Mobile Sensing

Networking and Services
Advanced studies
Year Semester Date Period Language In charge
2015 spring 10.03-30.04. 4-4 English Petteri Nurmi


Time Room Lecturer Date
Tue 16-18 B222 Petteri Nurmi 10.03.2015-30.04.2015
Thu 16-18 B222 Petteri Nurmi 10.03.2015-30.04.2015

Exercise groups

Group: 1
Time Room Instructor Date Observe
Thu 18-20 BK107 Samuli Hemminki 12.03.2015—12.03.2015
Thu 18-20 B222 Samuli Hemminki 19.03.2015—30.04.2015

Information for international students

All lectures and exercise sessions will be in English.


PROJECT WORK: If you haven't already, please send information about your group and topic of the project to the lecturer and both course instructors by Tue 21.04.


Home Exam

EXERCISE POINTS: We have finished scoring the exercises and those who had less than the required 50% of exercise points have been contacted by email. If you were contacted, we warmly suggest doing the last (BONUS) task in home exam to supplement missing points. Everyone not contacted should skip the last (BONUS) task.





Mobile (phone) sensing refers to the use of sensors embedded on mobile devices to obtain meaningful information about the user of the device or the current environment. Common examples of mobile sensing include step counting and other types of mobility monitoring applications on mobile phones as well as different kinds of ambient "fingerprinting" applications, e.g., applications that can automatically detect places that the user has visited.


Mobile Sensing is an advanced level course that covers algorithmic and system-level aspects in designing and developing mobile sensing applications. A rough outline of the course topics is as follows:

  • Basics of Mobile Sensing: Sensors, System Pipelines, Evaluation Metrics
  • Ambient Sensing: WiFi, Audio, and other sensors.
  • Inertial Sensing: Accelerometer and Gyroscopes
  • Psychological and Social sensing: App sensors, prosodic sensing
  • Emerging topics

Lecture I - Fundamentals

Lecture II - Energy-Efficiency

Lecture III - Signal Processing for Sensor Data Analysis

Lecture IV - Activity Recognition

Lecture V - Motion Analysis

Lecture VI - Orientation Estimation

Lecture VII - Audio Sensing

Lecture VIII - Social and Psychological Sensing

Lecture IX - Prosodic Sensing

Lecture X - Physiological Sensing


(Moderated) discussion forum for exercise questions:

I - Lab Exercises

Compulsory exercises:

Exercises II - Selected Solutions

Exercises III - Selected Solutions

Exercises IV - Selected Solutions

Exercises V - Selected Solutions

Exercises VI - Selected Solutions

Exercises VII (Due Mon 27.04)


Completing the course

The course involves a home exam and weekly exercise sessions (starting already during the FIRST week). Successful completion requires achieving at least half of the points in both the exam and the exercises.

The course involves also a compulsory project work where a small-scale mobile sensing application will be developed.

Project Work Instructions

Project work can be done either individually or in small groups, max 4 / group. The scale of the topic should match with the size of the group. Topic needs to be agreed with the lecturer and instructors. Only person/group can have a specific topic.

Three parts:

  1. Implement the sensing pipeline for a selected mobile sensing app. Preferably on mobile device, if not possible, it is also possible to consider an existing dataset of measurements.
  2. Evaluate the implementation. Take into consideration classification accuracy, including robustness against main sources of variability, and energy-effiency.
  3. Report the implementation and the results.

Deadline for project work is on 31.05.

You can pick a project topic from the following list, or suggest your own topic.

Project topic examples: Step Counter, Accelerometer-based Activity recognition (e.g., stationary, walking, running), Magnetometer-based Activity recognition (e.g., car, cycle, metro), Simple game using inertial sensors, Running or Cycling coach (report features such as step frequency, cadence and variation in them), Magnetic field anomality detector, audio sensing methods described in papers: SoundSense, or Sound of Silence, Proximity and orientation based phone placement identifier (..) (TBU)


Groups and Topics


Literature and material

The course does not follow any specific textbook. Links to relevant papers and other material will be given in the lecture slides.