| Author: | Hao, Jiale |
| Title: | Edge learning on energy-constrained embedded systems |
| Advisors: | Zhang, Jun (EIE) |
| Degree: | M.Sc. |
| Year: | 2021 |
| Subject: | Edge computing Real-time data processing Neural networks (Computer science) Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Electronic and Information Engineering |
| Pages: | 50 pages : color illustrations |
| Language: | English |
| Abstract: | With the development of information technology, more and more videos have been created and need to be processed immediately. Cloud computing has been unable to fulfill some critical conditions. Nowadays, data will be sent to the cloud center for processing and the results will be sent back to the client. However, the process needs high bandwidth and will cause high latency, which is not enough for some cases like UAVS and self-driving. On the other hand, user data may be not safe when transmitting, so it is better to process at local for safer privacy [2]. Besides, with the increase of cloud computing users, it will result to a burden to cloud computing center. Edge computing has been proposed to solve these problems. Instead of send the data to the cloud center for computing and receive the result, edge computing processes locally and could get the results without transmission latency. This dissertation focuses on real time video detection and prediction and deploys compressed DNN on raspberry pi. In order to use DNN on resource-constrained device raspberry pi3, this dissertation use MobileNet and optimize this network to fit this device. This system can process locally without much loss of accuracy. With low latency, good accuracy and small DNN module, this real-time object prediction system can be used in UAVS, robots, avoid obstacles and so on. |
| Rights: | All rights reserved |
| Access: | restricted access |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 5658.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 2.13 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/11182