ResearchHighlightsStudent projects

Offered projects for internship students and final year graduation projects


2019-10-17

Development of a general-purpose object detection system

In this project, the student will be investigating how to perform multiple objects detection in both image and real-time video streaming using the pre-trained models. The student will also be exploring how to retrain the model to detect custom objects.

Advisor: Pakorn Ubolkosold

Description:
Expected outcome: a prototype system on Nividia Jetson platform that is flexible and easy to use for re-train the model for new different objects.
Duration: 3-4 months
Programming used: Python
Platform used: Laptop, Nvidia Jetson nano development board, camera
Some related information: https://developer.nvidia.com/embedded/jetson-nano-developer-kit

2019-10-17

Development of a warehouse localization system using visible light communications

Short summary: The student will develop and test an indoor localization system prototype using visible light communications (VLC) through LED lamps. Its potential application is in an industrial warehouse where mobile robots are deployed and location information is needed for robot guiding. Individual LED lamps transmit distinct signature signals, e.g., frequencies or codewords, that can be exploited by a VLC receiver to estimate its current location within a warehouse.

Advisor: Poompat Saengudomlert

Description:
Expected outcome: A prototype system with an LED panel and a mobile VLC receiver with a display of estimated location. In addition, some experimental results on the localization accuracy are expected.
Duration: 3-6 months
Hardware: LED lamps, photodiodes, circuit boards and components, microcontroller devices (e.g., Arduino Uno, ESP8266) Software: Arduino IDE

2019-10-17

Pyramidal fly-eye reception of Visible Light Communication (VLC) signals

Short summary: To allow for low-cost, all-round reception of VLC signals from LED lamps, three photodiodes will be placed in a 3-sided pyramidal structure that allows for angular diversity detection. In this way, the photodetectors view the VLC channel in a way similar to how a fly observes the surroundings through its facetted eye. At the start of the project, the student will receive tutoring on VLC transceiver design, as well as hands-on training on Printed Circuit Board (PCB) design using the Electronic Design Environment (EDA) RS DesignSpark PCB. The student will then develop a PCB that contains three VLC receivers, one for each photodetector. The PCB also contains three signal strength detector circuits that will be used to select the strongest signal, which will be processed by a Field Programmable Gate Array (FPGA). Since the microprocessor on an Arduino Uno board includes various Analogue-to-Digital Converters (ADCs), such board will be used as interface between the VLC reception PCB and the FPGA. The FPGA will also be used to control the Automatic Gain Control (AGC) of the receiver that captures the strongest signal. In case the student is unfamiliar with Arduino and/or FPGA programming using the Verilog Hardware Description Language (HDL), they will be tutored on these, and receive hands-on training.

Advisor: Karel Sterckx

Description:
Expected outcome: A prototype system that consists of a 3-sided pyramidal fly-eye photodetector antenna that connects to a PCB with three VLC receivers which signals are processed and controlled by an FPGA development board with an Arduino Uno interface.
Duration: 3-6 months
Number of students: 1-2
Prerequisite: Student should be familiar with Op Amp circuitry and digital electronics (logic gates, flipflops, registers and counters). Previous experience with automated PCB design, Arduino programming and/or FPGA programming is helpful though not a must.
Hardware: Components for the reception PCB, Arduino Uno Board, Intel FPGA Development Board Software: RS DesignSpark PCB, Arduino IDE (Integrated Development Environment), Verilog HDL, Intel Quartus Prime Lite Edition

2019-10-17

Extending the transmission bandwidth of Light Emitting Diodes (LEDs)

Due to their capacity of switching on and off rapidly, LEDs for general lighting can also be used to transmit communication signals, a concept that is known as Visible Light Communication (VLC). The human eye does not detect the switching and sees the average emitted light intensity. A photodetector, on the other hand, is able to capture the transmitted communication signal. The bandwidth of LED chips is about 10-20 MHz. In order to extend this bandwidth with external electronic circuitry, the equivalent circuit of LEDs need to be determined. Actually, the equivalent circuit of LEDs is known and the values of the components can be determined by means of a network analyser. However, a network analyser is expensive and the measurement method complicated. As such, a simpler procedure with less expensive measurement equipment needs to be developed. After the specifics of the LED’s equivalent circuit has been determined an electronic circuit that facilitates the measurement of LED equivalent circuitry allows for the bandwidth extension of the emitted light needs to be developed and implemented on PCB. Towards this end, hands-on training on Printed Circuit Board (PCB) design using the Electronic Design Environment (EDA) RS DesignSpark PCB will be provided. The developed circuitry will be tested with various LEDs and LED lamps.

Advisor: Karel Sterckx

Description:
Expected outcome: A low-cost method to determine the values of the components in the equivalent circuit of LEDs. A development board that allows for the low-cost measurement of the components in the equivalent circuit of LEDs, and provides for a test environment to evaluate the effectiveness of the external circuitry that extends the bandwidth of the LED under test.
Duration: 3-4 months
Number of students: 1
Prerequisite: Student should be familiar with RLC and Op Amp circuitry. Previous experience with automated PCB design is helpful though not a must.
Hardware: Components for the PCB of the development board Software: RS DesignSpark PCB

2019-10-17

Implementation of Fast Fourier Transform (FFT) and Inverse Fast Fourier Transform (IFFT) on Field Programmable Gate Array (FPGA) through employment of the CORDIC algorithm.

The CORDIC algorithm allows for the effective digital hardware implementation of trigonometry, exponential and logarithmic functions amongst other mathematical operation. Some of these mathematical functions form the core of the FFT and the IFFT, which is the reverse of the FFT and can be implemented as such. FFT and IFFT are computational effective algorithms that allow for the implementation of Fourier and Inverse Fourier Transforms. The former computes the frequency contents of a signal, whereas the latter constructs the signal from its frequency components. FFT circuitry forms the core of spectrum analysers whereas the IFFT can be used to generate arbitrary waveforms. FFT and IFFT are also used in Orthogonal Frequency Division Modulation (OFDM), which is a multicarrier modulation format that is widely used in telecom applications. As the implementation of FFT/IFFT involves a significant amount of parallel processing, it is especially suited for implementation on FPGA. At the start of the project, the student will receive tutoring, as well as hands-on training on FPGA programming using the Verilog Hardware Description Language (HDL). This will be followed by tutorials on CORDIC and FFT/IFFT. Afterwards, the student will have to implement FFT and IFFT circuitry on FPGA.

Advisor: Karel Sterckx

Description:
Expected outcome: Low-cost spectrum analyser and arbitrary waveform generator implemented on an FPGA development board.
Duration: 3-6 months
Number of students: 1
Prerequisite: Student should be familiar with digital electronics (logic gates, flipflops, registers and counters). Previous experience with FPGA programming is helpful though not a must.
Hardware: Intel FPGA Development Board
Software: Verilog HDL, Intel Quartus Prime Lite Edition

2019-10-17

Smart Interactive Museum Display

The internship is part of an on-going project aimed to develop systems and algorithms to increase interactive display for Museum. The project overall goal is to provide users interactions with hologram images representing real artefacts to study it more in depth overcoming the no touching policy applied to real ones. Currently, the hologram is displayed using 3 views, 4 views acrylic pyramid and acrylic cone. The aim of the internship is to develop algorithms to interact with the displayed hologram. There are two elements to develop for the interaction. One is to develop algorithm using Kinect to control the rotation/zoom and additional textual information of the hologram. The other interaction uses a fisheye camera positioned over the hologram display. This part of the project needs to develop image processing algorithms for the detection of the user around the display and rotate the hologram depending on the location of the user. Depending on the advance on the project, further implementation can be done. The project mixes VR and image processing

Advisor: Romuald Jolivot

Description:
Expected outcome: Algorithms using depth sensor camera to control hologram display. Algorithm for the detection of a visitor using fisheye camera to simulate real object in a hologram display.
Duration: 3-6 months
Number of students: 1
Prerequisite: Image processing, programming skills
Hardware: Kinect, Raspberry Pi, Raspberry Pi Fisheye Camera, Display using
Software: Python or C#/C++, OpenCV

2019-10-17

Development of a low-cost Multispectral Camera

RGB camera mimics the human eye vision but provides only limited spectral information. Multiple fields (medical, food, agriculture, etc.) can benefit by the analysis of narrower band of the visible spectrum. To this extent, the development of a low-cost multispectral camera is the main focus point of this research. The project runs on Raspberry Pi to control a motor that control the rotation of a diffraction grating. A white LED and a fixed aperture are used as light generating system. The rotation of the grating changes the wavelength that passes through the aperture. The goal are: to develop both the optical and mechanical system using 3D printing, to calibrate the wavelength bands and control the rotation speed to match specific waveband of the visible light. The final part is to synchronize the rotation with the camera acquisition. Further processes will be added depending on the advancement of the project with a specific applications

Advisor: Romuald Jolivot

Description:
Expected outcome: Low-cost multispectral with controlling algorithm and image processing algorithms.
Duration: 3-6 months
Number of students: 1
Prerequisite: Programming skills, basic image processing knowledge
Hardware: Raspberry Pi, Monochrome camera, while LED
Software: Python, C#, openCV

2019-10-17

3D printed optics

Testing of 3D printing of periodical structure of periods as small as 50 um in transparent raisin. This can be used for sensing applications. The work includes Learning the Digital light projection 3D printing technology and designing of small periodical structure.

Advisor: Waleed S Mohammad

Description:
Expected outcome: Fabrication of a diffraction grating, characterization of the printed structure through optical microscopy, Measuring the diffraction efficiency of the printed structure and possibly imprinting the structure in PDMS.
Duration: 3-6 months
Number of students: 1-2
Prerequisite: Basic knowledge in optics
Hardware: DPL 3D printer, optical microscope, optical detection, optical mounting and PLA 3d printer.
Software: Python

2019-10-17

Flexible grating sensors

The project aims to characterize grating structure fabricated in flexible material and test the optical response with respect to bending and pressure. The work involves learning about one dimensional gratings with periods down to 640 nm. and use the currently fabricated grating to test the diffraction efficiency of the first order.

Advisor: Waleed S Mohammad

Description:
Expected outcome: Building a setup for testing the bending and pressure effect on the grating response and design an apparatus for real life application of the actual sensor.
Duration: 3-6 months
Number of students: 1-2
Prerequisite: Basic knowledge in optics
Hardware: Optical microscope, optical detection, optical mounting and PLA 3d printer.
Software: Python

2019-10-17

Solar powered environmental monitoring system

The project aims to utilize the existing system using 6W polysilicon off-shelf solar panel with one axis tracking system for environment monitoring. That includes the implementation of different sensors and the data are sent via cellular network to be stored in a database. The work involves understanding the open loop tracking system that depends on the GPS location and calculations of the sun location at specific time on Arduino platform.

Advisor: Waleed S Mohammad

Description:
Expected outcome:
1- Writing an Arduino code to perform:Solar tracking calculations along one axis, Motor control, record all the sensors readings and send data to the server.
2- Write a client software in Python or Javascript to handle and visualize the data.
Duration: 3-6 months
Number of students: 1
Hardware: Arduino board, Narrow band communication board, in house built solar tracking system, solar panel.
Software: Arduino sketch, Javascript, Python

2019-09-06

Appliance control for smart home through visible light communications

Using visible light communications to control home appliances in smart home environments

Advisor: Poompat Saengudomlert

Description: The work includes implementing LED-based transmitters and photodiode-based receivers for transmitting commands for controlling home appliances. The work involves creating circuits and programming microcontrollers.

2018-02-15

Cloud-based Emergency Vehicle to Traffic Light Communication for Responsive Rescue Missions

Applying V2V, V2I, and Cloud computing to optimize emergency vehicle’s route and traffic light schedule for Responsive Rescue Missions.

Advisor: Chakkaphong Suthaputchakun

Description: The work includes studying Vehicle-to-Vehicle communication, Vehicle-to-Infrastructure communication, such as vehicle to traffic light communication, and cloud computing with optimization algorithms.

2018-02-15

Cloud-based Emergency Vehicle to Traffic Light Communication for Responsive Rescue Missions

Applying V2V, V2I, and Cloud computing to optimize emergency vehicle’s route and traffic light schedule for Responsive Rescue Missions.

Advisor: Chakkaphong Suthaputchakun

Description: The work includes studying Vehicle-to-Vehicle communication, Vehicle-to-Infrastructure communication, such as vehicle to traffic light communication, and cloud computing with optimization algorithms.

2017-01-09

Security and Privacy Preservation in Vehicular Networks

Development of security and privacy framework for communication between vehicles.

Advisor: Chakkaphong Suthaputchakun

Description: The ultimate objective of the project is to provide security and privacy to drivers and vehicles during communication in both Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I).

2017-01-09

Security and Privacy Preservation in Vehicular Networks

Development of security and privacy framework for communication between vehicles.

Advisor: Chakkaphong Suthaputchakun

Description: The ultimate objective of the project is to provide security and privacy to drivers and vehicles during communication in both Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I).

2016-02-26

FPGA-based MIMO Transmissions

Implementation of Alamouti code using FPGA based software defined communication systems.

Advisor: Pakorn Ubolkosold

Description: This project aims to implement the well-known Alamouti Space-Time Block Code (STBC) using BU-CROCCS software defined communication systems based on FPGAs. The work includes hardware programming (Verilog) to perform Alamouti encoding/decoding on the FPGAs.

2016-02-26

FPGA-based MIMO Transmissions

Implementation of Alamouti code using FPGA based software defined communication systems.

Advisor: Pakorn Ubolkosold

Description: This project aims to implement the well-known Alamouti Space-Time Block Code (STBC) using BU-CROCCS software defined communication systems based on FPGAs. The work includes hardware programming (Verilog) to perform Alamouti encoding/decoding on the FPGAs.

2016-02-25

Meat Quality Grading System

Development of image analysis software package that can predict the quality of the meat.

Advisor: Pakorn Ubolkosold

Description: The aim of this project is to develop an image analysis software package that can be used to grade the meat quality through images. The key component of the system is the chain of image processing algorithms including image pre-processing, features extraction, features selection, and training/learning algorithms.

2016-02-25

Meat Quality Grading System

Development of image analysis software package that can predict the quality of the meat.

Advisor: Pakorn Ubolkosold

Description: The aim of this project is to develop an image analysis software package that can be used to grade the meat quality through images. The key component of the system is the chain of image processing algorithms including image pre-processing, features extraction, features selection, and training/learning algorithms.

2016-02-20

Image Processing with Snakes and Active Contour

The implementation of Snake Active Contour with Python

Advisor: Wisarn Patchoo

Description: Snakes active contour model is image processing techniques that are widely used in several applications, for example, bio-medical applications, object tracking, security applications. It is also used as pre-processing part for high-level image and video processing, i.e., segmentation, identification and classification, object recognition, etc. Snake algorithm is a variational method that tries to minimize energy functions, derived from image and interesting features, by using external constraint forces and internal image forces. The word "snakes" is used to exhibit the behavior of contour that will pull itself toward interesting features such as edges, lines, or regions. To implement snake model for practical applications, we have to concern with realization of partial differential equations which is based on two well-known techniques: Finite Difference Method (FDM) and Finite Element Method (FEM). This project intends to implement various version of snake algorithms based on such techniques with Python program languages. It is also interested to compare how efficiency of both techniques is. Students should have some background in computer programing languages, especially Python is preferred. Duration of this project should be about 1-2 months.

2016-02-20

Image Processing with Snakes and Active Contour

The implementation of Snake Active Contour with Python

Advisor: Wisarn Patchoo

Description: Snakes active contour model is image processing techniques that are widely used in several applications, for example, bio-medical applications, object tracking, security applications. It is also used as pre-processing part for high-level image and video processing, i.e., segmentation, identification and classification, object recognition, etc. Snake algorithm is a variational method that tries to minimize energy functions, derived from image and interesting features, by using external constraint forces and internal image forces. The word "snakes" is used to exhibit the behavior of contour that will pull itself toward interesting features such as edges, lines, or regions. To implement snake model for practical applications, we have to concern with realization of partial differential equations which is based on two well-known techniques: Finite Difference Method (FDM) and Finite Element Method (FEM). This project intends to implement various version of snake algorithms based on such techniques with Python program languages. It is also interested to compare how efficiency of both techniques is. Students should have some background in computer programing languages, especially Python is preferred. Duration of this project should be about 1-2 months.

2016-02-16

Vehicular Cloud and Communications (VCloud)

Contribute to the standard for vehicular cloud architecture and platform

Advisor: Chakkaphong S.

Description: The goal of the vCloud project is to develop a vehicular cloud architecture and platform which enables resource limited vehicles to have seamless access to data sharing in temporary vehicular cloud and have access to large pool of data service in the permanent cloud for enhanced road safety and passenger comfort.

2016-02-16

Mobile Ad Hoc Network (MANET) and Wireless Sensor Network (WSN)

Developement of WSN based Application

Advisor: Chakkaphong S.

Description: The goal of the project is to apply wireless sensor network and mobile ad hoc network concepts to solve any problems such as agriculture, environment, and traffic monitoring. Thus, this topic also includes vehicular ad hoc network (VANET) and communications such as vehicle-to-vehicle (V2V) and vehicle-to-infrastructure communications (V2I).

2016-02-16

Vehicular Cloud and Communications (VCloud)

Contribute to the standard for vehicular cloud architecture and platform

Advisor: Chakkaphong S.

Description: The goal of the vCloud project is to develop a vehicular cloud architecture and platform which enables resource limited vehicles to have seamless access to data sharing in temporary vehicular cloud and have access to large pool of data service in the permanent cloud for enhanced road safety and passenger comfort.

2016-02-16

Mobile Ad Hoc Network (MANET) and Wireless Sensor Network (WSN)

Developement of WSN based Application

Advisor: Chakkaphong S.

Description: The goal of the project is to apply wireless sensor network and mobile ad hoc network concepts to solve any problems such as agriculture, environment, and traffic monitoring. Thus, this topic also includes vehicular ad hoc network (VANET) and communications such as vehicle-to-vehicle (V2V) and vehicle-to-infrastructure communications (V2I).

Development of a general-purpose object detection system

Development of a warehouse localization system using visible light communications

Pyramidal fly-eye reception of Visible Light Communication (VLC) signals

Extending the transmission bandwidth of Light Emitting Diodes (LEDs)

Implementation of Fast Fourier Transform (FFT) and Inverse Fast Fourier Transform (IFFT) on Field Programmable Gate Array (FPGA) through employment of the CORDIC algorithm.

Smart Interactive Museum Display

Development of a low-cost Multispectral Camera

3D printed optics

Flexible grating sensors

Solar powered environmental monitoring system

Appliance control for smart home through visible light communications

Cloud-based Emergency Vehicle to Traffic Light Communication for Responsive Rescue Missions

Cloud-based Emergency Vehicle to Traffic Light Communication for Responsive Rescue Missions

Security and Privacy Preservation in Vehicular Networks

Security and Privacy Preservation in Vehicular Networks

FPGA-based MIMO Transmissions

FPGA-based MIMO Transmissions

Meat Quality Grading System

Meat Quality Grading System

Image Processing with Snakes and Active Contour

Image Processing with Snakes and Active Contour

Vehicular Cloud and Communications (VCloud)

Mobile Ad Hoc Network (MANET) and Wireless Sensor Network (WSN)

Vehicular Cloud and Communications (VCloud)

Mobile Ad Hoc Network (MANET) and Wireless Sensor Network (WSN)