You are on page 1of 4

Summary Report on

The Mobile Computing Framework Research for Avionics


Cloud

Submitted as Assignment 2 of

Mobile Computing Course

by

Name of the Student Class Roll No.

Ruksa Gazi BED 60

Needhi Kore BED 65

Raj Jani BED 62

Under the guidance of

Mr. Tejas Hirave

DEPARTMENT OF COMPUTER ENGINEERING

SHAH AND ANCHOR KUTCHHI ENGINEERING COLLEGE

CHEMBUR, MUMBAI – 400088.

2021 – 2022
Paper Title: The Mobile Computing Framework Research for Avionics Cloud.

Published in: 2018 National Conference on Networking and Network Application.

Publication year: 2018.

Authors: Xie Jianchun, Wang Zhonghua, Li Yahui

OBJECTIVE OF PAPER

In Avionics cloud computing, different tasks are divided by logically or physically. These
tasks need to be cooperated with each other to complete certain task. Battle cloud computing
which is a cloud computing of how to share and gather different information and data from
different military devices i.e., plane, vehicle, satellite and boat etc. The traditional way of
achieving this task is done by using resources from different physical platform, sensors which
are connected by a network. Many researchers try to implement cloud computing into
military area.

The objective of this paper is to address the problems with the existing cloud computation
environment which is, as in avionics application more and more artificial intelligence devices
are being used cloud system will face more distributed cooperated task, such as radar detect,
optical image processing, electric surveillance etc.

SOLUTION DETAILS(METHODOLOGY)

This paper presents set of frameworks to accomplish above mentioned task. This framework
contains hardware and software models, with task distribution.

1. Computing Model:
This runs in cluster form; this is used to utilize the use of memory to reduce number
of I/O process and data access time. The complete computing model can be divided
into 5 stages:
a. Cluster Management and task initialization.
b. Application computing instance initialization.
c. Creating the Direct Acyclic Graph (DAG) for computing.
d. Task Allocation
e. Compute.

2. Hardware Model of the Framework:


This model includes preprocess pool, signal process pool, AI process pool and storage
pool. In these pools, the hardware includes FPGA (Field-Programmable Gate Array),
GPU, CPU, AI special processor, memory, high speed bus and network. This model
support parallel computing, iteration computing etc.
a. General pretreatment pool: Depending on hardware config and application
loading this pool can preprocess algorithms i.e., signal filtering, image
convolution etc.
b. Signal process pool: This pool support signal processing algorithms i.e., radar
detection.
c. Image/video process pool: This includes GPU process module and supports
multimedia process algorithm i.e., target detection/recognition etc.
d. General Information pool: This includes CPU and have general information
process and storage transfer ability. It supports avionics information process
algorithms i.e., mission planning, mission scheduling.
e. AI process pool: This include TPU (Tensor Processing Unit) and supports AI
algorithm i.e., neural networks, deep learning etc.
f. Storage pool: This hardware pool includes large capacity storage disk and
RAID (Redundant Arrays of Inexpensive Disks) mechanism.

3. Software Model of the Framework:


Software model is used to provide resource sharing conveniently from heterogenous
nodes or platforms in avionics cloud. According to the service, this software model
can be integrated to any layer of hardware model.
Layer Service

Application Serve Layer General ability and Mission ability

Transfer Serve Layer Data Transfer and Shared ability

Platform Serve Layer Platform Management ability

IO Serve Layer IO Device Standardization and Virtualized

Operating System Layer Operating Resource Virtualized,


Management of Virtual Machine ability

In the avionics cloud system, all the airborne physical platform, devices and weapons
should be treated as separate functional node. All these nodes are connected to
through software model and virtualized logically.

RESULT

As use of smart devices increased, using only cloud is not enough. To overcome this problem
a framework is introduced which is a distributed framework to accomplish task between
different nodes which are planes, boats etc. This framework consists of 3 models computing
model, hardware model and software model to share data and resources between nodes in
avionics.

IMPLICATION

For the distributed process demand in avionics cloud system, a distributed system is
presented in this paper. In this framework cluster pattern compute model is adopted to make
compute task management optimized, hardware pool model makes the hardware integration,
extension and upgrade more convenient. The software pool model also makes the system
resource visualize and scheduling more convenient.

You might also like