Signal Processing

Signal processing is a discipline that deals with the transformation and manipulation of signals for information extraction, signal estimation, and efficient representation of signals. Digital signal processing (DSP) has a wide range of applications and has become a critical component of almost all modern electronic devices. DSP finds applications in speech and audio signal processing, image processing, digital communications, control and robotics, medical devices, and many other industrial and consumer electronics. The application of DSP in speech and audio processing includes speech compression, speaker identification and verification, noise cancellation, music equalization and manipulation, etc.

DSP techniques have made a wide range of image processing applications possible, such as face recognition, image enhancement, and image compression. The recent rapid growth and expansion of wireless communication devices, systems, and networks are made possible because of the sophisticated DSP algorithms. Similarly, radar systems and robotics owe their existence and growth to DSP. Also, recently, biomedical DSP has spurred the development of many life-saving wearable medical devices.

The UW Bothell EE graduate curriculum includes a sequence of two courses on digital signal processing courses, B EE 511 and B EE 512, and a course on image processing, B EE 515. The DSP courses cover the basic signal processing concepts such as analysis of signals and systems, design of filters, multi-rate processing and design of polyphase filters, and advanced signal processing algorithms such as the design of optimum filters, adaptive filters, Kalman filters, and spectrum estimation. For deeper knowledge, students are encouraged to pursue B EE 600 Independent Student or B EE 700 Thesis with a faculty member.

Career pathways

Experiences from classes in this key focus area of Signal Processing will prepare students for wide employment opportunities in the field of signal processing. DSP engineers work in various industries, including telecommunications, aerospace, semiconductor, consumer electronics, etc. Students should be prepared for competitive consideration of employment positions such as:

  • Signal & Image Processing Engineer
  • Camera Algorithm Engineer
  • Principle Active Noise Cancellation Engineer
  • Camera Systems Engineer
  • Financial Signal Processing Engineer
  • Speed and Audio Processing Engineer
  • Machine Learning and Artificial Intelligence Engineer

The Institute of Electrical and Electronics Engineers Signal Processing Society shared “3 Reasons Why Signal Processing is the Career of the Future”:

  • Signal Processing Plays a Key Role in Multiple Industries: Signals are used in finance, to send messages about and interpret financial date. The most exhilarating new movies are made possible by multi-resolution signal processing, making entertainment a lucrative market for people with this skill set.
  • Signal Processing is the Technology of the Future: The latest breakthroughs in health care are enabled by signal processing engineers, who are developing ways to process medial images more quickly and accurately.
  • Signal Processing Can Serve a Social Purpose: Signal processing allows for the expansion of computing power and data storage capabilities, making signal processing engineers indispensable for understanding and tackling our biggest global problems.

Faculty in Signal Processing

Harry Aintablian

Kaibao Nie

Tadesse Ghirmai

Research at UW Bothell

Dr. Harry Aintablian’s research experience is in power electronics for space systems. At UW Bothell he has supervised several student research projects in power conversion, reliability analyses of electronic systems, and power management of photovoltaic systems with energy storage. His current research involves the development of high-voltage, high-frequency power supplies for electrohydrodynamic applications such as Unmanned Aerial Vehicles (UAVs) and a COVID respirator mask.

Dr. Kaibao Nie’s research focuses on signal processing in cochlear implants. A cochlear implant is an electronic device to restore hearing to people with profound hearing loss. It can convert sounds to electrical current pulses for directly stimulating the auditory nerve. The research topics on cochlear implants span a wide range of techniques including noise reduction, speech coding, collection of electrically evoked action potentials, cortical EEG processing and auditory modeling. View a conference paper Dr. Nie coauthored on speech processing in hearing devices, speech perception with cochlear implants, and keywords spotting.

Study Signal Processing

The curriculum reflects depth and breadth of faculty research expertise and provides graduate students with a solid foundation in signal processing and digital image processing. Relevant courses include:

Learning objectives

In this technical area, students will learn:

  • Industry standard simulation tools such as Matlab, Spiece, and LabView. Those tools are extensively used in industry as well as in academia for the simulation of design and analysis in several disciplines of engineering such as signal processing, circuit design, system fault analysis, and instrumentation interface.
  • Basic probability and random processes and their applications to engineering.
  • Digital signal processing techniques for analysis of systems and designing of digital filters.
  • Statistical signal processing which deals with random signals, their modeling, characterization, and transformation to extract useful information about the underlying mechanism that generates them.
  • Advanced image processing techniques, image filtering design, and its applications to images acquired from various imaging techniques.
  • Concepts of predictive learning algorithms for supervised and unsupervised learning tasks.

Emphasis on project-based learning through class projects

Many of the courses listed above provide class projects that will enhance student learning. Particularly, students in a team-oriented project learn important skills such as collaboration, communication, and presentation. Both DSP courses (B EE 511 and 512) have significant design homework problems and projects, where MATLAB is extensively used as a tool.

[B EE 510] Probability and Random Process

Handwriting recognition is an important machine learning problem that has many practical applications. In this design project, students will develop learning and classification algorithms to identify the number on an image of a handwritten digit (0 to 9). Students create programs in Python or Matlab to perform handwritten digit recognition using the Naïve Bayes model and the Gaussian mixture model. The purpose of this project is to introduce students to the basic concepts of Bayes classifier and machine learning.

[B EE 511] Signal Processing I

Students can work on pre-processing and feature extraction of electrocardiogram (ECG) signals; design lowpass, highpass and notch filters to remove noise and power-line interference from the ECG signal; and develop algorithms to determine the average heart rate and heart rate variability from the ECG signal.

[B EE 512] Signal Processing II

Students can work on speech signal modeling and linear prediction. The first part of the project requires students to represent a speech signal using parameters of an all-pole filter and then synthesize back the speech signal from the filter parameters. In the second part of the project, students predict future stock values using a linear predictor.