Skip to main content

Adaptive acoustic monitoring system

Primary Investigators:
Akos Ledeczi
Brief Description of Project:
Real-time acoustic signal processing is a well-studied field of cyber-physical system. More recently, machine learning (ML) techniques have been applied to enhance the accuracy of acoustic event classification and speech recognition. However, carrying out these tasks in embedded systems are more challenging due to resource constraints. The goal of this research project is to investigate sophisticated resource-management algorithms that adapt to the environment they are placed in, with a special focus on acoustic applications. The integration of ML and well-known adaptive methods can lead to data-driven resource-aware solutions. The tasks include the examination of different resource-optimization techniques, the implementation and analysis of several related approaches, and the hardware-level utilization of one such method.

Desired Qualifications:
Background in signals and systems, software design, optimization, signal processing, embedded systems, Python and C/C++
Nature of Supervision:
Work closely with research scientist
A Brief Research Plan (period is for 10 weeks):
Weeks 1-2: literature survey
Weeks 3-5: design and prototype on desktop
Weeks 6-9: implementation on embedded platform
Week 10: evaluation
Number of Open Slots: 1
Contact Information:
Name: Akos Ledeczi
Department: Computer Science