Traditional Kalman filter algorithms used for predicting drone positions rely solely on the historical positioning information of the drone itself, which makes it challenging to accurately estimate the position of the drone over longer time periods. In this paper, an improved IMM-KF (Interacting Multiple Models-Kalman Filter) algorithm is proposed. This algorithm incorporates a priori information about the drone’s planned flight route and the limited states of the drone, such as position, velocity, and acceleration, to achieve real-time prediction of drone position information. The advantages of interacting multiple models are also leveraged in the algorithm. Mathematical simulation results validate that this improved algorithm outperforms traditional Kalman filters in predicting drone positions, with the improvement becoming more pronounced as the prediction time horizon increases.
With the rapid development of UAV technology, the research topic of remote sensing image segmentation has gradually attracted more and more attention. Whether the image can be accurately segmented is a measure of the goodness of the algorithm. In recent years, machine learning methods have been applied to a large number of fields, and deep neural network technology has also been widely used in the field of UAV remote sensing image segmentation. This paper introduces the specific applications of various deep learning methods in remote sensing image segmentation, and briefly analyzes the development of neural networks in this problem according to the research status of several typical deep neural networks in UAV remote sensing image segmentation. Research shows that on the basis of proper adjustment of optimizer, learning rate and loss function, using deep learning method to segment UAV remote sensing images, the accuracy rate can reach more than 96%.
Hyperspectral images contain large amounts of information and have high spectral resolution. The performance of hyperspectral images in describing and distinguishing target categories has been greatly improved. With the development of unmanned aerial vehicles (UAVs), a lightweight, adaptable and low-cost way has greatly expanded the application field of hyperspectral images. This paper proposes a spatial spectrum attention mechanism based on Deep Deterministic Policy Gradient (DDPG) for hyperspectral classification. This attention mechanism is combined with 3DCNN to assign different weights to different channels in the classification process. The classification accuracy is improved by activating the useful spatial spectrum information and suppressing the useless spatial spectrum information in the hyperspectral image. A large number of experiments have been carried out to prove the effectiveness of the structure.
With the rapid development of drones, unmanned vehicles and robotics industries, VLAM has become a hot technology. In particular, the birth of 5G-powered UAV has promoted the emergence of more industrial applications, making it the most core and indispensable role in many scenarios. The loop closure detection can decrease the accumulative total of error during the process of VSLAM. Former loop closure detection methods always rely on artificially features, which are not robust, making it hard to deal with changing complex scenarios. The later deep learning-based methods are considered to be better solutions for loop closure detection. However, due to the simple network structure, there is still a lot of room for improvement. This paper proposes a more complex neural network to achieve loop closure detection. This approach adopts a fish-shaped deep neural network backbone, which can extract and fuse data features at different levels. Experiments demonstrate the feasibility and effectiveness of this backbone in loop closure detection problems.
With the increase of carrier frequency, bandwidth and antenna size of communication system, the communication system and radar system are gradually approaching. Integration of communication and sensing technology has become a research hotspot in recent years, and can be applied to many fields such as UAV low altitude airspace monitoring and traffic management. This paper provides a comprehensive overview of the latest technologies in the integration of sensing and communication. For different types of integration schemes, the classification and improvement details are given. This paper refines the development and application of communication and sensing in UAV, and provides abundant resources on methods, including the overview of advanced algorithms, systems and applications. This paper can be used as a hands-on guide for understanding, using, and developing methods. This paper comprehensively considers the difficulties in the field of unmanned aerial vehicles, summarizes the challenges faced by the integration of sensing and communication technology, and points out the development direction of this rapid development field.
In the actual operation scenario, we usually want to measure the trajectory of the UAV performing the task. This can be achieved by many different means. For example, we can obtain the motion state estimation of UAV by obtaining sensor information such as GPS, RTK and IMU. Alternatively, the velocity and acceleration of the UAV can be measured and its displacement can be calculated by integration. The device (including hardware and algorithm) to complete this motion estimation is called odometry. Considering the efficiency and cost, we hope to complete the devise of odometry visually. The target of visual odometry (VO) is to estimate the motion of the camera according to the captured image. This paper presents a method based on feature extraction and incremental pose estimation, which can accurately calculate motion state of UAVs through aerial images taken by its own.
UAV has become a promising development direction in 5G era because of its flexible deployment and economic efficiency. UAV with communication function can serve many scenes, such as traffic congestion, limited base station conditions, emergency rescue and so on. However, UAV has limited airborne energy, throughput and energy efficiency are the main bottlenecks of UAV as an air base station. Based on the consideration of various factors such as channel, user, UAV speed and transmission power, this paper constructs a reinforcement learning model for UAV energy efficiency, and puts forward the description of environment matrix to quantify the environmental parameters and participate in the action value evaluation. Firstly, based on the existing conditions, a constrained model is established to maximize the information throughput per unit energy consumption by combining historical empirical data with the exploration of a certain degree of freedom. In addition, we establish strong constraints on UAV energy to avoid unnecessary consumption as much as possible. The experimental results show that the algorithm proposed in this paper shows good performance in the simulation stage and excellent stability in the open environment.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.