BME/CENG/ELEC/SENG 499 Projects
Supervisor: Mihai Sima, email@example.com (Jan. 2010)
MS1 FPGA-based solutions for Discrete Fourier/Cosine/Wavelet Transform
Problem definition. Reviewing the theory of the discrete Fourier/Cosine/Wavelet transform. Research on the literature regarding implementation schemes. Selecting the appropriate implementation with respect to a particular FPGA family (e.g., Xilinx). Writing VHDL code followed by a place and route procedure. Simulation of the resulting design. Identifying the advantages and drawbacks of the FPGA-based solution versus full-custom and pure-software solutions. Team: 3 students. Prerequisites: Good marks and enthusiasm.
MS2 Voice Dialer on a low-end digital signal processor
Problem definition. Reviewing the theory of linear prediction and spectrum analysis. Building a minimal acoustical database. Implementing a classification algorithm based on Dynamic Time Warping. Testing the final design and extracting recognition rate figures. Team: 3 students. Prerequisites: Good marks and enthusiasm.
MS3 Magnetic field measurement by means of Hall sensors
Problem definition. Reviewing the theory of the Hall effect. Selecting the appropriate Hall sensor with respect to the design parameters. Designing and implementing the instrumentation amplifier. Data acquisition and displaying the results usinga microcontroller development board. Team: 2 students. Prerequisites: Good marks and enthusiasm.
The Driver Intent Collision Evasion System is an algorithmic approach to assist vehicle operators in avoiding potentially damaging accidents while driving. This system will consist of a series of sensors used to collect information about how a driver is operating his vehicle, an algorithm to convert the raw data into an estimation of intent, a communication protocol to alert and coordinate other vehicles nearby to the driver as well as an interface to alert these other drivers of possible upcoming collision possibilities. It is conceivable that any number of algorithms could be written to predict any collision scenario but for the purposes of this project only a few of the most common will be inspected. These common scenarios include running a red light, rear ending a vehicle and merging into traffic. The system will consist of two major parts: a simulator for determining the effectiveness of different algorithms and a scale model using remote control cars to mimic the system in the real world. The simulator will take input from either a keyboard/mouse pair or a racing wheel joystick. This input will simply change the vehicle being controlled in the simulator or will be propagated to the remote controlled cars to imitate actually driving the car. Using this setup, the algorithms can be evaluated both purely in software as well as in the physical domain.
The goal of this project is to create a steering wheel augmented with piezoelectric elements and sensors to provide feedback to a driver without having to take their attention away from the road to look at displays. A major outcome of this project is to mount embedded adhoc wireless systems into two vehicles and have the drivers steering wheels vibrate in the direction of the other vehicle. This system would allow drivers to detect oncoming cars, or cars in their blind spots without taking their attention away from the road itself, as well as providing a safety mechanism in case the driver has not noticed a vehicle leading to a high-risk-of-collision situation.