Research Interests:

Modern signal processing relies heavily on underlying statistical models, and the performance of the resulting algorithms is largely determined by the accuracy of the assumptions made within these models. To make the most of modern sensor data it is crucial to include available information regarding the physics of the sensor into the model, while limiting other potentially costly assumptions regarding the model structure. Nonparametric Bayesian approaches to statistical modeling allow us to include uncertainty regarding not only the parameters of the model but also the structure of the model, thereby limiting the number of necessary assumptions and creating flexible model structures that are automatically adapted to meet specific operational goals. Kenneth Morton's current research concerns the development and application of nonparametric Bayesian models to signal processing and machine learning problems in a variety of application areas. He is actively working on problems concerning acoustic signal classification, the detection of buried landmines in ground penetrating radar, and the analysis of laser induced break down spectroscopic data for the detection of explosive contaminants and discrimination of geological samples.

Specialties:

Signal Processing
Sensing and Sensor Systems
Land Mine Detection
Acoustics