Recently, inferring lane change intention has received considerable attention. Due to the high nonlinearity and complexity of traffic contexts, traditional methods cannot satisfy requirements long-term prediction tasks lack ability capturing nonlinear temporal dependencies. This paper proposes an inference model based on Recurrent Neural Networks (RNN), tackle time series problems. Considering dynamic interaction among surrounding vehicles, our takes sequence motion information vehicles as...

We study the problem of response selection for multi-turn conversation in retrieval-based chatbots. The task involves matching a candidate with context, challenges which include how to recognize important parts and model relationships among utterances context. Existing methods may lose information contexts as we can interpret them unified framework are transformed fixed-length vectors without any interaction responses before matching. This motivates us propose new that sufficiently carry at...

In order to decide a safe and reliable trajectory for autonomous driving vehicles, the threat of surrounding vehicles need be assessed quantitatively consider potential risk. This paper proposes novel integrated assessment algorithm decision-making system. First, motion vehicle is predicted probabilistic based on interact multiple model (IMM) threat. Then, we build an function assess in each state objectively, which synthesizes existing time-to-collision (TTC), time-headway (TH), original...

In order to improve the real-time and computational efficiency of autonomous vehicles' decision-making process, this paper draws on behavior human drivers with motivation as core proposes a planning method based risk assessment. On one hand, it analyzes determines motivations that cause driving state change for planning. other basis lateral trajectory prediction surrounding vehicles, longitudinal propensity different is added construct assessment model can reflect future time domain. Based...

On-ramp merging scenario has a great impact on traffic efficiency and fuel economy. At present, most research on-ramp focuses the optimization of sequence in single main lane scenario, which fails to give full play capacity multi-lane roads. To overcome this problem, an efficient strategy (ORMS) is proposed coordinate vehicle traffic. First, we built model unevenness flow between lanes. Based model, established selection by reinforcement learning for coordination vehicles Before enter zone,...

The demand for accurate estimation of aboveground biomass (AGB) at high spatial resolution is increasing in grassland-related research and management, especially those regions with complex topography fragmented landscapes, where grass shrub are interspersed. In this study, based on 519 field AGB observations, integrating Synthetic Aperture Radar (SAR; Sentinel-1) high-resolution (Sentinel-2) remote sensing images, environmental topographical data, we estimated the mountain grassland...

To make autonomous vehicles consider driver's personalized characteristics, this paper proposes an integrated model and learning combined (IMLC) algorithm to realize human-like driving. It includes the driving behavior modeling ensure basic safety characteristic further imitate human style. Firstly, is built according operation logics, including lane advantage assessment, target selection acceleration determination. The assessed by five features, like safety, efficiency, cooperativity, etc....

Radar human motion recognition methods based on deep learning models has been a heated spot of remote sensing in recent years, yet the existing are mostly radial-oriented. In practical application, test data could be multi-aspect and sample number each very limited, causing model overfitting reduced accuracy. This paper proposed channel-DN4, few-shot method. First, local descriptors introduced for precise classification metric. Moreover, episodic training strategy was adopted to reduce...

With the help of specific emitter identification (SEI), control efficiency satellite communication systems can be effectively improved by discriminating individual satellite. In recent years, deep learning has been introduced into SEI to enhance performance because its powerful classification capability. However, classical real-valued neural networks exhibit some limitations in extracting radio frequency fingerprint (RFF) features from complex signals, limiting improvement accuracy. Thus, we...

The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts Interest none was reported.

This work proposes an active collision avoidance between autonomous driving vehicle and pedestrian with motion uncertainty under urban road. A candidate trajectory planning method considering spatial time sequences is proposed, which combines the polynomial path velocity variable safety velocity. Then, a pedestrian-vehicle interaction model constructed, takes pedestrian's uncertain as superposition of Markov process without interference caused by vehicle, predicts probabilistically. On these...

In order to improve the efficiency and comfort of autonomous vehicles while ensuring safety, decision algorithm needs interact with human drivers, infer most probable behavior then makes advantageous decision. This paper proposes a Nash-Q learning based motion consider interaction. First, local trajectory surrounding vehicle is predicted by kinematic constraints, which can reflect short-term trend. Then, future action space built that consists five basis actions. With that, process be...

To tackle the inherent unknown deformation (e.g., translation, rotation and scaling) of inverse synthetic aperture radar (ISAR) images, a deep polar transformer-circular convolutional neural network, i.e., PT-CCNN, is proposed to achieve robust ISAR image automatic target recognition (ATR) in this article. Inspired by human visual system canonical coordinate Lie-groups, we adopt transformer module transform images log-polar representations, before which conventional network (CNN) utilized...

In the switching process of driving authority for man-machine cooperative driving, it should be ensured that driver can take over vehicle safely, stably and efficiently. This paper proposes a allocation strategy based on real-time domain (RAD), which mainly includes establishment RAD dynamic optimization. The is comprehensive internal, whose range determined by corresponding values driver's cognitive state, muscle state environment state. optimization to search optimal in at any time, its...

Comparing algorithms are crucial for enhancing the accuracy of remote sensing estimations forest biomass in regions with high heterogeneity. Herein, Sentinel 2A, 1A, Landsat 8 OLI, and Digital Elevation Model (DEM) were selected as data sources. A total 12 algorithms, including 7 types learners, utilized estimating aboveground (AGB) Pinus yunnanensis forest. The results showed that: (1) optimal algorithm (Extreme Gradient Boosting, XGBoost) was meta-model (referred to XGBoost-stacking)...

In order to solve the accuracy problem of future motion prediction surrounding vehicles with different types drivers, this paper proposes a comprehensive lateral method that combines driver intention and vehicle behavior recognition. For optimization models are established: personal optimal model system model. Then, driver’s probability is obtained through game theory. Different from traditional theory, we apply for on premise getting type instead using Nash equilibrium all players....

Lane changes require dynamic decision-making and rapid behavior planning, which are challenging for traffic modeling. We propose a two-dimensional following lane-changing framework (2DF-LC) that exploits the benefits of car-following (CF) models computational efficiency, collision avoidance, human-like behavior. This uses sigmoid-based intelligent driver model (SIDM) with both longitudinal lateral following. To avoid excessive acceleration at start-up, we develop an SIDM ensures smooth...

Soil environmental quality related to the residents’ life, health, and safety, has been hotspot issues in science of ecological environment protection. Evaluating distribution characteristics, risk, source heavy metals farmland is important for protecting soil resources. The agricultural area Lianhua town, Gongcheng County, Guilin a typical karst landform. In response problem metal pollution complex sources this area, characteristics profiles from farmland, abandoned land, forest were...

It is crucial to develop a comprehensive method for estimating the aboveground biomass (AGB) of trees, shrubs, grasslands, and sparse tree areas in ecologically fragile dry, hot valley regions with vertical zonation. Multi-source remote-sensing data can fulfill this requirement, providing help monitoring health ecosystems basis regional biodiversity conservation restoration. Sentinel-2A satellite imagery was used classify forests, grasslands Yuanmou County, Chuxiong Yi Autonomous Prefecture,...

The on-ramp area usually produces congestion, high energy consumption, and emission. In order to improve the efficiency of multi-lane heterogeneous traffic composed vehicles with different dynamic characteristics, we propose an eco-friendly merging strategy for connected automated in traffic. Firstly, optimal lane decision method is proposed via conditional proximal policy optimization algorithm optimize flow each avoid local congestion Then, considering travel time, emission, problem...

Manganese (Mn), an essential trace element for plants in which it is involved redox reactions as a cofactor many enzymes, represents important factor environmental contamination. Excess Mn can lead to toxicity conditions natural and agricultural sites. one of the most severe growth limiting factors acid soil, accounts 21% total arable lands China. The more significant part Mn-toxicity its interactions with other mineral elements, particular phosphorus (P), calcium (Ca) iron (Fe). application...

The prediction and estimation of the lane-changing state host car surrounding cars are important parts an advanced driving assistant system, which mainly depend on understanding driver behavior. To learn maneuver well, this article provides a novel stochastic model based improved input-output hidden Markov (IOHMM) framework. First, IOHMM is proposed to address deficiency that traditional cannot remember previous data describe continuous output. Then, framework, established considering...