In this paper, we propose a novel localization methodology to enhance the accuracy from two aspects, i.e., adapting uncertain noise of microelectromechanical system-based inertial navigation system (MEMS-INS) and accurately predicting INS errors. First, an interacting multiple model (IMM)-based sequential two-stage Kalman filter is proposed fuse information MEMS-INS, global positioning (GPS), in-vehicle sensors. Three bias filters are built with different covariance matrices cover wide range...

How to achieve reliable and accurate positioning performance using low-cost sensors is one of the main challenges for land vehicles. This paper proposes a novel fusion strategy vehicles in GPS-denied environments, which enhances simultaneously from sensor methodology levels. It integrates multiple complementary not only incorporating GPS microelectromechanical-based inertial measurement unit, but also "virtual" sensor, i.e., sliding-mode observer (SMO). The SMO first synthesized based on...

In large-scale long-term dynamic environments, high-frequency objects inevitably lead to significant changes in the appearance of scene at same location different times, which is catastrophic for place recognition (PR). Therefore, how eliminate influence achieve robust PR has universal practical value mobile robots and autonomous vehicles. To this end, we suggest a novel semantically consistent LiDAR method based on chained cascade network, called SC_LPR, mainly consists semantic image...

In this paper, we propose a cost-effective localization solution for land vehicles, which can simultaneously adapt to the uncertain noise of inertial sensors and bridge Global Positioning System (GPS) outages. First, three Unscented Kalman filters (UKFs) with different covariances are introduced into framework Interacting Multiple Model (IMM) algorithm form proposed IMM-based UKF, termed as IMM-UKF. The IMM provide soft switching among UKFs therefore characteristics. Further, two IMM-UKFs...

As an effective means of solving collision problems caused by the limited perspective on board, cooperative roadside system is gaining popularity. To improve vehicle detection abilities in such online safety systems, this paper, we propose a novel multi-sensor multi-level enhanced convolutional network model, called architecture (MME-YOLO), with consideration hybrid realistic scene scales, illumination, and occlusion. MME-YOLO consists two tightly coupled structures, i.e., inference head...

Abstract In this paper, a novel multi-information fusion methodology is proposed for vehicle pose estimation. The purpose to improve the estimation accuracy during global navigation satellite system (GNSS) outages, mainly from two aspects: (1) extra observations of accurate velocities and angular rates without cumulative errors; (2) adaptive intelligent fusion. Firstly, multidimensional motion perception network (MMPN) designed in order estimate rates. At same time, state also output. inputs...

In this paper, we propose a reliable hybrid positioning methodology by combining the advantages of H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> filter and extreme learning machine (ELM), which addresses GPS outages uncertain nonlinear drift MEMS INS simultaneously. A novel parallel-dual-H filtering (PDHF) mechanism is proposed to prevent from diverging during make full use supplementary observations. The PDHF composed an enhanced...

The dynamic interaction among Connected and Automated Vehicles (CAVs) is becoming increasingly complex, encompassing factors such as topology the states of multiple CAVs. Existing cooperative control methods struggle to explicitly represent interaction, which can lead dangerous behavior, severe congestion, even accidents. In this paper, we propose a methodology that aims improve safety efficiency in merging zone by deeply representing Our proposed methodology, named GMA-DRL, utilizes spatial...

Many intelligent transportation system applications require accurate, reliable, and continuous vehicle position information whether in open-sky environments or Global Positioning System (GPS) denied environments. However, there remains a challenging task for land vehicles to achieve such positioning performance using low-cost sensors, especially microelectromechanical (MEMS) sensors. In this paper, novel cost-effective fusion methodology bridge GPS outages is proposed applied the Inertial...

Place recognition plays a crucial role in simultaneous localization and mapping. Unfortunately, however, changes viewpoints conditions large-scale environments impose tricky challenges for PR. To this end, article specifically proposes an explicit points-of-interest driven PR method, which consists of road segmentation module based on grid-wise patch U-transformer regions interest siamese transformer NetVLAD (RI_STV). Especially RI_STV, the individual dimension, it is dedicated to exploring...

The safety and efficiency of the merging zone is particularly important for traffic networks. Although autonomous vehicle improves from view, controlling in mostly focus on improving roadside view. Lacking detailed driving recommendation, it ignores where commercial pose a high collision risk real traffic. This paper proposes decision-making methodology to simultaneously improve zone. We have built two modules, namely assessment decision-making. Assessment module takes advantage Bayesian...

As a significant issue in computer vision, pedestrian detection has achieved certain achievements with the support of deep learning. However, congested scenes still encounters challenging problem feature loss and obfuscation. To address issue, we propose network based on correlation-and-correction fusion attention mechanism. First, multimask correction module is proposed, which generates visible part masks pedestrians, enhancing region’s features correcting false one. Besides, preserves...

LiDAR-based semantic segmentation, particularly for unstructured environments, plays a crucial role in environment perception and driving decisions unmanned ground vehicles. Unfortunately, chaotic especially the high-proportion drivable areas large-area static obstacles therein, inevitably suffer from problem of blurred class edges. Existing published works are prone to inaccurate edge segmentation have difficulties dealing with above challenge. To this end, paper proposes real-time...

The cooperative positioning method based on wireless sensors has become a popular solution for the intelligent vehicle in recent years, especially GNSS-denied environments. However, error due to Non-line-of-sight (NLOS) propagation is main problem which affects accuracy noticeably, thereby affecting vehicle's functions. In order circumvent aforementioned problem, this paper proposes multi-sensor fusion methodology achieve reliable position NLOS Initially, identification module consistency of...

Driving decision-making determines the safety and rationality of autonomous commercial vehicles. Aiming at issue safe driving decision-making, herein, a rear anti-collision methodology based on deep reinforcement learning (RAD-DRL) was creatively proposed. Firstly, aiming dynamic coupling factors for driving, decision network an actor-critic framework proposed sensor data processing. Then, inspired by idea multi-objective optimization, refined reward function is developed. It comprehensively...

Loop closure detection is the crucial issue of simultaneous localization and mapping in field autonomous driving robotics. In outdoor large-scale complex environments, existing LiDAR-based methods still inevitably suffer from viewpoint, condition changes, perceptual aliasing. To effectively fill aforementioned drawbacks, this article, a novel multimodule cascaded Siamese convolutional neural networks developed, named MMCS-Net, which simulates human-eye mechanism to extract more...

It has great significance to acquire vehicle speed for active safety system. This paper presents a methodology identifying by obtaining sparse optical flow from image sequences. Distinct corners can be detected Harris corner detector after enhancement. Then, Lucas-Kanade method calculation is utilized match the feature set of one frame on consecutive frame. In order improve accuracy flow, RANSAC algorithm introduced optimize matched corners. Finally, determined averaging all speeds estimated...

Reliable estimation of vehicle lateral position plays an essential role in enhancing the safety autonomous vehicles. However, it remains a challenging problem due to frequently occurred road occlusion and unreliability employed reference objects (e.g., lane markings, curbs, etc.). Most existing works can only solve part problem, resulting unsatisfactory performance. This paper proposes novel deep inference network (DINet) estimate position, which adequately address challenges. DINet...

In complex environments with long-term changes such as light, seasonal, and viewpoint changes, robust, accurate, high-frequency global positioning based on light detection ranging (LiDAR) map is still a challenge, which crucial for autonomous vehicles or robots. To this end, novel observation model that relies the siamese multitask convolutional neural networks (CNNs) multimodule cascaded creatively presented in article. particular, new pseudoimage representing LiDAR submap designed to...