Wayne T. Padgett

An undergraduate fixed point DSP course

A course is proposed to cover fixed point, quantization and overflow issues in DSP programming for undergraduates. The course covers the material using a simulation approach in a case study of a SSB communication system. The course is needed in the curriculum because most new graduates in DSP jobs need these skills immediately to be productive on the more common fixed point processors.

Blur identification and correction for a given imaging system

Estimating the blur function is the first step in all the image restoration techniques. A priori knowledge of the blur phenomenon helps improve the quality and certainty of image restoration. In this paper, a method of estimating the blur function for imaging setups that are not subject to change, for example surveillance cameras, is developed using the Wiener filter. A random noise image is used as a test image. Deconvolution is performed between the original test image and its observation taken from the given imaging system, to obtain the blur estimate for the system. Necessary preprocessing steps are implemented to compensate for the non-ideal nature of the imaging environment. Blur correction is then implemented on different observations taken from the same imaging system. The performance of the restoration is evaluated using mean square error and signal to noise improvement criteria.

Teaching design through design competition

The use of competition both inside and outside the classroom, but particularly the use of an external design competition is examined. Although there are a number of difficulties in integrating this type of experience into an engineering curriculum, design competitions remain an extremely motivational and useful tool in the hands of engineering faculty members. This paper details some of the benefits and difficulties associated with various design competitions and experiments with in-class competition.

Extending the usefulness of the Signals and Systems Concept Inventory (SSCI)

The Signals and Systems Concept Inventory is an extremely useful assessment tool that has been widely used and analyzed. The authors have applied it in a variety of situations outside its typical usage. These situations include comparing international students to US students, introducing concept inventories to international faculty, supporting an ABET evaluation report, evaluating the effect of electives on conceptual understanding, analyzing the effect of time between courses on the conceptual understanding, and considering the effect of visual learning style on student performance on a largely graphical test. The SSCI has proven to be a flexible tool that makes student understanding visible and quantifiable in all these situations.

Teaching Fixed-point Algorithm Development in a Systems Context

A course on fixed-point algorithm design using a single sideband (SSB) communication project has been offered three times. The course emphasizes system cost and performance to simulate industry constraints. Students have found the course motivational, and there is a significant need for the topic in industry. Efforts to assess the effect of the course on conceptual understanding are discussed, as is a plan for an associated text

Trends in Signal Processing Education [In the Spotlight]

This summary captures the ideas presented in the Trends session on Signal Processing Education (SPED) at ICASSP 2011. In that session, we surveyed a number of key areas of educational innovation and solicited other trends from the audience. While we strive to be as comprehensive as possible, there are quite likely to be gaps in our treatment.

DSP instead of circuits?-transition to undergraduate DSP education at Rose-Hulman

We assert that digital signal processing (DSP) can and should be taught early (sophomore-junior) in the electrical and computer engineering curricula. This paper looks at the impact this has on the rest of the curriculum, both in electrical and computer engineering and in other engineering curricula. While the early introduction of basic DSP makes it possible to offer better senior electives and graduate courses in DSP, the biggest benefit is the ability to build on DSP core concepts just as we have traditionally built on circuits core concepts in the past. Further, motivational examples in DSP lend themselves to multimedia and are often more familiar to todays students than basic circuits.

Efficient parallel implementation of the LMS algorithm on a multi-ALU architecture

A block exact formulation [1] of the LMS algorithm is well suited for efficient implementation on multiple-ALU architectures. The method can be applied to any number of ALUs. An example analysis is given for the StarCore SC140 core which shows a 33% speed increase for large filters.

Detection of nonstationary random signals in colored noise

This paper describes a novel method for detecting nonstationary signals in colored noise. A first order complex autoregressive, or AR(1), signal model is used which restricts the application of the detector to low order signals, i.e., those which are well modeled by a low order AR process and have only a single spectral peak. The detector assumes the noise covariance is stationary and known. The likelihood function is estimated in the frequency domain because the model simplifies, and the nonstationary frequency estimate can be obtained by an algorithm which approximates the Viterbi algorithm. The AR model parameters are then used to form the appropriate covariance matrix and the approximate likelihood is calculated. Therefore, the detector uses efficient approximations to approximate the generalized likelihood ratio test (GLRT). Simulation results are shown to compare the detector with the known signal likelihood ratio test.<<ETX>>

Performance analysis of a detector for nonstationary random signals

The detection of nonstationary random signals is an important sonar problem which also has potential applications in diverse areas such as biomedical signal processing and spread spectrum communications. The primary problem with applying a powerful test like the generalized likelihood ratio test (GLRT) is the computational effort required to search for the maximum likelihood model parameters for the observed signal. The computation required is multiplied many times over when a signal parameter is nonstationary. A computationally efficient detector of nonstationary Gaussian random signals based on the GLRT was presented at ICASSP94 [1]. A slightly enhanced version of the detector is described below, along with new simulation results demonstrating that the detector performs nearly optimally and is quite robust to signal model inaccuracy.