Realizing human-like perception is a challenge in open driving scenarios due to corner cases and visual occlusions. To gather knowledge of rare occluded instances, federated learning assisted connected autonomous vehicle (FLCAV) has been proposed, which leverages vehicular networks establish deep neural (DNNs) from distributed data captured by vehicles road sensors. Without the need aggregation, FLCAV helps preserve privacy while reducing communication costs compared with conventional...

Vehicle detection and localization in complex traffic scenarios pose significant challenges due to the interference of moving objects. Traditional methods often rely on outlier exclusions or semantic segmentations, which suffer from low computational efficiency accuracy. The proposed SSF-PAN can achieve functionalities LiDAR point cloud based object detection/localization SLAM (Simultaneous Localization Mapping) with high accuracy, enabling map-free navigation frameworks. novelty this work...

With the increase in coal mining depth, many rock bursts triggered by fold structures have been observed, but disaster mechanism is not clear. This research aims at investigating influence of upright structure(UFS) on burst mechanisms. The mechanical model and numerical UFS are constructed, stress evolution law working face a area studied utilizing engineering practice. shows that occurrence disasters structure has obvious regional characteristics. horizontal vertical distribution similar to...

The significant difference between recycled aggregate and natural is the content of attached mortar layer. With increase replacement rate aggregate, shrinkage creep concrete significantly increased. In this paper, 180-day tests with different water–cement ratios were designed in order to analyze effect substitution ratio on properties. results show that strain a 50% 100% at 180 days 26% 48% higher than ordinary concrete, respectively, growth group II 22% 47%, respectively. When load was old,...

The tail-sitter VTOL UAV boasts not only high-speed cruising and air hovering capabilities, but also its unique tail-sitting vertical takeoff landing attitude enable aerial operations with an exceptionally small cross-sectional area. This feature effectively broadens the scope of application for in intelligent agriculture, encompassing tasks such as agricultural inspection, production monitoring, topographic mapping. Given necessity frequent modal transitions, this paper is grounded a...

Safety is the primary concern in motion planning and decision-making of virtual driver that provides prescriptions to real human even performs self-driving absence take-over. For such an issue, traditional reinforcement learning methods, limited by their mechanisms, suffer from a slow convergence model training as well less consideration for early warning possible accidents. To address above deficiency, this paper proposes new method based on counterfactual evolutionary reasoning can be used...

State of charge (SOC) estimation lithium batteries is one the most important unresolved problems in field electric vehicles. Due to changeable working environment and numerous interference sources on vehicles, it more difficult estimate SOC batteries. Particle filter not restricted by Gaussian distribution process noise observation noise, so suitable for application estimation. Three main works are completed this paper taken LFP (lithium iron phosphate) battery as research object. Firstly,...

In order to solve the problems of high sampling rate which could shorten lifetime wireless sensor networks in wild relics security system, power spectrum reprocessing is applied feature extraction seismic wave signal. This method used target recognition problem, and three data sets are verify its effectiveness. The result shows that can reduce communication consumption prolong networks. also increase accuracy classification. method, indicating good performance, had been system Museum...

With the rapid development of urbanization, construction industry consumes a lot cement and produces large amount waste. To overcome this situation, rational use recycled aggregate produced from waste concrete is one solutions. In some countries, building has approved coarse aggregates in concrete, with limits. However, practically all existing standards regulations prohibit fine (RFA) concrete. Therefore, study on improving performance RFA vital. study, effects fly ash GGBS were...

Recently, Graph Convolutional Network (GCN) has become a novel state-of-art for Collaborative Filtering (CF) based Recommender Systems (RS). It is common practice to learn informative user and item representations by performing embedding propagation on user-item bipartite graph, then provide the users with personalized suggestions representations. Despite effectiveness, existing algorithms neglect precious interactive features between pairs in process. When predicting user's preference...

Abstract In order to reveal the pattern of oil–water movement during water flooding development in ultra-low permeability reservoir, flow units division and multiple methods such as rate-controlled mercury injection, nuclear magnetic resonance, real sandstone model, test are used perform quantitative characterization reservoir from two aspects, geological property coupling micro-mechanism. Radius size number-percentage mainstream throats moveable fluid saturation for different clusters have...

Multi-object tracking (MOT) with camera-LiDAR fusion demands accurate results of object detection, affinity computation and data association in real time. This paper presents an efficient multi-modal MOT framework online joint detection schemes robust for autonomous driving applications. The novelty this work includes: (1) development end-to-end deep neural network correlation using 2D 3D measurements; (2) a module to compute occlusion-aware appearance motion affinities space; (3)...

Navigating a nonholonomic robot in cluttered environment requires extremely accurate perception and locomotion for collision avoidance. This paper presents NeuPAN: real-time, highly-accurate, map-free, robot-agnostic, environment-invariant navigation solution. Leveraging tightly-coupled perception-locomotion framework, NeuPAN has two key innovations compared to existing approaches: 1) it directly maps raw points learned multi-frame distance space, avoiding error propagation from control; 2)...

Automatic segmentation of pulmonary arteries and veins in CT has great clinical significance. Because the growth range a single vessel is vast, have barely identical intensity values on grow very close to or even interleaved, accurate them requires intricate vascular texture information long-distance trunk as basis for artery vein classification. In order meet these two requirements simultaneously, we design residual W-Unit, which concatenated U-shaped structures. It allows network become...

In the context of Stirling polynomials, Gessel and Stanley introduced definition permutation, which has attracted extensive attention over past decades. Recently, we permutation code provided numerous equidistribution results as applications. The purpose present work is to further analyse code. First, derive an expansion formula expressing joint distribution types A B descent statistics hyperoctahedral group, also find interlacing property involving zeros its coefficient polynomials. Next,...