各位同學,
Dear students,
Our department intends to open an intensive short-term course. If you are willing to select this course, please come to ECE office and find Ms. Liao for filling out the course selection form.
The course description is as follows:
Course Code & Name: ET5927701電子工程專論(七)
Special Topics on Electronic Engineering (7)
Topic: Signal analysis and vital sign extraction based on non-obstrusive, wearable sensors
Instructor: Dr.-Ing. Jan Geldmacher
Lemonbeat GmbH, subsidiary of innogy SE, Dortmund, Germany.
Position Senior Engineer, Project Management for Embedded Systems, Product Owner for
Embedded Stack Development
Class dates & Class hours:
19th Nov. 2019 (Tue.)~ 22nd Nov. 2019 (Fri.)
18:25- 21:05(第A,B,C節) (Every day)
23rd Nov. 2019 (Sat.)
09:10-12:10 &13:20-16:40 (第2,3,4節&第6,7,8節) (One day)
Total time : 18 hours
Credit: 1
Course language: English
Class Room: IB607
Abstract:
Sensors that can measure motion signals with high resolution and accuracy are widely available today: A common smartphone features acceleration and motion sensors that can be used to deliver manifold information about its user and its environment. For example, a simple one-channel motion signal from a sensor that is attached to a persons body or bed during sleep can reveal information about heart rate and respiration rate with very good accuracy. Even medical conditions and disorders can be predicted from these data sets to some extend.
In this lecture we will do a case study and introduce and apply digital signal processing techniques to analyze data from non-obtrusive motion sensors. In the course of the lecture the students will develop algorithms that can extract parameters like heart rate or respiration rate from these signals. This is done based on real-world data sets, which have been collected in clinical studies using motion sensors and ECGs.
This lecture will recapitulate some required basics of linear algebra. Based on this it will be shown how different kinds of signal processing problems can be represented using linear algebra techniques. Each course thus consists of a theoretical part, where concepts and background of certain DSP problems are explained. Equipped with this knowledge, the students will then approach a practical problem where these concepts can be applied. Real-world data will be provided and the implementation will be done using MATLAB (or GNU Octave). MATLAB is used since it is a popular solution for fast prototyping of DSP algorithms and since it is well suited for implementation of linear algebra based algorithms.
The course consists of 6 parts. Each part consists of a lecture, a short tutorial and practical work using MATLAB. The students will work in groups on the assignments and will be provided with advise and pointers if needed. There will be a written exam, which will cover basics of the introduced concepts.