Showing papers by "Universite de technologie de Belfort-Montbeliard published in 2020"

Online learning for 3D LiDAR-based human detection: experimental analysis of point cloud clustering and classification methods

Abstract: This paper presents a system for online learning of human classifiers by mobile service robots using 3D LiDAR sensors, and its experimental evaluation in a large indoor public space. The learning framework requires a minimal set of labelled samples (e.g. one or several samples) to initialise a classifier. The classifier is then retrained iteratively during operation of the robot. New training samples are generated automatically using multi-target tracking and a pair of “experts” to estimate false negatives and false positives. Both classification and tracking utilise an efficient real-time clustering algorithm for segmentation of 3D point cloud data. We also introduce a new feature to improve human classification in sparse, long-range point clouds. We provide an extensive evaluation of our the framework using a 3D LiDAR dataset of people moving in a large indoor public space, which is made available to the research community. The experiments demonstrate the influence of the system components and improved classification of humans compared to the state-of-the-art.

Distributed power generation via gasification of biomass and municipal solid waste: A review

Abstract: The access to electricity has increased worldwide, growing from 60 million additional consumers per year in 2000–2012 to 100 million per year in 2012–2016. Despite this growth, approximately 675 million people will still lack access to electricity in 2030, indicating that electricity demand will continue to increase. Unfortunately, traditional large fossil power technologies based on coal, oil and natural gas lead to a major concern in tackling worldwide carbon dioxide emissions, and nuclear power remains unpopular due to public safety concerns. Distributed power generation utilizing CO2-neutral sources, such as gasification of biomass and municipal solid wastes (MSW), can play an important role in meeting the world energy demand in a sustainable way. This review focuses on the recent technology developments on seven power generation technologies (i.e. internal combustion engine, gas turbine, micro gas turbine, steam turbine, Stirling engine, organic rankine cycle generator, and fuel cell) suitable for distributed power applications with capability of independent operation using syngas derived from gasification of biomass and MSW. Technology selection guidelines is discussed based on criteria, including hardware modification required, size inflexibility, sensitivity to syngas contaminants, operational uncertainty, efficiency, lifetime, fast ramp up/down capability, controls and capital cost. Major challenges facing further development and commercialization of these power generation technologies are discussed.

Influence of post-heat treatments on the mechanical properties of CX stainless steel fabricated by selective laser melting

Abstract: High-performance CX stainless steel was successfully manufactured using selective laser melting (SLM) technology, and different types of post-heat treatments were adopted for ameliorating the mechanical properties of as-built specimens. The microstructure evolution process (i.e., cell structures, cellular dendritic grains and blocky grains containing substructures) was explained using the rapid solidification theory after SLM. Nanoprecipitates and their hardening behavior in the SLM CX stainless steels in the as-built and solution-aged state were detected by transmission electron microscope (TEM). The results of high-resolution TEM showed that the massive needle-like nanoprecipitates with a size range of 3–25 nm (as-built sample) and 7–30 nm (solution-aged sample) were evenly distributed in the martensite matrix. In the meantime, the strengthening mechanism was analyzed and discussed. Moreover, various post-heat treatments exhibited a great influence upon the mechanical performances of the SLM CX stainless steel samples. The average micro-hardness of the SLM CX stainless steel parts was found to extremely improve from 357 HV0.2 (as-built sample) to 514 HV0.2 (solution-aged sample). On the contrary, the total impact energy (Wt) of the SLM CX stainless steel parts decreased from 83.8 J in the as-built condition to 5.3 J in the solution-aged condition.

Fusing Transformed Deep and Shallow features (FTDS) for image-based facial expression recognition

Abstract: In this paper, we propose combining between the transformed hand-crafted and deep features using PCA to recognize the six-basic facial expressions from static images. To evaluate our approach, we use three popular databases (CK+, CASIA and MMI). We introduce the use of the Pyramid Multi Level (PML) face representation for facial expression recognition. The hand-crafted features are obtained with such representations. Initially, we determine the optimal level of the PML features of three hand-crafted descriptors (HOG, LPQ and BSIF) using CK+, CASIA and MMI databases. After the optimal level of the PML is found for each descriptor, we combine them together with the transformed final VGG-face layers (FC6 and FC7) in order to get a compact image descriptor. In within-database experiments, our approach achieved higher accuracy than the state-of-art methods on both the CK+ and CASIA databases, and competitive result on the MMI database. Likewise, our approach outperformed the static methods in all six experiments of cross-databases. (C) 2020 Published by Elsevier Ltd.

Traffic Signs Detection and Classification for European Urban Environments

Abstract: Traffic signs play an important role for Advanced Driver Assistance Systems (ADAS) as well as for autonomous driving vehicles. Most of the works done focus on recognizing symbol based signs leaving apart important information provided by other type of signs like complementary panels or text based signs. In this paper, we include detection and classification of both symbol and text based signs focusing on the most common ones found in European urban environments. The system consists of three stages, traffic sign detection, refinement and classification. The detection and refinement is performed using Mask R-CNN while the classification is achieved with a proposed Convolutional Neural Network (CNN) architecture. We introduced the extended version of the German Traffic Sign Detection Benchmark (GTSDB), labeled in a pixel manner (masks) with 164 classes grouped into 8 categories. It is used for the detection and classification steps. Experimental results on German environments show that our proposed system is capable of detecting all categories of traffic signs while at the same time recognizing them with high accuracy achieving comparable performance with the state of the art.

Quantifying the Impact of Solar Photovoltaic and Energy Storage Assets on the Performance of a Residential Energy Aggregator

Abstract: Demand response (DR) and renewable energy sources have opened new avenues for end-users to lower their energy expenses via energy management systems. Aggregators facilitate the participation of end-users by acting on their behalf and interacting with bulk electricity markets. In this paper, an energy management algorithm is presented to investigate the impact of distributed photovoltaic (PV) and central energy storage system (ESS) assets on the economic performance of an energy aggregator in the residential sector. To enable DR, the aggregator provides a competitive incentive price to end-users, and centrally optimizes the central ESS assets and schedule of committed customer elastic loads. Thus, customers reduce their daily electricity bill while the aggregator decreases the aggregated peak consumption and earns profits as a return for providing DR services. The scope of this paper pertains to the economic impact of distributed PV and central ESS assets on aggregator profits and customer savings resulting from DR, including ESS degradation. Simulation results showed that the central ESS increases the income of the aggregator, whereas residential PV reduces the impact of DR.

Using Digital and Physical Simulation to Focus on Human Factors and Ergonomics in Aviation Maintainability.

Abstract: Objective:This research aimed to evaluate the differences in the assessments made by three simulation tools used in a maintainability design office to perform human factor/ergonomics (HFE) analysis...

EU Long-term Dataset with Multiple Sensors for Autonomous Driving

Abstract: The field of autonomous driving has grown tremendously over the past few years, along with the rapid progress in sensor technology. One of the major purposes of using sensors is to provide environment perception for vehicle understanding, learning and reasoning, and ultimately interacting with the environment. In this paper, we first introduce a multisensor platform allowing vehicle to perceive its surroundings and locate itself in a more efficient and accurate way. The platform integrates eleven heterogeneous sensors including various cameras and lidars, a radar, an IMU (Inertial Measurement Unit), and a GPS-RTK (Global Positioning System / Real-Time Kinematic), while exploits a ROS (Robot Operating System) based software to process the sensory data. Then, we present a new dataset (https://epan-utbm.github.io/utbm_robocar_dataset/) for autonomous driving captured many new research challenges (e.g. highly dynamic environment), and especially for long-term autonomy (e.g. creating and maintaining maps), collected with our instrumented vehicle, publicly available to the community.

A comprehensive comparative study of handcrafted methods for face recognition LBP-like and non LBP operators

Abstract: Pattern recognition and computer vision fields experienced the proposal of several architectures and approaches to deal with the demands of real world applications including face recognition. They have almost the same structure, based generally on a series of steps where the main ones are feature extraction and classification. The literature works interessted in face recognition problems, insist on the role of texture description as one of the key elements in face analysis, since it greatly affects recognition accuracy. Therefore, texture feature extraction has gained much attention and became a long-standing research topic thanks to its abilities to efficiently understand the face recognition process, especially in terms of face description. Recently, several literature researches in face application proposed new architectures based on pattern description proved by their discriminative power when extracting the feature information from facial images. These advantages combined with an outstanding performance in many classification applications, allowed the LBP-like descriptors to be one of the most prominent texture description method. Given this period of remarkable evolution, this research work includes a comprehensive analytical study of the face recognition performance of 64 LBP-like and 3 non-LBP texture descriptors recently proposed in the literature. To this end, we adopted a face recognition framework composed of four stages: 1) image pre-processing using gamma correction; 2) feature extraction using texture descriptors; 3) histogram calculation and 4) face recognition and classification based on the simple parameter-free Nearest Neighbors classifier (NN). The conducted comprehensive evaluations and experiments on the challenging and widely used benchmarks ORL, YALE, Extended YALE B and FERET databases presenting different challenges, indicate that a number of evaluated texture descriptors, which are tested for the first time on face recognition task, achieve better or competitive compared to several recent systems reported in face recognition literature.

Real-Time Simulation of Power Electronic Systems Based on Predictive Behavior

Abstract: The field-programmable gate array (FPGA) based hardware-in-the-loop (HiL) test, which minimizes the time-step of the real-time simulation below 500 ns, is an enabling technology for the development of the control unit of high-power electronic systems (HPE). In order to improve the time performance of FPGA-based HiL, this article proposes a novel parallel structure using the predictive behavior of the power electronic system. With this structure, we design an improved system-level solver applied to HPE. A piecewise insulated-gate bipolar transistor (IGBT) model is used to determine the state of the switch, offering a quasi-realistic model of the power converter. A parallel integration method is also implemented to solve the status of the circuit elements. Moreover, the proposed parallel structural can execute both the IGBT model and circuit element model at the same time, thus, reducing the simulation time-step significantly. The numerical accuracy of the solution, the architecture design, and the issue of the parallel computation are discussed in detail. An ac–dc–ac topology is presented as a case study. At last, a 25 ns time step in the National Instruments FlexRIO platform is achieved. Results comparison with the reference model is also identified and discussed.

Manufacturing feature recognition with a 2D convolutional neural network

Abstract: Feature recognition is critical to connect CAX tools in automation via the extract of significant geometric information from CAD models. However, to extract meaningful geometric information is not easy. There are still a couple of problems, such as lack of robustness, inability to learn, limited feature types, difficult to deal with interacting features, etc. To fix these problems, a new feature recognition method based on 2D convolutional neural networks (CNNs) is proposed in this paper. Firstly, a novel feature representation scheme based on heat kernel signature is developed. Then, the feature recognition problem is transferred into a graph learning problem by using a percentage similarity clustering and node embedding technique. After that, CNN models for feature recognition are trained via the use of a large dataset of manufacturing feature models. The dataset includes ten different types of isolated features and fifteen pairs of interacting features. Finally, a set of tests for method validation are conducted. The experimental results indicate that the proposed approach not only performs well on recognizing isolated features, but also is effective in handling interacting features. The state-of-the-art performance of interacting features recognition has been improved.

Detection of the key steps during Liquid Resin Infusion manufacturing of a polymer-matrix composite using an in-situ piezoelectric sensor

Abstract: This article investigates the interest of a novel promising approach, dealing with the use of an in-situ piezoelectric (PZT) disk to monitor the whole manufacturing process of a glass fiber/polyester Polymer-Matrix Composite (PMC) plate using the Liquid Resin Infusion (LRI) technique. The real-time in-situ Process Monitoring (PM) is conducted using the electrical signature (capacitance) variation of the embedded PZT transducer, which has never been done so far in the literature for such a purpose. In order to understand the capacitance response, an internal/external multi-instrumentation (Infrared Thermography, thermocouples, Acoustic Emission, Z-displacement and pressure sensing devices) was set on the infusion systems, so that it was possible to make multi-physical couplings between the various obtained measurements and the PZT capacitance curves. It was shown that the PZT capacitance is sensitive to all several steps of the infusion process, especially the gelation and vitrification phases, and can also have Structural Health Monitoring (SHM) applications as it makes the resulting composite part “smart”.

LaNoising: A Data-driven Approach for 903nm ToF LiDAR Performance Modeling under Fog

Abstract: As a critical sensor for high-level autonomous vehicles, LiDAR’s limitations in adverse weather (e.g. rain, fog, snow, etc.) impede the deployment of self-driving cars in all weather conditions. In this paper, we model the performance of a popular 903nm ToF LiDAR under various fog conditions based on a LiDAR dataset collected in a well-controlled artificial fog chamber. Specifically, a two-stage data-driven method, called LaNoising (la for laser), is proposed for generating LiDAR measurements under fog conditions. In the first stage, the Gaussian Process Regression (GPR) model is established to predict whether a laser can successfully output a true detection range or not, given certain fog visibility values. If not, then in the second stage, the Mixture Density Network (MDN) is used to provide a probability prediction of the noisy measurement range. The performance of the proposed method has been quantitatively and qualitatively evaluated. Experimental results show that our approach can provide a promising description of 903nm ToF LiDAR performance under fog.

Self-Motion-Assisted Tensor Completion Method for Background Initialization in Complex Video Sequences

Abstract: The background Initialization (BI) problem has attracted the attention of researchers in different image/video processing fields. Recently, a tensor-based technique called spatiotemporal slice-based singular value decomposition (SS-SVD) has been proposed for background initialization. SS-SVD applies the SVD on the tensor slices and estimates the background from low-rank information. Despite its efficiency in background initialization, the performance of SS-SVD requires further improvement in the case of complex sequences with challenges such as stationary foreground objects (SFOs), illumination changes, low frame-rate, and clutter. In this paper, a self-motion-assisted tensor completion method is proposed to overcome the limitations of SS-SVD in complex video sequences and enhance the visual appearance of the initialized background. With the proposed method, the motion information, extracted from the sparse portion of the tensor slices, is incorporated with the low-rank information of SS-SVD to eliminate existing artifacts in the initiated background. Efficient blending schemes between the low-rank (background) and sparse (foreground) information of the tensor slices is developed for scenarios such as SFO removal, lighting variation processing, low frame-rate processing, crowdedness estimation, and best frame selection. The performance of the proposed method on video sequences with complex scenarios is compared with the top-ranked state-of-the-art techniques in the field of background initialization. The results not only validate the improved performance over the majority of the tested challenges but also demonstrate the capability of the proposed method to initialize the background in less computational time.

Large-Signal Stable Nonlinear Control of DC/DC Power Converter with Online Estimation of Uncertainties

Abstract: The Passivity-Based Control (PBC) is recognized as an effective energy shaping approach to guarantee the asymptotic stability of the whole system by using the passivity property. However, the model-based and sensor-based characteristics limit its development and application. The combination of the PBC and online estimation technique can solve these problems. The purpose of this paper is to propose a controller and an observer, which are designed simultaneously based on Hamiltonian framework and Lyapunov criterion. It leads to the system design without separation of the dynamics of the controller and the observer. The uncertainties in the model and parameters are considered as equivalent voltage and current sources. To reduce the number of sensors, input voltage, output current, and equivalent sources are estimated together. The steady-state error is eliminated by using this estimation technique. The exponential stability of the whole system (converter, controller, and observer) is proved by using a proper Lyapunov function. Simulation and experimental results from a 3 kW 270–350 V DC/DC boost converter with a Constant Power Load (CPL) are performed to confirm the proposed control algorithm. Since the system parameter values may vary with temperature and the equilibrium point, the robustness of the proposed method is verified without and with parameters uncertainties.