Proposed Projects - 2015

Document created by Stefan Alexandru Malinici Employee on Apr 7, 2015Last modified by Stefan Alexandru Malinici Employee on Feb 7, 2016
Version 29Show Document
  • View in full screen mode

 

Internet radio – Adrian Stoica

This project aims to create an user space application to search and connect to internet radio servers, displays all the radio posts found (ideal to touchscreen display), menu for selection. The connection to internet must be done wireless.wire. The output must be done via Bluetooth/wire.

 

Palm Recognition – Radu Brasoveanu

 

The project scope is to build an application that should run in embedded Linux environment and should implement an algorithm that is able to distinguish individuals by palm scanning.

Scanning should be done using an simple web cam, modified or not to be able to use IR light or not.

Software should control acquisition of images in different light exposures and implement the algorithm to detect the similarities with an known individual pattern.

 

Quick references:

http://www3.ntu.edu.sg/home/adamskong/publication/A_Survey_of_Palmprint%20Recognition_Journal_10.pdf

http://www.researchgate.net/profile/Prabir_Bhattacharya/publication/221146566_Palm_Vein_Extraction_and_Matching_for_Personal_Authentication/links/09e4150b0f4ae3ef85000000.pdf

http://research.ijcaonline.org/volume50/number4/pxc3880827.pdf

http://www.sersc.org/journals/JSE/vol6_no1_2009/6.pdf

 

Debug agent via – WIFI – Marius Grigoras

 

This project purpose is to have a debug agent controlled via WIFI. Linux application for defining the agent and controlling the WIFI will be needed.

More or less complex debug scenarios including trace and performance analysis can be created as well.

 

Server cloud – WIFI – Marius Grigoras

 

This project aims at having a server or multiple servers available over WIFI for cloud based applications.

 

 

Smart house – Procopciuc Ghennadi

 

This project is more or less an umbrella for any Linux application which could control various aspects of the house - lights, water, temperature, video cameras, fridge and so on.

 

 

Generation of C/C++ header files from ARM pdf file - Procopciuc Ghennadi

 

This project is aiming at generating C/C++ header files by parsing various documents. While less spectacular maybe in its aim it could be a very useful tool.

 

IOT Gateway - Procopciuc Ghennadi

 

This project aims similar to the smart house Linux applications that can control various sensors and take decisions. There is an endless number of applications that can be done.

See the general topic of IoT for example at: Internet of Tomorrow - IoT Tour - Home | Freescale

 

Game Console/Media Player – Florina Terzea

This project aims at creating a game or turning the i.MX into a real media player by creating and combining existing applications.

 

The idea would be to use existing applications running over Linux or Android in order to create a games console based on the chosen board. The games should be integrated in a common interface and one game could be developed from zero taking the accelerometer sensors into account.

The board will have to be enhanced with at least an LCD panel and some control buttons (could use the ones on board already but some bigger hardware buttons could work better if they are fit in a case).
Sources of inspiration for this kind of projects would be the available on the internet raspberry PI projects:

Best Raspberry Pi gaming creations | Vox Magazin

 

The media player option could use existing software for Linux or Android in order to create an interface similar to one of the options already available for raspberry PI:

4 Great media center software for Raspberry Pi . Whether it's about porting some existing code and changes here and there or creating the interface from zero and integrate the available programs into it in order to offer a unified experience to the end user of the media center. This project does not need any additional hardware.

 

The project is subject to changes from students, taking their interest into account.

 

Feeding My Pet - IoT – Mihaela Panescu

 

The project's aim is to create an automated feeder for your pet while you are not at home.

This can be designed to be remotely controlled (web/mail server) or to periodically release food/water.

Other fun features can be added to it like a live camera so that the pet can be monitored.

 

Here is an example of a cat feeder:

http://www.cnet.com/pictures/the-most-ambitious-raspberry-pi-projects-pictures/3/

 

 

Control Kinetis Zumo Robot – Daniel Scurtu

This project combines two distinct parts:

- a Zumo Robot with infrared sensors controlled by a Freescale KL25Z on a Freedom Board communicating via Bluetooth with

- an i.MX6 core running Linux environment for the host application part.

 

This setup - which will be described below with links to its components - is incredibly powerful in creating a large number of application. Basically the robot moves around scanning its way in infrared and transmits or listens via Bluetooth. On the other side the applications written in Linux can do all kind of "tricks":

- start/stop/speed up/make robot sing (as it has a small speaker as well) from a microphone on the host (Linux) via voice recognition app

- scan a labyrinth to figure out the shortest way out

- make it park on its own

 

Once the setup is put in place and understood only one's imagination is limit for the applications which can be written. Additional sensors - like an ultrasound scanning can be added if required for more fun.

 

Almost complete details available at the following links:

- Purchase Zumo Robot for Arduino; v1.2 (Assembled with 75:1 HP Motors)

- Freescale Freedom Development Platform for Ki|Freescale

 

 

Intelligent traffic lights – Andrei Trandafir

This project aims to simulate a network of traffic lights which monitor and attempt to optimize traffic flow inside a city. The ultimate goal would be to reduce or eliminate long traffic queues, which are extremely undesired since they cause delays, excessive pollution and congestion. In order to achieve this we combine IoT and Automotive concepts by having one intelligent traffic light for each street that flows in an intersection. We assume that all vehicles are also intelligent and can communicate with that traffic light as they approach the intersection. The traffic light can then monitor in real-time the size of the vehicle queue and track how the queue splits after it leaves the intersection. By then communicating with neighboring traffic lights from adjacent intersections, it can dynamically adjust its own timings for the Red, Yellow and Green lights. By having this done throughout the day by the entire network of intersection lights, the traffic flow can be optimized and the average travel time can be reduced. This, in turn, saves time, reduces pollution and fuel costs.

 

The team working on this project will implement:

- a city road network, described as a digraph

- a set of vehicles, described as points flowing on the edges of the digraph on random or guided paths

- a set of traffic lights that reside on the nodes of the digraph

- an algorithm that runs continually on every traffic light and attempts to optimize the traffic flow and reduce congestion

- an alternative, naive algorithm which uses static traffic light timings, whose performance should be compared with the dynamic algorithm in order to show the benefits of the latter

 

Useful links:

- Intelligent Traffic Management Systems - YouTube

- SMART Signal: Monitoring the Real-Time Performance of Arterial Traffic Signal Systems - YouTube

- Freescale brings the Internet of Things to the vehicle - IoT Conference

 

 

Intelligent GPS for calculating the fastest route - Andrei Trandafir

A modern GPS is capable of generating a route from a start point to a given destination which can be optimized for distance or time - the latter taking into account prior knowledge of the existing traffic in order to generate a route that is the fastest, even if not necessarily the shortest. In most cases, the latter feature is more desirable. This project aims to simulate a GPS with time-optimized routing. To achieve this, the GPS will connect to a simulated data-center which keeps track of the real-time traffic distribution inside a geographic area (like a large city). The data-center may also use machine learning techniques to anticipate traffic on a given street direction based on the current time of day. When asked for a route, the GPS will query the data-center and calculate the fastest route, taking into account the traveling interval during the day (morning, afternoon, evening etc.) in order to estimate which streets will be the quickest to transit.

 

The team working on this project will implement:

- a road network, described as a digraph

- a set of vehicles, described as points flowing on the edges of the digraph on random or guided paths

- the data-center, which periodically analyses the traffic on the network and updates road transition times (seen as digraph edge costs)

- the GPS navigator, which connects to the data-center and works with it to calculate a route based on the current traffic layout

- a naive alternative of a GPS navigator which only calculates a route based on distance; this is to be compared with the time-based solution, showing the benefits of the latter

 

Useful links:

- Global Positioning System - Wikipedia, the free encyclopedia

- Garmin nuvi 2797LMT: Navigation Settings with GPS City - YouTube

 

 

Software simulator for vehicle fuel management system - Andrei Trandafir

Every modern vehicle is equipped with an intelligent fuel delivery system, governed by the vehicle's ECU to tightly control the fuel flow to the engine. As part of direct injection systems, this leads to much reduced fuel consumption when compared to older, carburetor-type variants. In addition, it allows the implementation of some important safety systems for the vehicle, such as Traction Control (which when active, among other things, prevents fuel flow to some cylinders in order to reduce engine power output and prevent wheel slippage) or automatic fuel cutoff during accidents. This project aims to simulate such a system, which takes various parameters as inputs (vehicle speed, current engine load, throttle and brake pedal position etc.) and calculates how much fuel must be instantaneously delivered to each cylinder. It also keeps track of how much fuel is left in the tank and can also be tuned to optimize the driving profile for reduced fuel consumption ("eco" driving mode) or maximum performance ("sport" driving mode).

 

The team working on this project will implement:

- a simulation of a vehicle, containing the engine, an automatic transmission and a simplified drivetrain

- the algorithm for the fuel delivery system, as described above

- a GUI that will provide the "driver" with information regarding the vehicle's status: current and average fuel consumption, fuel left, estimated vehicle range etc.

 

Useful links:

- Fuel injection - Wikipedia, the free encyclopedia

- How Fuel Injection Systems Work - HowStuffWorks

 

 

HMI - Human Machine Interface

 

 

An HMI is a car system that allows the driver to easily use different information and entertainment car services. The HMI is usually found at the center of the car, between the driver and the passenger. A typical HMI can be a touchscreen allowing access to Radio & Audio subsystems as well as Navigation and Phone Connectivity.

 

The teams chosing this project will need to design a graphical interractive application that allow the user to:

Play media from various external devices (SD card or smartphone)

Adjust audio and video settings

Make and receive phonecalls from a smartphone connected to the Wandboard

Use the navigation system

 

Examples of HMI below:

 

 

Audio Post Processing Filter

Audio post processing, is the intentional alteration of auditory signals, or sound, often through an audio effect or effects unit. When listening to radio or music in a car, the surroundings around the listener can reduce the quality of listening experience. Traffic or wheel noise, people talking can contribute a lot of noise -- so much noise that some of the soft pieces in the music may become inaudible.Audio post processing can be used to improve the user experiencer by supressing the effects produced.by various noises.The teams chosing this project will have to show sample audio streams recorded with noises and the resulted output after the post processing alghorithms have been applied. Check out the links below dor more details:

 

Real Time Collision Detection

A collision avoidance system is a system designed to reduce the severity of an accident. Also known as precrash system, forward collision warning system or collision mitigating system, it uses radar and sometimeslaser and camera sensors to detect an imminent crash. Once the detection is done, these systems either provide a warning to the driver when there is an imminent collision or take action autonomously without any driver input (by braking or steering or both).One of the first phases in collision avoidance collision detection or detecting an approaching obstacle. A typical application will process images received through a camera connected to the Wandboard, detect objects that are getting closer and closer (having a certain speed as a threshold) and warn by playing a sound or showing a warning message on the screen.  For more details, visit:

 

AFS - Adaptive Front-light System

 

Approximately 70% of vehicle to pedestrian accident occured in night time.

The AFS adapts automatically to the traffic situation and weather conditions for increased safety and greater driving comfort.The teams chosing this project will have to create a prototype for controlling the front-light to different situations: curves, bumps, other traffic vehicles.Check out the links below for more information:

Attachments

    Outcomes