580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How are hardware architectures optimized for efficient real-time digital signal processing in embedded systems?
This theme investigates the design considerations, architectural innovations, and resource optimizations for implementing DSP algorithms on hardware platforms such as FPGAs, DSP microprocessors, and ASICs. The research focuses on enabling fast, low-resource, and flexible execution of computationally intensive algorithms integral to DSP applications, including image decoding, motor control, and signal modulation techniques, while often aiming to reduce external dependencies and improve throughput.
Key finding: Demonstrates a novel, largely on-chip implementation of first-order sigma-delta ADCs exploiting FPGA LVDS input buffers as comparators, minimizing external analog components. This approach achieves audio-range conversion with... Read more
Key finding: Presents a hardware architecture for digital image decoding using components optimized for logic and memory minimization, including entropy decoders, probability estimators, dequantizers, and inverse subband transformers... Read more
Key finding: Introduces a hardware digital architecture on FPGA for space vector PWM controlling three-phase inverters, enabling real-time variation of carrier and switching frequencies, phase, and amplitude without external references or... Read more
Key finding: Develops an Application Specific Instruction Set Processor (ASIP) with a high-speed multiplier-accumulator unit employing Wallace tree structures and fast two’s complement algorithms optimized via adaptation vectors. FPGA... Read more
Key finding: Comprehensively details fundamental DSP system elements, emphasizing computationally efficient algorithms for real-time digital processing, including ADC/DAC components, sampling theory, FFT structures, and digital filters.... Read more
2. What are the state-of-the-art methods for real-time digital control and signal processing algorithms implemented on DSP platforms?
This theme focuses on the software and algorithmic implementations of critical DSP functionalities on commercial DSP platforms, including digital filtering, modulation, adaptive control, and communications protocols. Research explores how to map complex signal processing and control algorithms—such as pulse Doppler radar processing, vector control for motor drives, and reconfigurable OFDM modulation—onto DSPs to achieve real-time operational performance with flexibility and precision.
Key finding: Describes a TMS320C50-based DSP board implementation of pulse-Doppler radar signal processing, enabling digital clutter suppression and advanced detection via replacement of analog Doppler filter banks with N digital filters.... Read more
Key finding: Presents closed-loop motor drive system implementation using dSPACE DS1104 DSP hardware with PWM control and PI speed regulator, targeting asymmetric converters and separately excited DC motors. Simulation and experimental... Read more
Key finding: Details the hardware deployment of dq-PI current control algorithm on TMS320F28335 DSP, synthesizing SPWM signals to drive a three-phase VSI. Validation through MATLAB/Simulink simulation and hardware testing confirms the DSP... Read more
Key finding: Demonstrates a DSP-based implementation of a flexible, reconfigurable OFDM transmitter/receiver system capable of rapid code updates and low latency. Practical measurements from spectrum analyzers and oscilloscopes confirm... Read more
Key finding: Although focused on fingerprint image enhancement, this work includes algorithmic methodologies closely related to DSP—such as statistical texture feature extraction and Gabor filter-based enhancement—serving as early... Read more
3. How can adaptive and stable observer-based control schemes be designed for sensorless motor drives using DSP hardware?
This theme explores observer design and stability analysis methods for sensorless induction motor drives, particularly under challenging conditions such as very low speed and regenerating operation modes. It highlights the role of advanced algorithms including rough set theory, genetic algorithms, and Lyapunov-based stability criteria in selecting feedback gains and state observer parameters. The practical deployment leveraging DSP boards further illustrates the connection between algorithmic development and real-time embedded control.
Key finding: Proposes a novel combined rough set and genetic algorithm based method for optimizing observer feedback gains in sensorless induction motor drives operating at very low speeds and in regenerating mode. The approach reduces... Read more