A single unit (head) is the conventional input feature extractor in deep learning architectures trained on multivariate time series signals. The importance of the fixed-dimensional vector representation generated by t...
详细信息
Nonlinear control has led to several great achievements for attitude control in virtual simulation environments, where the system model can be simplified to meet the application condition of those traditional nonlinea...
详细信息
ISBN:
(数字)9781728176871
ISBN:
(纸本)9781728176888
Nonlinear control has led to several great achievements for attitude control in virtual simulation environments, where the system model can be simplified to meet the application condition of those traditional nonlinear control algorithms. However, only a few of those algorithms can be applied to a practical Hypersonic Aircraft system, which suffers from uncertainties, multi-constraints, and various noises. To solve those problems effectively, in this paper, a novel reinforcement learning framework based on PI2 is proposed, wherein three kernel techniques are introduced. Firstly, a generalized path- integral-control approach is proposed to obtain the numerical solution of a stochastic dynamic model, wherein the calculation of the gradient and matrix inversion is avoided to ensure fast and reliable training convergence. Secondly, an action smooth method is used to strengthen the ability of the controller to resist noise. Thirdly, a novel RL algorithm combined with soft constrained techniques is illustrated to address the constrained nonlinear control problem. Several experiments are carried out and analyzed to demonstrate the outstanding performance of the CS-PI2.
To solve the problem of multiple Unmanned Aerial Vehicles (UAVs) with intermittent communication topologies simultaneously attack a stationary target, a time cooperative guidance law is studied in this paper. First, a...
详细信息
As the application of unmanned forklifts becomes more and more widespread, logistics scenarios are constantly evolving. This paper focuses on a scenario where the pick-up of pallets perpendicular to forklift in narrow...
As the application of unmanned forklifts becomes more and more widespread, logistics scenarios are constantly evolving. This paper focuses on a scenario where the pick-up of pallets perpendicular to forklift in narrow aisles is required. The narrow aisles only allow the forklift to complete the turning, performing pallet localization in front of it will be difficult. We proposes a method for localizing pallets in narrow aisles for unmanned forklifts. By installing the camera at an oblique angle above the pallet, the collected images are uploaded to the image processing system. The YOLOv5 is used to detect the pallet and extract the bounding boxes of the pallet class. Based on the bounding box coordinates, the original image is cropped. Convert the cropped image to grayscale for Canny edge detection, and then perform Hough line detection to extract the right edge of the pallet. This right edge will be compared with the reference pallet edge to obtain the relative localization between the pallet and the unmanned forklift.
Regional cooperative search is an important scenario for Unmanned Aerial Vehicle (UAV) fleet. It is difficult for the UAV fleet to adapt to the dynamic environment by using traditional search method, which needs pre-p...
详细信息
ISBN:
(数字)9781728180250
ISBN:
(纸本)9781728180267
Regional cooperative search is an important scenario for Unmanned Aerial Vehicle (UAV) fleet. It is difficult for the UAV fleet to adapt to the dynamic environment by using traditional search method, which needs pre-planning. In this paper, a reinforcement learning approach is employed to train the UAV agents to search without pre-determined strategies. By designing reward function, state and action spaces, agents can learn to make decisions all by themselves based on their own observations and cooperate with each other effectively. The proposed method is experimentally validated by numerical simulations, and the result show that the UAV fleet can complete task with a pretty high coverage and few repetitions at the same location.
To solve the simulation problem of uav-borne bi-static synthetic aperture radar (SAR) for sea surface targets, a scattering signatures modeling method of sea-surface target for bi-static imaging radar is proposed, bas...
详细信息
ISBN:
(数字)9781728180250
ISBN:
(纸本)9781728180267
To solve the simulation problem of uav-borne bi-static synthetic aperture radar (SAR) for sea surface targets, a scattering signatures modeling method of sea-surface target for bi-static imaging radar is proposed, based on the coupling scattering mechanism between the ship and sea surface. A bistatic scattering signature signal model for sea-surface target in conjunction with high frequency asymptotic techniques for electromagnetic (EM) scattering calculation, the multi-path EM scattering model, the time-evolving complex reflection coefficient model and the ship motion dynamics. The time varying scattering signatures for a typical ship on time evolving sea-surface are simulated and analyzed.
This article deals with model- and data-based consensus control of heterogenous leader-following multi-agentsystems (MASs) under an event-triggering transmission scheme. A dynamic periodic transmission protocol is de...
详细信息
With the widespread application of Automated Guided Vehicles (AGVs) in warehousing and logistics systems, the optimization of multi-AGV path planning has become a critical issue. Current methods primarily focus on min...
详细信息
ISBN:
(数字)9798331507992
ISBN:
(纸本)9798331508005
With the widespread application of Automated Guided Vehicles (AGVs) in warehousing and logistics systems, the optimization of multi-AGV path planning has become a critical issue. Current methods primarily focus on minimizing operating time and energy consumption but often overlook spatiotemporal distances between paths, leading to frequent path conflicts and interferences in the system. This paper proposes a neural network based time window estimation model, designed to generate time windows that better align with dynamic environments. Based on this model, we introduce a collaborative multi-AGV path planning method that optimizes spatiotemporal distances between paths. Compared to the Conflict-Based Search (CBS) method, this approach significantly improves the spatiotemporal distance between AGVs during task execution, enhancing the robustness of AGV clusters while only marginally increasing time costs.
In this paper, an autonomous positioning method based on coarse-to-fine multi-modal image matching is proposed for UAV navigation in GPS denied environment. Coarse image matching refers to roughly determining the appr...
详细信息
ISBN:
(数字)9781728190044
ISBN:
(纸本)9781728190051
In this paper, an autonomous positioning method based on coarse-to-fine multi-modal image matching is proposed for UAV navigation in GPS denied environment. Coarse image matching refers to roughly determining the approximate position of the real-time image in the reference image. Fine image matching refers to iterative matching around the position obtained by coarse image matching to achieve more accurate matching results. Besides, RANSAC is adopted to eliminate the outliers of image matching. After getting the result of image matching, we use PNP to estimate the position of the aircraft. The proposed position method has the advantages of high accuracy and reliability. The experimental results show that the average error of the position of the proposed method is 59%.
In order to solve the problem that the point feature tracking is not robust enough to reduce the accuracy of the system in a low-texture environment, this paper proposes a visual inertial odometry system based on poin...
详细信息
ISBN:
(数字)9781728186351
ISBN:
(纸本)9781728186368
In order to solve the problem that the point feature tracking is not robust enough to reduce the accuracy of the system in a low-texture environment, this paper proposes a visual inertial odometry system based on point-line features (PL-VINS). Firstly, the system improves the extraction and matching algorithm of line features to increase the speed and accuracy of the algorithm; then the line features are added to the front-end tracking and back-end optimization. Finally, the EuRoc public data set is used as the experimental object to prove the effectiveness of the algorithm.
暂无评论