OpenStreetMap (OSM) is widely used in outdoor navigation research recently, which publicly available and can provide a wide range of road information for robot navigation. In this article, aiming at the problem that map error OSM will cause global path to be inconsistent with real environment, we propose an OSM-based method combines network local perception information. As map, provides obtain by Dijkstra algorithm. Multisensor (including 3D-LiDAR Charge-coupled Device (CCD) camera) fusion...

This work presents a real-time system for multiple objects tracking in dynamic scenes. A unique characteristic of the is its ability to cope with long-duration and complete occlusion without prior knowledge about shape or motion objects. The produces good segment results at frame rate 15-20 fps image size 320 /spl times/ 240, as demonstrated by extensive experiments performed using video sequences under different conditions indoor outdoor occlusions changing background.

In complex real-world scenarios, wheel-legged robots with maneuverability, stability, and reliability have addressed growing research attention, especially in material transportation, emergency rescue, as well the exploration of unknown environments. How to achieve stable high-level movement payload delivery simultaneously is main challenge for robot. this article, a novel hierarchical framework flexible motion six robot considered experimental results. First, wheeled motion, speed consensus...

This paper presents the mechanical and control system design of latest humanoid robot platform, BHR-5, from Beijing Institute Technology. The was developed as a comprehensive platform to investigate planning for fast responsive motion under unforeseen circumstances, example, playing table-tennis. It has improvement on structure, stiffness, reliability. An open architecture based concurrent multichannel communication mode CAN bus is proposed upgrade real-time performance expansibility system....

Purpose When it comes to the high accuracy autonomous motion of mobile robot, is challenging effectively control robot follow desired trajectory and transport payload simultaneously, especially for cloud system. In this paper, a flexible tracking scheme developed via iterative learning manage distributed (BIT-6NAZA) under delivery scenarios. Design/methodology/approach Considering relationship six-wheeled independent steering in BIT-6NAZA an controller implemented reliable with...

Due to the rapid growth of Internet broadband access and proliferation modern mobile devices, various types multimedia (e.g., text, images, audios, videos) have become ubiquitously available anytime. Mobile device users usually store use contents based on their personal interests preferences. challenges such as storage limitation have, however, introduced problem overload users. To tackle this problem, researchers developed techniques that recommend for In paper, we examine importance...

Purpose This paper aims to focus on the environmental perception of unmanned platform under complex street scenes. Unmanned has a strict requirement both accuracy and inference speed. So how make trade-off between speed during extraction information becomes challenge. Design/methodology/approach In this paper, novel multi-scale depth-wise residual (MDR) module is proposed. makes full use separable convolution, dilated convolution 1-dimensional (1-D) which able extract local contextual...

Small- and medium-sized enterprises (SMEs) play a vital role in the global economy, driving innovation economic growth, despite constraints on their financial operational resources. In competitive landscape of modern markets, continuous product design improvement has become essential for SMEs to meet dynamic user requirements, enhance satisfaction, maintain competitiveness. Online reviews have emerged as valuable sources feedback, offering real-time, large-scale insights into preferences....

Lane detection and tracking is the basic component of many intelligent vehicle systems. In this study, a robust multilane method proposed. Using measurements provided by an in‐vehicle mono‐camera forward‐looking LIDAR, algorithm can address challenging scenarios in real urban driving situations. The proposed approach makes use steerable filters for lane feature detection, LIDAR‐based image drivable space segmentation marking points validations RANdom SAmple Consensus technique model fitting....

With the appearance of Global Navigation Satellite Systems spoofers, anti-spoofing has become a pressing issue. Intermediate spoofing is performed at power only slightly higher than that an authentic signal; therefore, it quite difficult to detect counterfeit signals in real time via current detection methods. Multimodal well-known method for detecting signals. However, often used acquisition stage, and thus, its effective very short. In this paper, we define searching process as when module...

Diabetic macular edema (DME) is a major cause of vision impairment in the patients with diabetes. Optical Coherence Tomography (OCT) an important ophthalmic imaging method, which can enable early detection DME. However, it difficult to achieve high-efficiency and high-precision extraction DME OCT images because sources are diverse quality not stable. Thus, still required design model improve accuracy images.A lightweight (DME-DeepLabV3+) was proposed for using DeepLabV3+ architecture. In...

In order to realise autonomous navigation of unmanned platforms in urban or off‐road environments, it is crucial study accurate, versatile and real‐time road detection methods. This proposes an adaptive method that combines lane lines obstacle boundaries, which can be applied a variety driving environments. Combining multi‐channel threshold processing, robust feature under various complex situations. Obstacle information extracted from the grid image constructed by 3D LIDAR point cloud used...

While urban areas and the corresponding traffic demand continue to grow, rural are experiencing significant growth as well. This is now resulting in congestion on facilities that previously did not have any. One area becoming a concern, particularly Florida, transitioning into more developed area. Access these usually by twolane highways, but within areas, there may be an occasional signal, possibly segments of multilane highway Currently, no analysis tool exists for analyzing two-lane with...

With the rapid development of Unmanned Aircraft System (UAS) technology, UAS has become indispensable in military and civilian field. Growth number activity required soaring demand for pilot/operator who a hot profession aviation. In this paper, classification unique characteristics operation are stated firstly. Then, considering current requirements about Pilot/Operator qualification issued by aviation authorities U.S, UK China, general special analyzed, which includes professional quality,...

In recent years, deep convolutional neural networks have made breakthrough progress in object recognition and detection tasks the field of computer vision, achieved great results both accuracy speed. However, small objects is still difficult detection, on common dataset MS COCO very low. This paper briefly reviews some work multi-scale algorithms, then proposes a method feature enhancement fusion based maps, improving COCO.

To improve the robustness of target tracking algorithms in a complex environment, this paper proposes moving detection and algorithm based on context information closed-loop learning. A region is composed its current neighboring background. For every frame that follows from video stream, long-term task principally decomposed into four parts synchronous operation: tracking, detection, integration, First, tracker obtains posterior probability location estimates state over succeeding frames by...