01.06.2017 Views

UWE Bristol Engineering showcase 2015

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

This project was a<br />

development of an algorithm<br />

to help control pilots of UAVs<br />

to perform a safe forced<br />

landing for any emergency for<br />

UAVs.<br />

For this project a smart<br />

mobile phone will be used to<br />

sense the data from the<br />

sensors integrated on the<br />

phone i.e. accelerometer,<br />

gyroscope and GPS. The<br />

mobile phone will be<br />

integrated on the UAV.<br />

Salim LOUNIS<br />

MEng Aerospace System <strong>Engineering</strong><br />

An additional function of the developed<br />

system is the ability to utilise the<br />

smartphones sensors for telemetry data.<br />

This allows the application to be used as<br />

a backup telemetry system should it be<br />

required.<br />

The advantage of using Google Maps is<br />

to eliminate the poor image quality due<br />

to the increase of the frame rate when<br />

the UAV fly are low altitude<br />

Forced Landing Algorithm Development And<br />

Implantation On An On-Board Android-Based Smartphone Device<br />

This Algorithm was developed<br />

on Simulink using Computer<br />

Vision Toolbox to analyse<br />

Google Maps images and<br />

display safe landing Sites.<br />

Algorithm Improvement<br />

Processing Time (s)<br />

1.8<br />

1.6<br />

1.4<br />

1.2<br />

1<br />

0.8<br />

0.6<br />

Application Development Architecture<br />

Resolution Vs Processing Time for Lower Quality Images<br />

Size [10 10]<br />

Size [40 40]<br />

size [70 70]<br />

Size [100 100]<br />

Size [130 130]<br />

Size[160 160]<br />

Size [190 190]<br />

Project Supervisor<br />

Dr Pritesh Narayan<br />

Project summary<br />

The main aim of this project is to develop a forced landing<br />

algorithm and implement it on an on-board android-based<br />

smart phone that allows the UAV to safe land at the<br />

closest landing sit at any time during the flight.<br />

Performance of the algorithm outcome was tested which<br />

concluded that hardware specification will dictates the<br />

image resolution of algorithm<br />

Project Objectives<br />

• To investigate how OpenCV library can be<br />

implemented on an Android operating system<br />

• To explore how to use edge-detection<br />

methodologies on an on-board smartphone<br />

• To examine and analyse the performance of the<br />

algorithm using real-time data<br />

• To investigate how to deploy and debug the<br />

algorithm on an Android operating device and<br />

other open source devices, such as Raspberry Pi2,<br />

and compare the performance of the devices and<br />

the algorithm<br />

Project Conclusion<br />

The aims of this project were to successfully<br />

develop an algorithm, which helps command<br />

pilots of UAVs to safely deal with forced<br />

landings in emergency situations, by using an<br />

android operating system. This project was a<br />

continuation of BEng project ,Improvement<br />

were successfully conducted to operate the<br />

algorithm in Near-Real-Time system. This<br />

algorithm was tested using a MacBook Pro(i7)<br />

to analyse it performance.<br />

This algorithm also been debugged on<br />

Android OS and Raspberry Pi2<br />

0.4<br />

Size[220 220]<br />

0.2<br />

0<br />

0 2 4 6 8 10 12<br />

Number of frames<br />

Raspberry Pi2 Development<br />

Processing time Vs. Image Resolution

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!