Speed plays a significant role in evaluating the evolution of traffic status, and predicting speed is one fundamental tasks for intelligent transportation system. There exists large number works on forecast; however, problem long-term prediction next day still not well addressed. In this paper, we propose multiscale spatio-temporal feature learning network (MSTFLN) as model to handle challenging task elevated highways. Raw data collected from loop detectors every 5 min are transformed into...

Treating each intersection as basic agent, multi-agent reinforcement learning (MARL) methods have emerged the predominant approach for distributed adaptive traffic signal control (ATSC) in multi-intersection scenarios, such arterial coordination. MARL-based ATSC currently faces two challenges: disturbances from policies of other intersections may impair and stability agents; heterogeneous features across complicate coordination efforts. To address these challenges, this study proposes a...

There have been many attempts to design brain-computer interfaces (BCIs) for wheelchair control based on steady state visual evoked potential (SSVEP), event-related desynchronization/synchronization (ERD/ERS) during motor imagery (MI) tasks, P300 potential, and some hybrid signals. However, those BCI systems cannot implement the navigation flexibly effectively. In this paper, we propose a scheme two-class MI four-class SSVEP tasks. It only provide multi-degree its user, but also allow user...

Locating multiple sources in an unknown environment based on their signal strength is called a multi-source location problem. In recent years, there has been great interest deploying autonomous devices to solve it. A particle swarm optimizer (PSO) widely employed source method. Yet most work this field focuses flat search space while ignoring height information. An unmanned aerial vehicle (UAV) coarser but wider view as it flies higher. Inspired by such facts, paper improving the efficiency...

A vehicle's license plate is the unique feature by which to identify each individual vehicle. As an important research area of intelligent transportation system, recognition vehicle plates has been investigated for some decades. An approach based on a visual attention model and deep learning proposed handle problem Chinese car traffic videos. We first use modified locate plate, then segmented into seven blocks using projection method. Two classifiers, combine advantages convolutional neural...

A learning automaton (LA) is a powerful tool for reinforcement learning. Its action probability vector plays two roles: 1) deciding when it converges, i.e., total computing budget has used, and 2) allocating among actions to identify the optimal one. These intertwined roles lead problem: mostly goes currently estimated due its high regardless whether such allocation can help true one or not. This work proposes new class of LA that avoids use allocation. Instead we only determine if converges...

In this paper, we present a new approach to detect traffic signs based on cascaded convolutional neural networks (CNNs). First, the local binary pattern (LBP) feature detector and AdaBoost classifier are combined extract regions of interest (ROI) for coarse selection. Next, CNNs employed reduce negative samples ROI sign recognition. Compared with conventional CNN, our CNN contains three layers its classification part is replaced by support vector machine (SVM). The German detection benchmark...

Geopolymers are novel cementitious binders that finding new industrial applications every day. One of the most important aspects for commercialization these products is their behavior in plastic state. The workability fresh geopolymer paste can be measured with several common tests used Portland cement concrete, like flow and slump; however, a more in-depth characterization rheology essential to understand its basic setting mechanisms. In this article, rheological prepared under different...

The connectivity of a large-scale heterogeneous Internet Vehicles (IoV) is an important challenge for the environment urban road scenes, which feature crossroads, buildings, and communication devices. uneven density wide distribution vehicles in city, building barriers, interference other technology will make weaker. To best our knowledge, there no formal method to model analyze by considering scenes. This paper presents such theoretical investigate four properties- i.e., possibility, data...

Accurate performance evaluation of discrete event systems needs a huge number simulation replications and is thus time-consuming costly. Hence, efficiency always big concern when simulations are conducted. To drastically reduce its cost conducting them, ordinal optimization emerges. further enhance the optimization, optimal computing budget allocation (OCBA) proposed to decide best design accurately quickly. Its variants have been introduced achieve goals with distinct assumptions, such as...

Being one of the key techniques for unmanned autonomous vehicle, traffic sign recognition is applied to assist autopilot. Colors are very important clues identify signs; however, color-based methods suffer performance degradation in case light variation. Convolutional neural network, as deep learning methods, able hierarchically learn high-level features from raw input. It has been proved that convolutional network–based approaches outperform ones. At present, inputs networks processed...

Traffic flow prediction is quite crucial for estimating the future traffic states, efficient and accurate models greatly contribute to smooth of road networks. However, existing methods mainly concentrate on short-term prediction. The challenging task long-term next day, as important reference management, still not well solved. In this paper, we present a residual deconvolution based deep generative network (RDBDGN) handle problem proposed method consists generator discriminator. composed...

Deep learning models for short-term travel speed prediction on urban expressways, such as the convolutional neural network (CNN), still present several limitations in multiscale spatiotemporal feature extraction. Hence, this paper, three hybrid CNN are proposed to improve basic model with regard target aspects (i.e., 5 min) expressways. More specifically, long memory (LSTM), AutoEncoder (AE), and Inception module incorporated into capture features of data effectively accuracy robustness...

Abstract Reservoir inflow prediction plays a significant role in the field of hydrological prediction. Accurate and reliable reservoir is key to flood control decision. In this paper, we combine deep belief network (DBN) with Long Short-Term Memory (LSTM) present hybrid model based on learning (HDL) for We take full consideration basin flow rainfall factors, which significantly affect flow. According data, divide corresponding data into two cases: rain no rain. The proposed approach consists...

Ensuring the reliability of deep neural networks (DNNs) is paramount in safety-critical applications. Although introducing supplementary fault-tolerant mechanisms can augment DNNs, an efficiency tradeoff may be introduced. This study reveals inherent fault tolerance networks, where individual neurons exhibit varying degrees tolerance, by thoroughly exploring structural attributes DNNs. We thereby develop a hardware/software collaborative method that guarantees DNNs while minimizing...

Due to the influence of global climate anomalies, abnormal weather conditions such as heavy rainfall have become more frequent in recent years, posing a significant threat operation transportation systems. An effective assessment resilience system before and during rain can alert department take necessary emergency actions. However, existing methods for assessing networks mostly suffer from following problems: (1) Simulation modeling impacts lack realism; (2) After-the-fact evaluations...

Parity symmetry is an important local feature for qualitative signal analysis. It strongly related to the phase of signal. In image processing parity a cue line-like or edge-like quality structure. The analytic well-known representation 1D signals, which enables extraction spectral representations as amplitude and phase. Its domain that complex numbers. We will give overview how can be generalized monogenic in $n$D case within Clifford valued domain. approach based on Riesz transform...

Along with the Internet of Vehicles, some intelligent systems can help medical vehicles transport supplies and patients. In terms emergency issues like catastrophic natural disasters or serious accidents, safe timely transportation for is particularly important. For assistance to on road, models position prediction provide accurate information ambient in next seconds. However, increasing number road changing environment, it an important challenge predict location correctly. Current usually...

Abstract Traffic data prediction offers a significant way to evaluate the future traffic congestion status; many deep learning based approaches have been widely applied in this field. Most current methods only consider short‐term forecasting; however, long‐term prediction, which supports optimized distribution of resources, is not well studied. Besides, multiple parameters enable stronger constraints for estimation, but correlation between them both spatial and temporal domains has...