scholarly journals A Cost-Efficient Autonomous Air Defense System for National Security

2021 ◽  
Vol 2021 ◽  
pp. 1-10
Author(s):  
Fazle Rabby Khan ◽  
Md. Muhabullah ◽  
Roksana Islam ◽  
Mohammad Monirujjaman Khan ◽  
Mehedi Masud ◽  
...  
Keyword(s):  
Neural Network ◽  
National Security ◽  
Moving Objects ◽  
Defense System ◽  
Camera System ◽  
Human Errors ◽  
Air Defense ◽  
Detection Stage ◽  
Average Loss ◽  
Cost Efficient

In a country, air defense systems are designed to reduce threats efficiently. An air defense system is a fundamental part of any country because it provides national security. This study presents an autonomous air defense system (AADS) development that will automatically detect aerial threats (e.g., drones) and target them without any human intervention. The AADS is implemented using radar, camera, and laser gun. The radar system dynamically emits microwaves and detects moving objects around it. It triggers the camera system if it senses the frequency of any aerial threat. The camera receives the radar’s signal and detects using a neural network algorithm whether it is a threat or not. Neural network algorithms are used for the detection and classification of objects. The laser gun locks its target if the live video feed classifies an object as a more than 75% threat. In the detection stage, an average loss of 0.184961 was achieved using YOLOv3 and 0.155 using the Faster-RCNN. This system will ensure that no human errors are made while detecting threats in a region and improve national safety.

scholarly journals Real-Time Adversarial Attack Detection with Deep Image Prior Initialized as a High-Level Representation Based Blurring Network

Electronics ◽  
2020 ◽  
Vol 10 (1) ◽  
pp. 52
Author(s):  
Richard Evan Sutanto ◽  
Sukho Lee
Keyword(s):  
Neural Network ◽  
Attack Detection ◽  
Detection Methods ◽  
Defense System ◽  
Image Prior ◽  
The Neural Network ◽  
Adversarial Examples ◽  
Deep Image ◽  
Adversarial Attack ◽  
High Level

Several recent studies have shown that artificial intelligence (AI) systems can malfunction due to intentionally manipulated data coming through normal channels. Such kinds of manipulated data are called adversarial examples. Adversarial examples can pose a major threat to an AI-led society when an attacker uses them as means to attack an AI system, which is called an adversarial attack. Therefore, major IT companies such as Google are now studying ways to build AI systems which are robust against adversarial attacks by developing effective defense methods. However, one of the reasons why it is difficult to establish an effective defense system is due to the fact that it is difficult to know in advance what kind of adversarial attack method the opponent is using. Therefore, in this paper, we propose a method to detect the adversarial noise without knowledge of the kind of adversarial noise used by the attacker. For this end, we propose a blurring network that is trained only with normal images and also use it as an initial condition of the Deep Image Prior (DIP) network. This is in contrast to other neural network based detection methods, which require the use of many adversarial noisy images for the training of the neural network. Experimental results indicate the validity of the proposed method.

Spatial Calibration of a Dual PTZ-Fixed Camera System for Tracking Moving Objects in Video

2013 ◽  
Vol 57 (2) ◽  
pp. 1-10 ◽  
Author(s):  
Grzegorz Szwoch ◽  
Piotr Dalka ◽  
Andrzej Czyżewski ◽  
Susan Farnand
Keyword(s):  
Moving Objects ◽  
Camera System ◽  
Spatial Calibration

scholarly journals A New Approach in High Precision Bone Surgery: Semi-Automatic Alignment of an Application-Specific Instrument Guide Adjusted by a Smart Screwdriver

10.29007/bp2d ◽  
2018 ◽  
Author(s):  
Lukas Theisgen ◽  
Sabine Jeromin ◽  
Manuel Vossel ◽  
Sylvain Billet ◽  
Klaus Radermacher ◽  
...  
Keyword(s):  
Kinematic Chain ◽  
Lateral Direction ◽  
Human Errors ◽  
New Approach ◽  
Surgical Workflow ◽  
Bone Surgery ◽  
Handheld Device ◽  
Automatic Alignment ◽  
Cost Efficient ◽  
Application Specific

Robotic surgical systems reduce the cognitive workload of the surgeon by assisting in guidance and operational tasks. As a result, higher precision and a decreased surgery time are achieved, while human errors are minimised. However, most of robotic systems are expensive, bulky and limited to specific applications.In this paper a novel semi-automatic robotic system is evaluated that offers the high accuracies of robotic surgery while remaining small, universally applicable and easy to use. The system is composed of a universally applicable handheld device, called Smart Screwdriver (SSD) and an application specific kinematic chain serving as a tool guide. The guide mechanism is equipped with motion screws. By inserting the SSD into a screw head, the screw is identified automatically and the required number of revolutions is executed to achieve the desired pose of the tool guide.The usability of the system was evaluated according to IEC 60601-1-6 using pedicle screw implementation as an example. The achieved positioning accuracies of the drill sleeve were comparable to those of SpineAssist from Mazor Robotics Ltd., Caesarea (IL) with -0.54 ± 0.93 mm (max: 2.08 mm) in medial/lateral-direction and 0.17 ± 0.51 mm (max: 1.39 mm) in cranial/caudal-direction in the pedicle isthmus. Additionally, the system is cost-efficient, safe, easy to integrate in the surgical workflow and universally applicable to applications in which a static position in one or more DOF is to be adjusted.

Identification and Recognition of Speaker Voice Using a Neural Network-Based Algorithm

2021 ◽  
pp. 278-289
Author(s):  
Neeraja Koppula ◽  
K. Sarada ◽  
Ibrahim Patel ◽  
R. Aamani ◽  
K. Saikumar
Keyword(s):  
Neural Network ◽  
Speech Signal ◽  
Moving Objects ◽  
Back Propagation ◽  
Data Sets ◽  
Back Propagation Algorithm ◽  
Neural Organization ◽  
Propagation Algorithm ◽  
Individual Voice

This chapter explains the speech signal in moving objects depending on the recognition field by retrieving the name of individual voice speech and speaker personality. The adequacy of precisely distinguishing a speaker is centred exclusively on vocal features, as voice contact with machines is getting more pervasive in errands like phone, banking exchanges, and the change of information from discourse data sets. This audit shows the location of text-subordinate speakers, which distinguishes a solitary speaker from a known populace. The highlights are eliminated; the discourse signal is enrolled for six speakers. Extraction of the capacity is accomplished utilizing LPC coefficients, AMDF computation, and DFT. By adding certain highlights as information, the neural organization is prepared. For additional correlation, the attributes are put away in models. The qualities that should be characterized for the speakers were acquired and dissected utilizing back propagation algorithm to a format picture.

scholarly journals A Heterogeneous Spiking Neural Network for Unsupervised Learning of Spatiotemporal Patterns

2021 ◽  
Vol 14 ◽  
Author(s):  
Xueyuan She ◽  
Saurabh Dash ◽  
Daehyun Kim ◽  
Saibal Mukhopadhyay
Keyword(s):  
Neural Network ◽  
Moving Objects ◽  
Short Term Memory ◽  
Back Propagation ◽  
Spike Timing ◽  
Spatiotemporal Patterns ◽  
Spiking Neural Network ◽  
Spatial And Temporal Patterns ◽  
Unsupervised Training ◽  
Improved Performance

This paper introduces a heterogeneous spiking neural network (H-SNN) as a novel, feedforward SNN structure capable of learning complex spatiotemporal patterns with spike-timing-dependent plasticity (STDP) based unsupervised training. Within H-SNN, hierarchical spatial and temporal patterns are constructed with convolution connections and memory pathways containing spiking neurons with different dynamics. We demonstrate analytically the formation of long and short term memory in H-SNN and distinct response functions of memory pathways. In simulation, the network is tested on visual input of moving objects to simultaneously predict for object class and motion dynamics. Results show that H-SNN achieves prediction accuracy on similar or higher level than supervised deep neural networks (DNN). Compared to SNN trained with back-propagation, H-SNN effectively utilizes STDP to learn spatiotemporal patterns that have better generalizability to unknown motion and/or object classes encountered during inference. In addition, the improved performance is achieved with 6x fewer parameters than complex DNNs, showing H-SNN as an efficient approach for applications with constrained computation resources.

Discovering Gathering Pattern Using a Taxicab Service Rate Analysis Method based on Neural Network

2020 ◽  
pp. 408-428
Author(s):  
Junming Zhang ◽  
Jinglin Li
Keyword(s):  
Neural Network ◽  
Moving Objects ◽  
High Volume ◽  
Mean Value ◽  
Service Rate ◽  
Trajectory Data ◽  
Rate Analysis ◽  
The Neural Network ◽  
Time Period ◽  
Effectiveness And Efficiency

Moving objects gathering pattern represents a group events or incidents that involve congregation of moving objects, enabling the analysis of traffic system. However, how to improve the effectiveness and efficiency of the gathering pattern discovering method still remains as a challenging issue since the large number of moving objects will generate high volume of trajectory data. In order to address this issue, the authors propose a method to discovering the gathering pattern by analyzing the taxicab demand. This paper first introduces the concept of Taxicab Service Rate (TSR). In this method, they use the KS measures to test the distribution of TSR and calculate the mean value of the TSR of a certain time period. Then, the authors use a neural network based method Neural Network Gathering Discovering (NNGD) to detect the gathering pattern. The neural network is based on the knowledge of historical gathering pattern data. The authors have implemented their method with experiments based on real trajectory data. The results show the both effectiveness and efficiency of their method.

scholarly journals Real-bogus classification for the Zwicky Transient Facility using deep learning

2019 ◽  
Vol 489 (3) ◽  
pp. 3582-3590 ◽  
Author(s):  
Dmitry A Duev ◽  
Ashish Mahabal ◽  
Frank J Masci ◽  
Matthew J Graham ◽  
Ben Rusholme ◽  
...  
Keyword(s):  
Deep Learning ◽  
False Positive ◽  
Large Scale ◽  
Moving Objects ◽  
False Negative ◽  
Efficient Manner ◽  
Astronomical Surveys ◽  
Comparable Performance ◽  
Cost Efficient ◽  
Initial Results

ABSTRACT Efficient automated detection of flux-transient, re-occurring flux-variable, and moving objects is increasingly important for large-scale astronomical surveys. We present braai, a convolutional-neural-network, deep-learning real/bogus classifier designed to separate genuine astrophysical events and objects from false positive, or bogus, detections in the data of the Zwicky Transient Facility (ZTF), a new robotic time-domain survey currently in operation at the Palomar Observatory in California, USA. Braai demonstrates a state-of-the-art performance as quantified by its low false negative and false positive rates. We describe the open-source software tools used internally at Caltech to archive and access ZTF’s alerts and light curves (kowalski ), and to label the data (zwickyverse). We also report the initial results of the classifier deployment on the Edge Tensor Processing Units that show comparable performance in terms of accuracy, but in a much more (cost-) efficient manner, which has significant implications for current and future surveys.

Sign in / Sign up

Export Citation Format

Share Document