As the restaurant industry is facing labor shortage issues, use of meal delivery robots instead waiters/waitresses not only allows customers to experience impact robot technology but also benefits business financially by reducing costs. Most existing employ magnetic navigation technologies, which require strip installation and changes decor. Once moving path changed, strips need be re-laid. This study proposes multisource information fusion, i.e., fusion ultra-wide band positioning with an...

For a SLAM system operating in dynamic indoor environment, its position estimation accuracy and visual odometer stability could be reduced because the can easily affected by moving obstacles. In this paper, algorithm based on Yolov4-Tiny network is proposed. Meanwhile, feature point elimination strategy traditional ORBSLAM Besides this, to obtain semantic information, object detection carried out when points of image are extracted. addition, epipolar geometry LK optical flow method employed...

Pipeline is one of the most important tools to transport gases or liquid substances from starting point destination. There are large numbers small diameter pipelines under urban areas, which require be inspected regularly prevent their leakage. Inspection Gauge (PIG) carries pipeline defects detection sensors and localization discover corresponding positions by its spiral motion inside pipeline. The low-accuracy property Micro-Electro-Mechanical System (MEMS) based inertial main limitation...

The artificial fish swarm algorithm (AFSA) is one of the state-of-the-art intelligent techniques, which widely utilized for optimization purposes. Fiber optic gyroscope (FOG) error parameters such as scale factors, biases and misalignment errors are relatively unstable, especially with environmental disturbances aging fiber coils. These uncalibrated main reasons that precision FOG-based strapdown inertial navigation system (SINS) degraded. This research mainly on application a novel (NAFSA)...

Based on the 3D Reduced Inertial Sensor System (3D-RISS) and Machine Learning Enhanced Visual Data (MLEVD), an integrated vehicle navigation system is proposed in this paper. In demanding conditions such as outdoor satellite signal interference indoor navigation, work incorporates smooth navigation. Firstly, a landmark set up both of its size position are accurately measured. Secondly, image with information captured quickly by using machine learning. Thirdly, template matching method...

The inertial navigation system (INS) and the global positioning systems (GPS) are most popular methods of navigation. However, INS is prone to severe drift errors, especially when dealing with commercial grade systems, GPS may not always be available, such as indoors in downtown canyons. This paper proposes a vision based landmark matching improve overall solution, particularly absence GPS. Landmarks known position need set advance. As appears visual images, it detected through template...

This subject studies the vision/inertial combined system assisted by neural network, and realizes function of suppressing accumulation inertial device errors in environment Global Navigation Satellite System (GNSS) signal rejection. topic is first improved on basis VGGNet model, a more efficient model for landmark recognition obtained. The template matching method used to obtain continuous time-location information distance update visual odometry. In order reduce complexity this study uses...

This paper presents a multi-task learning architecture named SegPoseNet for 3D hand estimation problem. Previous methods usually gather image features from one perspective, which are unable to make full use of information. In this work, we dig into the part segmentation task and introduce semantic guidance information help pose task. The is novel but effective in estimation. Since predicting directly hard, design two-stage network, consists 2D-StageNet 3D-StageNet. For an RGB image, carries...