Single Scenario Research Articles

Multi-Agent Collaborative Inference via DNN Decoupling: Intermediate Feature Compression and Edge Learning

Recently, deploying deep neural network (DNN) models via collaborative inference, which splits a pre-trained model into two parts and executes them on user equipment (UE) and edge server respectively, becomes attractive. However, the large intermediate feature of DNN impedes flexible decoupling, and existing approaches either focus on the single UE scenario or simply define tasks considering the required CPU cycles, but ignore the indivisibility of a single DNN layer. In this paper, we study the multi-agent collaborative inference scenario, where a single edge server coordinates the inference of multiple UEs. Our goal is to achieve fast and energy-efficient inference for all UEs. To achieve this goal, we design a lightweight autoencoder-based method to compress the large intermediate feature at first. Then we define tasks according to the inference overhead of DNNs and formulate the problem as a Markov decision process (MDP). Finally, we propose a multi-agent hybrid proximal policy optimization (MAHPPO) algorithm to solve the optimization problem with a hybrid action space. We conduct extensive experiments with different types of networks, and the results show that our method can reduce up to 56% of inference latency and save up to 72% of energy consumption.

A modified mode shape data-based method for beams structural damage detection

Damage detection methods based on changing modal parameters have received considerable attention in engineering applications due to the satisfactory results and the low associated cost compared with other techniques. The Mode Shape Data Based Indicator (MSDBI) is a damage indicator available in the literature, being used to identify damage in beam structures from the mode shape, mode shape slope and mode shape curvature, in the undamaged and damaged configurations of the element under study. However, in some situations, the configuration of the displacement mode shape of the ith mode, of the undamaged structure compared to the damaged one, presents mirroring. The damage identification algorithm could be a better indicator when these situations occur. Them, this paper presents a proposal to modify this method, called MSDBIM. The proposed modified method (MSDBIM) and the traditional method (MSDBI) were applied in two numerical examples that were elaborated in commercial software of finite elements, namely a simply supported concrete beam and a fixed-end steel beam in different single and multiple damage scenarios with sensitivity studies. A new discretization for the fixed-end beam was performed to assess whether there is a direct influence on the damage identification method. The results show that the proposed method (MSDBIM) performs better than the traditional method (MSDBI).

Optimizing performance of water-cooled photovoltaic-thermal modules: A 3D numerical approach

To evaluate and improve the efficiency of photovoltaic solar modules connected with linear pipes for water supply, a three-dimensional numerical simulation is created and simulated via commercial software (Ansys-Fluent). The optimization utilizes the principles of the 1st and 2nd laws of thermodynamics by employing the Response Surface Method (RSM). Various design parameters, including the coolant inlet velocity, tube diameter, panel dimensions, and solar radiation intensity, are systematically varied to investigate their impacts on energetic and exergitic efficiencies and destroyed exergy. The relationship between the design parameters and the system responses is validated through the development of a predictive model. Both single and multi-objective optimizations are performed using the predictive model to optimize the thermal and electrical productivity under different scenarios. The findings indicate the significance of the thermal exergy effectiveness, as evidenced by its low P-value for all solar system responses, indicating its crucial role in the predictive model. For single-objective optimization, the desirability is equal to 1 in cases where only heat transfer efficiency, whole energy effectiveness, or thermal exergy efficiency is maximized or only destroyed exergy is minimized. The improvements in energy and exergy efficiencies range from 3.55% to 69.13%, with the amount of destroyed exergy reduced by 81.47% compared to the base case. For multi-objective optimization, desirability values exceeding 0.829 and 0.655 are obtained for single and multi-objective scenarios, respectively, indicating that the expected performance is within desirable limits. The findings provide valuable insights for designing high-efficiency photovoltaic/thermal systems and addressing their challenges and limitations.

Scalability evaluation of forecasting methods applied to bicycle sharing systems

Public Bicycle Sharing Systems (BSS) have spread in many cities for the last decade. The need of analysis tools to predict the behavior or estimate balancing needs has fostered a wide set of approaches that consider many variables. Often, these approaches use a single scenario to evaluate their algorithms, and little is known about the applicability of such algorithms in BSS of different sizes. In this paper, we evaluate the performance of widely known prediction algorithms for three sized scenarios: a small system, with around 20 docking stations, a medium-sized one, with 400+ docking stations, and a large one, with more than 1500 stations. The results show that Prophet and Random Forest are the prediction algorithms with more consistent results, and that small systems often have not enough data for the algorithms to perform a solid work.

Severity of Schistosoma haematobium co-infection with malaria in school-children is potentially modulated by host CD14 gene variants

ObjectiveSchistosomiasis remains a chronic disease of global importance, especially in many rural areas of the world where co-infection with Plasmodium falciparum is common. It is critical to decipher the role of single or co-infected disease scenarios on immune system regulation in such individuals and how such co-infections can either ameliorate or complicate immune response and the consequent disease outcome. First, 10 ml of urine samples, collected between 10:00 am and 2:00 pm, was filtered for diagnosis of schistosomiasis, while egg count, indicative of disease severity, was determined by microscopy. Furthermore, genomic DNA samples extracted from dried blood spots collected on filter paper from one hundred and forty-four Schistosoma haematobium-infected school-children was tested for P. falciparum parasite positivity by an allele-specific nested-PCR analysis of merozoite surface protein (msp)-1 and -2 genes and a real-time PCR assay. In addition, among P. falciparum parasite-positive individuals, we carried out a Taqman SNP genotyping assay to extrapolate the effect of host CD14 (-159 C/T; rs2569190) genetic variants on schistosome egg count.ResultsOf the 144 individuals recruited, P. falciparum parasite positivity with msp-1 gene were 34%, 43% and 55% for MAD20, RO33 and K1 alleles respectively. Of the co-infected individuals, CD14 genetic variants ranged from 18.8% vs 21.5%, 33.3% vs 44.4%, 9.7% vs 11.8% for single versus schistosome co-infection for the wild type (CC), heterozygous (CT) and mutant (TT) variants respectively. Though the mean egg count for co-infected individuals with CD14 wild type (33.7 eggs per 10 ml of urine) and heterozygote variants (37.5 eggs per 10 ml of urine) were lower than that of schistosome infection alone (52.48 and 48.08 eggs/10 ml of urine respectively), it lacked statistical significance (p-value 0.12 and 0.29), possibly reflecting the benefit of the CD14 activation in schistosome plus malaria co-infection and not schistosome infection alone. In addition, the lower mean egg count in co-infected individuals reveal the benefit of downstream Th1 immune response mitigated by CD14 innate activation that is absent in schistosome infection alone.

Design recommendations for the rehabilitation of an urban canyon in a subtropical climate region using aerial thermography and simulation tools

Outdoor public space is the setting for everyday social interaction where activities take place in order to satisfy collective urban needs. It is important to achieve climate-friendly urban design in order to ensure acceptable urban thermal comfort, especially in over-urbanised city centres during summer periods. In this research, an urban canyon street located in the historical centre of the subtropical city of Huelva (Spain) was analysed. After a survey carried out by in situ meteorological sensors and airborne infrared sensors (equipped on an unmanned aerial vehicle, UAV), the current thermal comfort was analysed in terms of PMV and PPD at different times of a typical summer day (11:00 h, 15:00 h and 19:00 h) with the aim of formulating design recommendations to improve its performance in terms of urban comfort. Then, thermal comfort was evaluated in different scenarios where feasible mitigation strategies (replacement of materiality, addition of vegetation and sun shading elements) were applied to classify them according to their effectiveness using the ENVI-met simulation tool. The results of the current scenario showed that, due to its N-S orientation and its aspect ratio (H/W), the urban comfort depends drastically on the day hour variation. A comfortable thermal environment is achieved at all points of the urban canyon as a result of the shade generated by the buildings during the morning and afternoon. However, in the central hours of the day the feeling of thermal discomfort was alarming (PMV values of +3 and PPD values above 90%). The proposed mitigation measures showed a considerable improvement in urban thermal comfort, with the addition of vegetation being the most effective solution (with an improvement in PMV value of 42% and a reduction in PPD value of 43%). The combination of all the proposed measures in a single scenario showed encouraging results in the rehabilitation of public spaces in use.

Novel Deep Feature Fusion Framework for Multi-Scenario Violence Detection

Detecting violence in various scenarios is a difficult task that requires a high degree of generalisation. This includes fights in different environments such as schools, streets, and football stadiums. However, most current research on violence detection focuses on a single scenario, limiting its ability to generalise across multiple scenarios. To tackle this issue, this paper offers a new multi-scenario violence detection framework that operates in two environments: fighting in various locations and rugby stadiums. This framework has three main steps. Firstly, it uses transfer learning by employing three pre-trained models from the ImageNet dataset: Xception, Inception, and InceptionResNet. This approach enhances generalisation and prevents overfitting, as these models have already learned valuable features from a large and diverse dataset. Secondly, the framework combines features extracted from the three models through feature fusion, which improves feature representation and enhances performance. Lastly, the concatenation step combines the features of the first violence scenario with the second scenario to train a machine learning classifier, enabling the classifier to generalise across both scenarios. This concatenation framework is highly flexible, as it can incorporate multiple violence scenarios without requiring training from scratch with additional scenarios. The Fusion model, which incorporates feature fusion from multiple models, obtained an accuracy of 97.66% on the RLVS dataset and 92.89% on the Hockey dataset. The Concatenation model accomplished an accuracy of 97.64% on the RLVS and 92.41% on the Hockey datasets with just a single classifier. This is the first framework that allows for the classification of multiple violent scenarios within a single classifier. Furthermore, this framework is not limited to violence detection and can be adapted to different tasks.

Conflation of modal strain energy, spectral finite element, and machine learning techniques for damage detection in multi-span steel girder bridges with variable sections

A new approach is proposed to detect micro and small damages in multi-span bridges constructed with variable sections by combining modal strain energy and spectral finite element through the use of machine learning techniques. The traditional finite element methods are inefficient in detecting and estimating structures' micro-damages. Therefore, wave propagation-based methods must be applied. Micro-damages can be modeled through the spectral finite element method, widely adopted in wave propagation simulation. Among the frequency-based methods used in damage identification, the modal strain energy is more sensitive, especially in complicated structures. However, this method is difficult to apply to steel girders with variable sections and different support conditions. By combining the modal strain energy and the spectral finite element and applying the support vector regression method, the location and severity of damage in steel girders could be accurately identified to overcome the difficulties of detection in complicated elements. To improve the ability of the support vector regression method, in the damage severity estimation, it was combined with feature selection. To evaluate the proposed method's ability, first, an element of the beam- type, with the characteristics of a spectral finite element, is defined in the OpenSees software. To assess the accuracy of the defined element in detecting micro-damages, a tested numerical model was adopted to measure the accuracy of the provided element. Next, to detect small damages, the model of the steel girder bridge with the defined element was prepared, and the modal properties of the structure were extracted. Due to the change in cross-section along each girder, the modified modal strain energy-based damage index (β) equation was used in the subject model. By applying the damages of different severities in steel girders in single and multiple damage scenarios, the modified damage index β was calculated. With the damage indexes calculated in the different scenarios and applying the machine learning method, the severity of damage was estimated. The obtained results indicate that this proposed method is accurate and satisfactory for detecting and estimating micro and small damages with high accuracy in bridges with variable section girders.

Queue-Aware Resource Allocation in Full-Duplex Multi-Cellular Wireless Networks

In this paper, we aim to tackle the challenges of resource block scheduling and power allocation in the context of multi-cell full-duplex wireless networks. This is a more realistic setting than the single cell scenario, and it better envisions how full-duplex wireless communications could eventually be implemented. We propose an optimal queue-aware joint scheduling and power allocation algorithm for full-duplex wireless networks in a multi-cell scenario. Because of its mathematical intractability, we decouple the problem and solve it for scheduling first, and for power allocation second. We consider both indoor and outdoor scenarios and show that the gains of multi-cell full-duplex wireless networks, with respect to their half-duplex counter-parts, are not always prevalent. Furthermore, we highlight the importance of inter-cell cooperation when it comes to scheduling resources and show that depending on the scenario at hand, interference mitigation from inter-cell cooperation can improve the performance of user equipment in terms of throughput and waiting delay. Finally, we show that power allocation can improve user equipment throughput with its efficiency being tied to the deployment scenario at hand.

An Energy Efficient Dual IRS-Aided Outdoor-to-Indoor Communication System

Intelligent reflecting surfaces (IRSs) have emerged as a promising technology to realize energy- and spectrally efficient low-cost wireless communication systems. In this work, we propose a dual IRS-aided communication model for outdoor-to-indoor wireless communication. The scenario comprises an outdoor base station that communicates with the indoor users with the aid of two distinct IRSs. Specifically, we derived the closed-form expressions of the spectral efficiency (SE), energy efficiency (EE), and outage probability (OP) for the dual IRS-aided communication system. The results show that utilizing distributed IRSs, i.e., by dividing the large IRS into two smaller IRSs, significant performance gains in the SE and EE can be achieved over the single IRS scenario. Although the multiplicative pathloss effect impacts performance, it can be mitigated by increasing the number of reflecting elements. Initially, we evaluated the performance of a single-user scenario; later on, we generalized it to the multiuser scenario with nonorthogonal multiple access. Further, the performance is also compared with the baseline direct cellular transmission. The analytical derivations of the SE, EE, and OP are corroborated through simulations.

RIS Aided NR-U and Wi-Fi Coexistence in Single Cell and Multiple Cell Networks on Unlicensed Bands

To address the scarcity of spectrum resource incurred by explosively increasing traffic load in existing fifth generation (5G) networks, we consider exploiting Reconfigurable intelligent surface (RIS) to harmonize the coexistence of cellular networks and WiFi system on the unlicensed bands. In particular, RIS is capable of not only improving the performance of new radio on unlicensed bands (NR-U) by increasing the signal to interference and noise ratio (SINR) of cellular users, but also decreasing the interference to the existing WiFi system caused by NR-U. Thus, we investigate the performance of RIS aided coexisting networks both in single cell and multiple cell scenarios. For the single cell scenario, we aim to improve the worst transmission rate of UEs in NR-U network. The problem is solved by a proposed alternating optimization-based solution. For the multiple cell scenario, cell cooperation is exploited jointly with multiple RISs to maximize the sum-rate of all UEs. Although the problem is more challenging than that of single cell case, we propose a multi-agent reinforcement learning (MARL) algorithm to solve it. Numeral results verify the effectiveness of our proposed schemes and show that deploying RISs is a promising solution for the enhancement of NR-U networks.

Changes in design precipitation over the Nordic-Baltic region as given by convection-permitting climate simulations

The increased risk of flooding due to global warming and subsequent heavy rainfall events in the Nordic-Baltic region, call for recommendations directed at long-term planning. One example of such recommendations are climate change allowances. These are often based on expected changes in design precipitation as given by climate model simulations, and are used as a buffer on top of current design values to avoid a future increased damage potential as a consequence of climate change. We here compute expected changes in precipitation design values, so-called climate factors (CFs), for the Nordic-Baltic region, based on 3 km convection permitting simulations from the Nordic Convection Permitting Climate Projections project. These have the advantage of explicitly resolving convection, which has been shown to be the main contributor to increased rainfall. We assess the dependence of the CFs on rainfall duration, return period, and geographical location, focusing on the summer (convective) season, short durations and the high emission scenario RCP8.5. We also compare these CFs to those computed from a non-convection permitting regional climate model ensemble.We found higher CFs for the longer return period, with only few exceptions, and distinctly higher CFs going from daily to sub-daily durations. However, the different simulations give conflicting results for very short-duration rainfall (<3 h). The huge difference in the climate sensitivity of driving GCMs dominates the magnitude of estimated return levels. Our analysis is shaped by the high computational costs of running convection permitting models, resulting in a very limited ensemble (3 members) representing a single emission scenario (RCP8.5). Therefore, we believe that combining results from convection permitting simulations with a larger ensemble (21 members) of non-convection permitting simulations adds value to the assessment of robust climate change allowances for heavy precipitation in the Nordic-Baltic region.

A gradient-enhanced physics-informed neural network (gPINN) scheme for the coupled non-fickian/non-fourierian diffusion-thermoelasticity analysis: A novel gPINN structure

This paper proposes a modified artificial intelligence (AI) approach based on the gradient-enhanced physics-informed neural network (gPINN) with a novel structure for the generalized coupled non-Fickian/non-Fourierian diffusion-thermoelasticity analysis. Previous successful application of gPINN in function approximation and single partial differential equation (PDE) problems gave us enough motivation to develop its application with a novel structure for solving a system of coupled PDEs. The governing equations are formulated for a copper-made strip based upon the Lord-Shulman theory of the generalized coupled thermoelasticity. In order to shed light on the capacity of the gPINN-based approach, three examples with different boundary conditions for a one-dimensional half-space are presented, as well as an example for a two-dimensional half-space. The primary goal is to investigate the transient behavior of field variables (i.e., non-dimensional molar concentration, non-dimensional displacement, and non-dimensional temperature). Variations of dimensionless stress are also examined in depth. Performance comparisons between the gPINN-based method and commonly used approaches support the efficacy of the gPINN with a novel structure. Detailed sensitivity analyses are provided to sufficiently explore the impact of neural network hyperparameters. Moreover, the effect of relaxation time on the behavior of mass concentration is addressed in a single scenario. Overall, the proposed gPINN-based method with a novel structure yields consistently encouraging and satisfactory results, and this approach entails the capability to be applied to more complicated problems. Numerical results underline the importance of the weights assigned to the derivatives of the loss terms to achieve more accurate predictions. Nonparametric statistical tests offer further evidence for exceptional agreement between gPINN solutions with certain weights and outcomes derived by commonly used analytical techniques.

Teaching Palliative Care to Emergency Medicine Residents Using Gamified Deliberate Practice-Based Simulation: Palliative Gaming Simulation Study.

Emergency departments (EDs) care for many patients nearing the end of life with advanced serious illnesses. Simulation training offers an opportunity to teach physicians the interpersonal skills required to manage end-of-life care. We hypothesized a gaming simulation of an imminently dying patient using the LIVE. DIE. REPEAT (LDR) format, would be perceived as an effective method to teach end-of-life communication and palliative care management skills. This was a gaming simulation replicating the experience of caring for a dying patient with advanced serious illness in the ED. The scenario involved a patient with pancreatic cancer presenting with sepsis and respiratory distress, with a previously established goal of comfort care. The gaming simulation game was divided into 4 stages, and at each level, learners were tasked with completing 1 critical action. The gaming simulation was designed using the LDR serious game scheme in which learners are allowed infinite opportunities to progress through defined stages depicting a single patient scenario. If learners successfully complete the predetermined critical actions of each stage, the game is paused, and there is a debriefing to reinforce knowledge or skills before progressing to the next stage of the gaming simulation. Conversely, if learners do not achieve the critical actions, the game is over, and learners undergo debriefing before repeating the failed stage with an immediate transition into the next. We used the Simulation Effectiveness Tool-Modified survey to evaluate perceived effectiveness in teaching end-of-life management. Eighty percent (16/20) of residents completed the Simulation Effectiveness Tool-Modified survey, and nearly 100% (20/20) either strongly or somewhat agreed that the gaming simulation improved their skills and confidence at the end of life in the following dimensions: (1) better prepared to respond to changes in condition, (2) more confident in assessment skills, (3) teaching patients, (4) reporting to the health care team, (5) empowered to make clinical decisions, and (6) able to prioritize care and interventions. All residents felt the debriefing contributed to learning and provided opportunities to self-reflect. All strongly or somewhat agree that they felt better prepared to respond to changes in the patient's condition, had a better understanding of pathophysiology, were more confident on their assessment skills, and had a better understanding of the medications and therapies after the gaming simulation. A total of 88% (14/16) of them feel more empowered to make clinical decisions. After completing the gaming simulation, 88% (14/16) of residents strongly agreed that they would feel more confident communicating with a patient and prioritizing care interventions in this context. This palliative gaming simulation using the LDR format was perceived by resident physicians to improve confidence in end-of-life communication and palliative care management.

Damage identification in steel plate structures considering bending and torsion moments using estimated seismic responses

This paper proposes a model-free damage identification method and suggests an effective index for damage diagnosis in steel plate-like structures considering both the torsion and bending moments using only acceleration data, with no need for geometric characteristics. The basis for damage detection is the difference between healthy and damaged structural responses to different ground accelerations. In doing so, two Peak Ground Acceleration (PGA) scales, 0.05 (g), and 0.5 (g), labeled for the Loma Prieta ground motion, are considered. Four-node meshes are applied to reflect the torsion moments in the steel plate behavior. A multi-objective function, Fisher Information Matrix, is utilized, and four sensor locations are optimized to identify the locations most sensitive to damage. AutoRegressive Moving Average with exogenous input (ARMAX) system identification method is employed to obtain the structural displacement from the acceleration data and, subsequently, the displacement state equations of the nodes. With the aid of the structural displacement and its derivatives obtained through the central finite difference approximation method, damage to the plate structure is detected at PGA = 0.05 (g) and PGA = 0.50 (g) considering thickness changes based on the peak values manifested in the proposed damage detection index. Furthermore, the recorded structural acceleration responses at the desired scales of PGAs are fed into the 2D wavelet transform to detect damage imposed on the plate structure using the diagonal component of the wavelet transform. The results confirm that the proposed nonmodel-based approach successfully detects the single and multiple damage scenarios and can be used for a quick and judicious assessment of the steel plate behavior immediately after damage.

Experimental investigation on alternate load path mechanisms in double-span beams with unsymmetrical span lengths and reinforcement ratios

In the design and assessment of reinforced concrete (RC) structures against progressive collapse via single column removal scenarios, also known as the “alternate load path (ALP) method”, the development of secondary mechanisms in double-span beams, i.e. compressive arch action (CAA) and catenary action (CA), has been thoroughly investigated and identified in many previous studies. The majority of the experimental tests were conducted on symmetrical double-span beams (identical span length and steel reinforcement ratio for both beams). However, variations in span length and reinforcement details of beams adjacent to one another are common in RC buildings due to architectural requirements. Hence, an experimental study with a total of 6 tests was conducted to investigate the development of ALP mechanisms (CAA and CA) in double-span RC frames with unsymmetrical 1) span length and/or 2) longitudinal reinforcement ratio. The 6 tests included 3 symmetrically-designed specimens that served as the baseline to facilitate comparison and quantification of the ALP development for the other 3 unsymmetrical continuous beams. The behaviour and development of ALP mechanisms in all the 6 specimens were analysed at structural level including general observations and measurements of the external forces, displacements, and strains. Last but not least, a simplified method that was initially developed based on symmetrical test data was employed to validate unsymmetrical test results, with the aim to provide a robust and simple tool for preliminary design considerations of the ALP mechanisms.

Simulation and Prediction of Greenhouse Gas Emissions from Beef Cattle

Greenhouse gas emission is a key issue in the sustainable development of agriculture. To effectively predict the greenhouse gas emissions of beef cattle, a model is proposed based on system dynamics and greenhouse gas emission calculation methods, and a scenario is set as a ‘Straw to Beef’ project in Jilin Province. The model was built on a baseline emission scenario (feed precision: 60%, breeding environment: dry fattening farm, corn straw utilization: burning straw), with single- and comprehensive emission reduction scenarios considered, predicting trends and reduction potentials in greenhouse gas emissions from cattle breeding and straw burning in Jilin Province from 2013 to 2028, measured in CO2-eq (CO2 equivalent). The model also explored the impact of 11 controllable variables on greenhouse gas emissions. Results showed that (1) From 2013 to 2022, greenhouse gas emissions from straw burning and cattle breeding in Jilin Province increased significantly and had an annual growth rate of 6.51% in 2020. (2) Single emission reduction scenarios showed an increasing trend in greenhouse gas emissions, while comprehensive emission reduction scenarios showed a decreasing trend. Among them, the S2.2.1 scenario (feed precision: 80%, breeding environment: livestock barn manure pit, corn straw utilization: burning straw) had the strongest emission reduction ability in the single reduction scenario, the S3.2.2 scenario (feed precision: 80%, breeding environment: livestock barn manure pit, corn straw utilization: Feed-processing straw) had the strongest emission reduction ability in the comprehensive reduction scenario, reducing emissions by 5.10% and 69.24%, respectively, compared to the baseline scenario. This suggests that the comprehensive emission reduction scenarios which utilized straw resources reasonably can greatly reduce agricultural greenhouse gas emissions. (3) The optimal emission reduction scenario indicated that the higher the proportion of digestible energy in beef cattle’s total energy intake, the more perfect the fecal treatment process, and the higher the utilization rate of straw feed, the lower the greenhouse gas emissions. Therefore, to effectively reduce greenhouse gas emissions from cattle breeding and straw burning in Jilin Province, it is important to implement comprehensive emission reduction scenarios prioritizing the efficient utilization of straw resources and improving beef cattle management practices.

DOA estimation and object localization based on multi-harmonic multi-beam characteristic of 1-bit time-modulated reflectarray

A method for direction of arrival (DOA) estimation and object localization based on 1-bit time-modulated reflectarray (TMRA) is presented. The multi-harmonic multi-beam (MHMB) characteristic of 1-bit TMRA is analyzed firstly to reveal the mapping relationship between target’s angular feature and received spectrum feature. Then the operating principle and perception procedures for single and multiple targets scenarios are discussed in detail. Benefited from the mapping relationship, the proposed method shows flexibility on simultaneous multi-target detection, which is difficult to realize by conventional methods. Furthermore, the proposed method enables the environment-perception capability of existing 1-bit TMRA without increasing the system complexity, making it suitable for the realization of a low-cost multifunctional reconfigurable intelligent surface (RIS).

Reduced-Space Multistream Classification Based on Multiobjective Evolutionary Optimization

In traditional data stream mining, classification models are typically trained on labeled samples from a single source. However, in real-world scenarios, obtaining accurate labels is very hard and expensive, especially when multiple data streams are concurrently sampled from an environment or the same process. To address this issue, multistream classification is proposed, in which a data stream with biased labels (called the source stream) is leveraged to train a suitable model for prediction over another stream with unlabeled samples (called the target stream). Despite the growing research in this field, previous multistream classification methods are mostly designed for single source stream scenarios. However, various source streams contain diverse data distributions, providing more valuable information for building a more accurate model. In addition, previous works construct classification models in the original shared feature space, ignoring the effect of redundant or low-quality features on the classification performance. This may produce inefficient knowledge transfer across streams. In view of this, a reduced-space multistream classification based on multi-objective evolutionary optimization is proposed in this paper. First, a multi-objective evolutionary optimization is employed to seek the most valuable feature subset shared in the source and target domains, with the purpose of narrowing the distribution difference between source and target streams. Following that, a Gaussian Mixture Model-based weighting mechanism for source samples is presented. More especially, two drift adaptation methods are proposed to address asynchronous drift. Experimental results on benchmark datasets show that the proposed method outperforms other comparative methods on classification accuracy and G-mean.

A dataset for the visually impaired walk on the road

As a special group, visually impaired people (VIP) find it difficult to access and use visual information in the same way as sighted individuals. In recent years, benefiting from the development of computer hardware and deep learning techniques, significant progress have been made in assisting VIP with visual perception. However, most existing datasets are annotated in single scenario and lack of sufficient annotations for diversity obstacles to meet the realistic needs of VIP. To address this issue, we propose a new dataset called Walk On The Road (WOTR), which has nearly 190 K objects, with approximately 13.6 objects per image. Specially, WOTR contains 15 categories of common obstacles and 5 categories of road judging objects, including multiple scenario of walking on sidewalks, tactile pavings, crossings, and other locations. Additionally, we offer a series of baselines by training several advanced object detectors on WOTR. Furthermore, we propose a simple but effective PC-YOLO to obtain excellent detection results on WOTR and PASCAL VOC datasets. The WOTR dataset is available at https://github.com/kxzr/WOTR