With the recent advancements in drone technology, there has been an increase in the development of human detection and tracking techniques for various applications, especially near borders. In this research, we propos...
With the recent advancements in drone technology, there has been an increase in the development of human detection and tracking techniques for various applications, especially near borders. In this research, we propose methods to enhance people detection performance in diverse outdoor scenarios. Our dataset design includes a wide range of lighting and color changes, different target distances, angles, and postures. The experimental data consists of images taken in various environmental situations, such as changing the drone’s flight height and capturing pictures in intensive light. To evaluate the performance of our proposed method, we enhanced the generic YOLOv5 model using the gathered data, and calculated key performance indicators, including loss functions, recall, accuracy, and mAP50. We compared the performance of our enhanced model against the standard YOLOv5 model and its versions on the same testing set.
This paper presents a signal processing framework for automatic anxiety level classification in a virtual reality exposure therapy system. Two types of biophysical data (heart rate and electrodermal activity) were rec...
详细信息
This article presents the problem of designing a nonlinear observer for an active magnetic suspension system. The design process of the nonlinear Luenberger observer (also known as the Kazantzis-Kravaris-Luenberger ob...
This article presents the problem of designing a nonlinear observer for an active magnetic suspension system. The design process of the nonlinear Luenberger observer (also known as the Kazantzis-Kravaris-Luenberger observer) is discussed. Particular attention was paid to the main nonlinearity of the system - the electromagnetic force, which was modeled applying the function describing the change in inductance as a function of the distance of the levitating object from the electromagnet surface. Theoretical analyses were confirmed by the results of experimental studies in which the task of moving the sphere between the given positions using current control was carried out. control tasks were conducted in the real-time regime on an embedded platform. The measured signals and estimated velocity were analyzed in the context of future implementations in control applications.
In a world in a continuous and rapid change, it is absolutely necessary for our students to keep up with the rapid progress of new technologies: Internet of Things (IoT), Robotics, Artificial Intelligence (AI), Virtua...
详细信息
Photovoltaic (PV) panel modelling and control is very important in renewable energy systems. Due to it variability, PV panel generation power should be maximized for the given climate conditions. This paper considers ...
详细信息
The paper’s purpose was to investigate some methods based on neural networks for the detection and classification of harmful insects for agriculture as the Halyomorpha Halys. The implementation of different object de...
详细信息
ISBN:
(数字)9781665406734
ISBN:
(纸本)9781665406741
The paper’s purpose was to investigate some methods based on neural networks for the detection and classification of harmful insects for agriculture as the Halyomorpha Halys. The implementation of different object detection networks for image categorization was analyzed. Images from the Maryland Biodiversity database were used for neural network training and testing. Rotation, scaling, blurring, mirroring, and other techniques were employed for data augmentation. For the detection and classification of Halyomorpha Halys, some neural networks that include multiple smaller networks were implemented and investigated. The networks used are the following: YOLOv5s, SSD with different backbones such as MobileNet V1, MobileNet V2, and ResNet-50, Faster R-CNN with ResNet-50 backbone, and EfficientDet-D0. Moreover, neural networks were evaluated and compared based on performance metrics such as accuracy and time. Performances like accuracy between 0.49 – 0.86 and time between 36 ms – 55 ms were obtained. The best results were obtained for YOLOv5s, in terms of accuracy, and EfficientDet-D0, in terms of time.
E-business offers to economic agents the opportunity to increase the efficiency of their activity, promote and offer their products and services on a large and diversified market. Often a single service, separately, d...
详细信息
Cyber-Physical Systems represent digital systems based mainly on the interconnection between the physical world and cyberspace. Physical processes are controlled and monitored through networked computers. Feedback loo...
Cyber-Physical Systems represent digital systems based mainly on the interconnection between the physical world and cyberspace. Physical processes are controlled and monitored through networked computers. Feedback loops between the physical and computing environment are involved, including sensors and actuators. Thus, the CPS design involves the analysis of the interrelated dynamical physical processes, computation resources, networks, and software. CPS is application-oriented, by its nature so the necessity of available services composition appears frequently. The CPS services, resources and capabilities can be placed at the system and node level. In this context, the Communication Cost is an important performance criteria for accessing their resources, even in a composite form. In this paper an interoperability framework, for Cyber-Physical Systems capabilities, using REST and SOAP web services is presented. The interoperability capabilities of the framework are ensured by both of the convergence of IP and non-IP networks for services access and by the services composition facilities. For a scenario of multiuser access at the composed services, the interoperability framework response time is evaluated, because it brings an important contribution in the determination of Communications Cost, in a QoS approach.
In this paper, a method of fusing distance information and camera data is presented, the objective being that of combining LIDAR measurements with the output of a segmentation neural network, applied on the merged ima...
In this paper, a method of fusing distance information and camera data is presented, the objective being that of combining LIDAR measurements with the output of a segmentation neural network, applied on the merged image from three frontal cameras, for navigating in an indoor environment. Usually, these types of environments are not structured, being diverse in their composition: for example, some buildings have more glass walls and doors, others might have large corridors and concrete structures. The proposed fusing method is used for reducing the errors generated by both techniques individually, which is then applied to control the trajectory of an electric wheelchair, as well as to gather a dataset of corrected and annotated images for reducing the dependency on the LIDAR data in the future.
The Internet of Things (IoT) is transforming industries by enhancing productivity and efficiency;however, energy availability remains a significant challenge due to the limited capacity of batteries and supercapacitor...
详细信息
暂无评论