded in smart devices or machines provide wireless connectivity and Global Positioning System (GPS) capability, and are co-utilized to acquire location-linked data which are algorithmically transformed into reliable and accurate location estimates. GPS is a mature and reliable technology for outdoor localization but indoor localization in a complex multi-storey building environment remains challenging due to fluctuations in wireless signal strength arising from multipath fading. Location-linked data from wireless access points (WAPs) such as received signal strength (RSS) are acquired as numerical sequences. By conceptualizing a fixed order sequence of WAP measurements as a sentence where the RSS from each WAP are words, we may leverage on recent advances in artificial intelligence for natural language processing (NLP) to enhance localization accuracy and improve robustness against signal fluctuations. We propose the hyper-class Transformer (HyTra), an encoder-only Transformer neural network which learns the relative positions of wireless access points (WAPs) through multiple learnable embeddings. We propose a second network, HyTra-HF, which improves upon HyTra by applying a hierarchical relationship between location classes. We test our proposed networks on public and private datasets varying in sizes. HyTra-HF outperforms existing deep learning solutions by obtaining 96.7\% accuracy for the floor classification task on the UJIIndoorloc dataset. HyTra-HF is amenable to deep model compression and achieves accuracy of 95.95\% with over ten-fold reduction in model size using Sparsity Aware Orthogonal (SAO) initialization and has the best-in-class accuracy for the sparse model.

The Zygomatic Anatomy-Guided Approach, Zygomatic Orbital Floor Classification, and ORIS Criteria-A 10-Year Follow-Up.

Presently, the management of patients with maxilla bone defects of the Cawood V or VI class is achieved using zygomatic or individual implants or through augmentation of the bone. For zygomatic implants, the ORIS criteria represent the most common factor in helping practitioners register success rates. The zygomatic anatomy-guided approach (ZAGA)and zygomatic orbital floor (ZOF) are factors that are crucial to examining the anatomy of a particular patient before the procedure. The aim of thisarticle is to find the statistical relationship between the abovementioned terms and other factors. A total of 81 patients underwent zygomatic implant procedures in different configurations. The ORIS, ZAGA, and ZOF parameters were compared with other factors such as type of surgery, sex, age, and the anatomy of the zygomatic bone. Most patients in this article were classified as ZAGA Class 2. The relationships between type of surgery and ZAGA classification, and ZAGA and sinus/maxilla zygomatic implant localization were statistically significant. The ZAGA and ZOF scales are practical and valuable factors that should be taken into account before surgery, whereas to date, criteria better than the ORIS scale have not been described in terms of the success of zygomatic implants. The ZOF scale might omit perforation of the orbit because this parameter warns a practitioner to be aware of the anatomy of the orbit.

Wi-Fi indoor 3D localization algorithm based on multi-classifier fusion

Indoor localization technology based on Received Signal Strength (RSS) fingerprint is widely used in life and industry. Compared with the traditional localization methods, the localization technology integrating multiple machine learning methods has better localization accuracy. However, in multi-floor localization, the existing fusion localization technology ignores the interaction between different floors in the localization process, resulting in low indoor localization accuracy. This paper proposes a Wi-Fi indoor 3D localization method based on multi-classifier fusion named FLMCF. Firstly, floor classification training is carried out to reduce the location deviation in the vertical direction. Secondly, for each floor, multiple classifiers are used for model training and training the optimal weight set by minimizing the average localization error. In this case, the advantages of each classifier can be fully integrated to improve the localization accuracy. Finally, the Reliability Fusion Weight Selection (RFWS) algorithm determines the weight and calculates the final location estimation. The experimental results indicate that FLMCF is nearly 14.1% better than DIFMIC in the 90th percentile of CDF.

Multi-Task Neural Network for Position Estimation in Large-Scale Indoor Environments

Pedestrian localization within large-scale multi-building/multi-floor indoor environments remains a challenging task. Fingerprinting-based approaches are particularly suited for such large-scale deployments due to their low requirements of hardware installments. Recently, the fingerprinting problem has been addressed by deep learning. Existing models are mostly task specific by providing floor classification or position estimation within a small area. A strategy to support localization within large-scale environments is to sequentially apply hierarchical models. This has several drawbacks including missing scalability and increased deployment complexity on smartphones. We propose a unifying approach based on training a single neural network that classifies the building/floor and predicts the position in a single forward-pass of the network. Our model classifies a grid cell and performs within grid cell regression, which solves the performance degradation of applying regression within large areas. To reduce the error in case of misclassified grid cells, we propose a novel technique called multi cell encoding learning (multi-CEL), where a model simultaneously learns several redundant position representations within an overlapping grid cell encoding. For three public WLAN fingerprinting datasets, we demonstrate that multi-CEL surpasses existing state-of-the-art multi-task learning neural networks and even outperforms regression neural networks explicitly trained for 2D-positioning by up to 17%.

Real-Time LSTM-RNN Classification of Floors With Different Friction Coefficients for a Walking Humanoid Robot Wearing a 3D Force System

In the study of biped humanoid robots it is crucial to achieve high precision and robustness in locomotion. Humanoid robots that operate in real world environments need to be able to physically recognize different grounds to best adapt their gait without losing their dynamic stability. This work proposes a technique to classify in real time the type of floor from a set of possibilities learnt off-line. Hence, the paper describes the collection and preparation of a dataset of contact forces, obtained with a wearable instrumented system, mixed with the information of the robot internal inertial sensor to classify the type of underlying surface of a walking humanoid robot. For this classification, the data are acquired for four different slippery floors at a rate of 100 Hz and it is used as input for a long short-term memory (LSTM) recurrent neural network (RNN). After testing different learning models architectures and tuning the models parameters, a good mapping between inputs and targets is achieved with a test classification accuracy greater than 92%. A real time experiment is presented to demonstrate the suitability of the proposed approach for the multi-classification problem addressed.

A Fast Online Classification Method of Solid Wood Floors Based on Stochastic Sampling and Machine Learning

Solid wood floors are widely used as an interior decoration material, and the color of solid wood surfaces plays a decisive role in the final decoration effect. Therefore, the color classification of solid wood floors is the final and most important step before laying. However, research on floor classification usually focuses on recognizing complex and diverse features but ignores execution speed, which causes common methods to not meet the requirements of online classification in practical production. In this paper, a new online classification method of solid wood floors was proposed by combining probability theory and machine learning. Firstly, a probability-based feature extraction method (stochastic sampling feature extractor) was developed to obtain rapid key features regardless of the disturbance of wood grain. The stochastic features were determined by a genetic algorithm. Then, an extreme learning machine—as a fast classification neural network—was selected and trained with the selected stochastic features to classify solid wood floors. Several experiments were carried out to evaluate the performance of the proposed method, and the results showed that the proposed method achieved a classification accuracy of 97.78% and less than 1 ms for each solid wood floor. The proposed method has advantages including a high execution speed, great accuracy, and flexible adaptability. Overall, it is suitable for online industry production.

A Study of the Direction of Evacuation Evaluation According to the Fire Alarm Method in Performance-Based Design

In performance-based design applied to domestic buildings, when evaluating evacuation by fire alarm, safety evaluation by single-floor horizontal evacuation contradicts the priority alarm method of the immediately upper and four higher floors, which is the fire alarm standard of the current firefighting design, making it difficult to improve evacuation safety through safety in performance-based design. Therefore, by analyzing the domestic fire alarm design standards and evaluating the evacuation simulation case by case for a single floor, immediate upper floor, and four floors directly above, the evacuation evaluation is performed through priority alarm evacuation according to the size of the building. It was judged to be suited to actual conditions. Therefore, a performance-based design is proposed as a method of evaluating evacuation safety by including the floor classification and providing an evacuation personnel calculation reflecting the priority alarm method above according to the size of the building and an analysis of congestion and evacuation delay phenomena.

A new coal mine floor rating system and its application to assess the potential of floor heave

Several underground coal mines have reported floor heave associated with the buckling mechanism. However, studies on the coal mine floor, especially the buckling failure mechanism in the floor and relevant analysis techniques, are still insufficient. This paper introduces a new floor classification system, the Coal Mine Floor Rating, based on the underlying failure mechanisms. The new floor mass rating system quantifies the stability of floor strata by considering two main factors: uniaxial compressive strength and discontinuity spacing of the floor units. The Coal Mine Floor Rating is incorporated into an empirical method, the Floor Heave Index, developed by statistical analysis of a database obtained from actual floor failure cases. The floor heave and non-floor heave cases are determined depending on the Coal Mine Floor Rating and the Horizontal Stress Rating, a proxy that represents the magnitude of horizontal stress. While obtaining underground floor data is challenging, the new floor classification system and the empirical method help assess the potential of significant floor heave on development in underground coal mine roadways for new mining projects or future workings.

New classification for pulp chamber floor anatomy of human molars.

Comprehensive knowledge of the entire root canal system is an essential prerequisite for successful routine endodontic treatments. The internal and external morphological configurations of roots and root canals are quite complex. Thus, several classifications have been put forth by researchers to appropriately understand the distinguishable features of root and root canal systems of multirooted teeth. Until now, the researchers have proposed new classification systems for a thorough understanding of the root canal systems, mainly concentrating on the root canal anatomy and anomalies. Moreover, with the advent of newer digital imaging systems, these classifications are more reproducible and relevant. They have further gained importance for the clinicians by acting as an aid in decision-making based on evidence-based dentistry. However, these classifications are primarily focused on the root canals, but none have assessed the anatomy of the pulp chamber floor. This review article delineates a novel Pawar and Singh molar pulp chamber floor classification© for the anatomy of maxillary and mandibular molars to ease the clinician's skills and further increase the prevailing literature for the benefit of researchers.

Feature selection on database optimization for Wi-Fi fingerprint indoor positioning

Indoor location-based services have become very popular, principally, because of its wide and valuable applications. On that context, Wi-fi fingerprinting based on the received signal strength indicator (RSSI) has become very popular, due the fact that RSSI values are easily acquired. On the Wi-fi fingerprint method, machine learning algorithms are trained on the constructed fingerprint database and then used on a new entry to give the indoor location based on its estimations. Choosing the correct machine learning algorithm is one of the main problems in the literature. However the database sizes used during the training phase is also one of the main concerns. In this paper, a proposed feature selection method used on the original UJIIndoorLoc database created a smaller version of it, with the 30 highest RSSIs after the APIDs responsible for then in descending order, and created even smaller database subsets. Both databases, the original UJIIndoor Loc database and ours, were split into smaller subsets that were used on the classification problem according the DESIP method proposed in [1]. Six machine learning algorithms were deployed for training and testing the two database subsets with the classification attributes modified for symbolic localization. The J48 with the AdaBoost iterative algorithm gave the best results on both database subsets. The minimized database subsets showed smaller elapsed time results for all the classifications that were done. The accuracy results show similar results for both database subsets, on building and floor classification. Although, on the region attribute, the database subset with 520 attributes got better accuracy results than the reduced one.

Teaching Forest Floor Classification with Blended Learning

Teaching Forest Floor Classification with Blended Learning

Blending Multimedia and Face-to-Face Teaching to Enhance Learning about the Forest Floor

The forest floor is essential to functional, healthy forests. It is important for forestry professionals to understand, describe, and classify forest floors. We developed a Forest Floor educational resource, blending web-based multimedia and face-to-face teaching. The objectives of this study were to (1) develop blended-learning activities to teach forest floor description and classification and (2) assess student perceptions of the blended-learning method using exploratory factor analysis and group interviews. We used a Likert scale survey instrument to assess student perceptions of their learning, and investigated underlying factors through exploratory factor analysis of survey results and the manifestation of factors in focus group interviews. Five implicit factors were interpreted: (1) satisfaction with the Forest Floor resource as a learning enhancement; (2) response to presentation of concepts using a blended learning method; (3) student self-assessment of learning; (4) student learning preferences in accessing materials; and (5) website usability. Ninety-four percent of students agreed or strongly agreed that the Forest Floor resource was helpful for learning forest floor concepts, 79% that describing samples in class was essential for understanding the properties of organic horizons, and 81% that they were able to relate information in the Forest Floor resource to samples used in a face-to face activity, demonstrating that students tended to prefer learning information from videos and in collaboration with other students, and felt positive about their knowledge of the new material.

Evolution of Sonar Survey Systems for Sea Floor Studies.

Approximately 71% of our planet is covered with oceans. It is also known that oceans are the last frontiers for the mankind’s survival and therefore it becomes pertinent that they are studied in great details. It has been found that the exploration of the oceans can be done more precisely using acoustics as one of the methods, as the acoustic waves can propagate over large distances and also using a broad spectrum of frequencies various issues of the ocean studies can be addressed more effectively than many of the other methods, both in terms resolution (using high frequency components) of measuring parameters and over large ranges (using low to very low frequency components). Currently with the technological advancement and improved computing algorithms, we have state of art systems for ocean exploration, which can provide information about the sea floor, sub-surface including ocean floor classification. These could be projected in 2-D and 3-D visualization to a great accuracy. Also available are acoustical methods wherein one can obtain an extremely important information about water column properties (both in terms of bioinformation and physical properties), and has great importance as this water column is the medium for transmission of all kind of energies(acoustic for short, medium and long ranges and some time light source for exploration over a very short distance) that are used for exploration on the oceans. It will therefore be interesting to understand the progress of underwater acoustics from its very primitive stage, where acoustic transmission through water medium was used for first time to the present day highly complex but very advanced acoustic sea-floor surveying systems. It will also be interesting to know, with a very old maritime history of using seas for transportation, as to what were the methods used by early time seafarers to understand depths of the oceans they were sailing. It has taken almost a century in developing an acoustic system to arrive at the present day advancement. An attempt has been made to present a perspective of evolution and advancement in underwater acoustics and related electronic, material and computational advancement, starting from the early attempts to the modern day acoustic equipments.

Universal response spectrum procedure for predicting walking-induced floor vibration

Floor vibrations caused by people walking are an important serviceability problem both for human occupants and vibration-sensitive equipment. Present design methodologies available for prediction of vibration response due to footfall loading are complex and suffer from division between low and high frequency floors. In order to simplify the design process and to avoid the problem of floor classification, this paper presents a methodology for predicting vibration response metrics due to pedestrian footfalls for any floor type having natural frequency in the range 1–20Hz.Using a response spectrum approach, a database of 852 weight-normalised vertical ground reaction force (GRF) time histories recorded for more than 60 individuals walking on an instrumented treadmill was used to calculate response metrics. Chosen metrics were peak values of 1s peak root-mean-square (RMS) acceleration and peak envelope one-third octave velocities. These were evaluated by weight-normalising the GRFs and applying to unit-mass single degree of freedom oscillators having natural frequencies in the range 1–20Hz and damping ratios in the range 0.5–5%. Moreover, to account for effect of mode shape and duration of crossing (i.e. duration of dynamic loading), the recorded GRFs were applied for three most typical mode shapes and floor spans from 5m to 40m.The resulting peak values as functions of frequency i.e. spectra are condensed to statistical representations for chosen probability of being exceeded over a wide range of applications. RMS (acceleration) spectra show strong peaks corresponding to the first harmonic of pacing rate followed by clear minima at approximately 3.5Hz, a second much smaller peak corresponding to the second harmonic and a steady decline with increasing frequency beginning around 5Hz. One-third octave spectra show asymptotic trends with frequency, span and damping.A comprehensive validation exercise focusing on the acceleration RMS spectra was based on a representative range of floor samples for which modal properties had been identified and walking response studied during experimental campaigns of vibration serviceability evaluation. Due to the statistical approach an exact validation would not be possible, hence measured peak RMS values were matched to distributions for the equivalent idealised structure. In the vast majority of cases the measured values, intended to represent worst-case conditions, fitted the upper decile of the corresponding simulated spectra indicating consistency with the proposed approach.

Probabilistic Modeling of Walking Excitation for Building Floors

Slender floor structures are becoming increasingly prone to excessive vibration due to human-induced walking excitation. To prevent discomfort of floor occupants and/or malfunctioning of sensitive equipment, it is necessary to have a reliable means of estimating floor vibration in the design phase. For accurate estimation of the floor vibration, both reliable excitation and structural models are required. This paper concentrates on the former by evaluating the performance of the existing force models and suggesting their improvement. For this a force model adopted in the United Kingdom by the Concrete Society was applied to four nominally identical floors using their experimentally identified modal properties. After comparison with experimental data the drawbacks of the force model were identified after which an improved model of the walking-induced dynamic force, based on the combination of two existing methodologies used separately for low- and high-frequency floors, is proposed. The improved model accounts for the intersubject variability in the walking force with respect to the pacing frequency, step length, and forcing magnitude. Moreover, it includes all relevant frequency components of the walking force into analysis, removing the need for classification of floors as low or high frequency. The proposed approach should help designers and building owners to make more informed decisions when evaluating vibration serviceability of floor structures.

Classification of Human Induced Floor Vibrations

This paper summarizes results from vibration tests on floors. Classification of floors into five headings is given based on vibration characteristics. Suggested criteria and limiting values are proposed. These values are based on tests carried out on site and in laboratory conditions. It was found that the point load deflection is the best indicator for floor vibration design, when the fundamental frequency is higher than 10 Hz. It is further suggested that the deformation caused by the point load is divided into a local deflection mainly caused by soft toppings and floating floors and a global deflection of the main floor joists.

Mapping the Seabed and Habitats in National Marine Sanctuaries—Examples from the East, Gulf and West Coasts

The National Marine Sanctuary System requires seabed and habitat maps to serve as a basis for managing sanctuary resources and for conducting research. NOAA, the agency that manages the sanctuaries, and the USGS have conducted mapping projects in three sanctuaries (Stellwagen Bank NMS, Flower Garden Banks NMS, and Channel Islands NMS) with an emphasis on collaboration of geologists and biologists from the two agencies and from academic institutions. Mapping of seabed habitats is a developing field that requires the integration of geologic and biologic studies and the use of swath imaging techniques such as multibeam and sidescan sonar. Major products of swath mapping are shaded-relief topographic imagery which shows seabed features in great detail, and backscatter imagery which provides an indication of the types of materials that constitute the seabed. Sea floor images provide an excellent basis for conducting the groundtruthing studies (using video, photo, and sampling techniques) that are required to collect the data necessary for making meaningful interpretative maps of the seabed. The compilation of interpretive maps showing seabed environments and habitats also requires the development of a sea floor classification system that will be a basis for comparing, managing, and researching characteristic areas of the seabed. Seabed maps of the sanctuaries are proving useful for management and research decisions that address commercial and recreational fishing, habitat disturbance, engineering projects, tourism, and cultural resources.

Development of a roof and floor classification applicable to collieries : Buddery, P; Oldroyd, D C Proc ISRM Symposium: Eurock'92, Chester, 14–17 September 1992P197–202. Publ London: Thomas Telford, 1992

Development of a roof and floor classification applicable to collieries : Buddery, P; Oldroyd, D C Proc ISRM Symposium: Eurock'92, Chester, 14–17 September 1992P197–202. Publ London: Thomas Telford, 1992

Evaluation and verification of bottom acoustic reverberation statistics predicted by the point scattering model

The point scattering model offers a parametrization of the reverberation envelope probability density function (pdf) in terms of the average number of scatterers contributing to the return and the presence of a coherent component in the received process. Computer simulations were used to verify model predictions and to evaluate their usefulness in the context of sea floor classification. Central to the verification strategy was the successful solution of the inverse problem based on synthetic reverberation data. The average number of scatterers was determined from estimates of the kurtosis of the instantaneous reverberation pdf. The magnitude of a coherent component embedded in the scattered return was recovered from envelope histograms with the assistance of the Kolmogorov goodness‐of‐fit test. Following the verification study, the model was perturbed by introducing clustered and ordered distributions of scatterers in addition to the standard Poisson. The initial results indicate that the reverberation envelope pdf differs significantly for the three scatterer distributions. The clustered distribution led to a rapid increase in kurtosis, while the ordered distribution displayed evidence of intermittent coherent scattering. The usefulness of this parametrization was further tested with real reverberation data representing several distinct seafloor regimes. [Research funded by ONR.]

Review of the Netherlands Study of Noise Control in Multiple Dwellings

A review will be given of the most interesting results of an extensive research program carried out under the auspices of the Organization T.N.O. in the Netherlands since 1945 with the objective to evaluate the insulation of sound in dwellings. The properties of a limited number of constructions were investigated by a subjective inquiry in 1200 flats in Rotterdam and The Hague. Following the preliminary international code for measuring the airborne and impact sound insulation between rooms a rather great number of floor and wall types have been studied in a set of 48 experimental flats in Rotterdam. Some typical results will be given. A new method for estimating the amount of flanking transmission has been found and will be described. A tentative Netherlands standard for sound insulation in dwellings will be sketched in outline. It is based on a classification of walls and floors as to their “over-all” capability of insulating sound. Future research will be concentrated among other things on statistically measuring sound spectra prevailing in dwellings. An automatically recording analyzer is being built for that purpose.