wearable antennas for patient health monitoring

Project Title

wearable antennas for patient health monitoring

Project Area of Specialization Artificial IntelligenceProject Summary

The title is'' Malicious Drone Detection for Public Safety Application ''. the title ''wearable antennas for patient health monitoring" was written unintentionally.

For the identification of malicious drones, we have used Descriptors and Deep Neural Network for their classification. The classification process includes the training of these networks with practical worst-case scenarios and updating the system model with the modified data-set after the experimentation process.

Project Objectives

Malicious Drone Detection for Public Safety Application 

This study aims to figure out what are the reasons
in the increased number of drones and the problems and threats directed to a common man’s privacy with the lack of restriction of the regulation of air space in the cities.

This study will shed light upon the complications caused by the lack of control and supervision of the UAVs and the potential harms that society might face in the case scenario of malicious drones roaming the streets.

There will be a brief discussion about the latest technologies being used in the market today to eliminate these
threats and the countermeasures will be taken to stop and
identify malicious and corrupted drones.

The limitations and shortcomings of these technologies will also be shared with the reader.

The main cause of the paper is to discuss the potential
the threat of a malicious drone and the detection system in order to identify the UAVs efficiently.

Project Implementation Method

The main rationale is to detect the Amateur Drone using
Descriptors and Deep neural networks. Several descriptors
such as GLCM, LBP, CJLBP, Letraset, LTP, and NRLBP were
used in the experimentation to see which descriptor perform
better. The accuracy of each descriptor was tested through
Polynomial, Linear, BPF Kernels and the output was shown
in the form of Confusion Matrix and percentage classification
table. But it has been deduced that the results were not
satisfactory while implementing the descriptors.
In the next phase of experimentation the Deep Neural
Networks were implemented to enhance the accuracy of the
classification system. Alexie, Resnet-50, VGG-19, and Inceptionv3
were implemented to see which DNN performed better.
The accuracy of each Deep Neural Networks was shown in the
form of Confusion Matrix and Classification Percentile Table
for the Polynomial, Linear, BPF Kernels. After the careful
scrutiny of the results, it was observed that Descriptor is not
performing as reliable to be used in a practical situation, on the other hand, Deep Neural Networks performed in a superior
manner in challenging environments. Resnet-50 Neural Network
provided reliable results so it was implemented in the
proposed Classification Model. The classification process is as
follows:
When the drone is in the restricted area, the picture of the
drone or UAV is taken from a security camera or taken
from a video frame.
Its features are extracted through using Resnet-50 as they
are similar to the training data that was fed to the DNN.
Compare its feature from that of the already extracted
training data using Resnet-50. This is done through SVM
Classifier. SVM Classifier Label the picture as 1 or 0. On
the basis of this label, it is determined whether the
drone is present or not.
Now to increase the accuracy of the model we have used
the Localization Technique. Suppose the picture taken from
the camera, contain many objects such as bird, drone, etc. Now
first it will detect the objects that are present and then, will
predict whether the object is a drone or not. This can be done
using the Localization Technique.

Benefits of the Project

This project will provide the following benefits to society:

i) Public Safety of Common Man from Stray Drones

ii) Detection of Possible Armed Drones in Public or Private Section  

iii) Prevention of Invading Privacy of Individuals 

Technical Details of Final Deliverable

Trained Deep Learning Model (Software based)

Camera setup plus Micro-controller Module Connected with Alarm System (Hardware-based)

Final Deliverable of the Project HW/SW integrated systemType of Industry Telecommunication Technologies Artificial Intelligence(AI)Sustainable Development Goals Sustainable Cities and CommunitiesRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	55300
Drone	Equipment	1	20000	20000
Raspberry Pi	Equipment	1	10000	10000
Camera	Equipment	1	20000	20000
USB 32GB	Miscellaneous	1	1800	1800
micro SD card	Miscellaneous	1	2000	2000
A4 page	Miscellaneous	250	2	500
Pen	Miscellaneous	25	20	500
DVD's	Miscellaneous	10	50	500