1995 TI DSP Solutions Challenge Regional Winners

Territory I (American Division)

• Western U.S.
• Eastern U.S.
• Canada/Latin American

Western U.S.

Entry Topic: 		"Adaptive Active Noise Control for 
			   Headphones"
University: 		University of California-Berkeley
Team Members: 		Angela K. Wang, PhD Candidate 
			Benedict Tse, Graduate
Advising Professor: 	Dr. Wei Ren, Assistant Professor, Electrical
			   Engineering & Computer Science

Abstract

Commercially available active noise control headphones rely on fixed analog controllers to drive "anti-noise" loudspeakers. Our design uses an adaptive controller to optimally cancel unwanted acoustic noise. This headphone would be particularly useful for workers who operate or work near heavy machinery and engines because the noise is selectively eliminated. Desired sounds, such as speech and warning signals, are left to be heard clearly.

The adaptive control algorithm is implemented on a TI TMS32OC30GEL DSP which drives a Sony CD55 headphone/microphone system. Our experiments indicate that adaptive control results in a dramatic improvement in performance over fixed control. This improvement is due to the availability of high-performance programmable DSPs and the self-optimizing and tracking capabilities of the adaptive controller in response to the surrounding noise.

Eastern U.S.

Entry Topic: 		"Real-time Implementation of a 
			   Wavelet-based Video Compression System"
University: 		University of Maryland,
Team Members: 		Hamid Jafarkhani, PhD Candidate 
			Jerome J. Johnson, Undergraduate 
			Rupulu Bhattachrya, Undergraduate
Advising Professor: 	Dr. Nariman Farvardin, Professor and Chair, 
			   Electrical Engineering

Abstract

The objective of the proposal is the implementation of a video coding system of QCIF format (180x144 pixels) on a TI TMS320C4x Parallel Processing Development System (PPDS). Our target bit rates are 64 to 384 kbit/s. The system uses the discrete wavelet transform to decompose video frames into 10 subbands. Each of the subbands is then encoded separately using uniform-threshold scalar quantizers followed by Huffman coding. The required rate for quantizing the coefficients of each subband is determined by an optimal bit allocation procedure. The video frames are reconstructed by performing the inverse filtering operation on the encoded coefficients. Our goal is to perform the decomposition, quantization, and reconstruction in real-time on the PPDS. The decomposition process is done on 2 DSP's; the other 2 DSP's are used for reconstructing the frames simultaneously.

Canada/Latin America

Entry Topic:		"A Microprocessor-Based System for 
			   Synchronized Voltage Phasor" 						Measurements"
University: 		University of Saskatchewan 
Team Members:		Craig Lukie, Undergraduate 
			Jacky Kim Chan, Undergraduate
Advising Professor: 	Dr. Tarlochan S. Sidhu, Associate Professor,
			   Electrical Engineering

Abstract

This project submission documents the work done in the design of a system for synchronized measurements of voltage phasors. It describes the purpose and requirements of the Synchronized Measurement System (SMS). The implementation of the system is outlined. The system consists of three separate blocks which are dealt with separately in this paper. The results of the testing of different blocks and the system as a whole are then presented and discussed in this paper. Finally, these results are used to compare with the original system requirements for determining the success of the project.

The Synchronized Measurement System (SMS) measures the magnitude and phase angle of voltage waveforms at two different locations across a power system simultaneously. Examples of voltage phasors are shown below.


Figure 1: Phasor Diagram

This information is determined by a computerized monitoring system at each location and is displayed and stored on each computer individually. This data could then be transmitted to a central location where the phasors are compared and processed, however this was not a part of this design project.

The SMS requires only two inputs at each remote location. These inputs are the voltage signal to be measured and the synchronizing signal. The synchronizing signal provides a common reference phasor between remote locations and is used to synchronize the local sampling clock. The system then provides two outputs which are the voltage magnitude (in rms volts) and phase angle (in degrees).