detectBiklio – detect bicycle usage with an Android smartphone app

September 21, 2021

The increasing concern related to air pollution levels caused by motorized vehicles and obesity levels, along with the technology evolution and its decreasing cost (making it more affordable), created new strategies to reduce the emission and obesity levels. One of them is using the existing sensing power available on users smartphones and determine his transportation mode rewarding her/him with gifts if she/he chooses bicycle as his transportation mode. Thus, the goal of this work is to improve the existing transportation detection algorithm accuracy of Biklio, a sustainable mobility encouraging app. We propose a machine learning (ML) algorithm for detecting if a user is using a bicycle. The ML algorithm is based on Random Forest. It replaces the current solution used in Biklio that uses the Activity Recognition API from Google; increasing the detection accuracy, we expect to minimize the number of cases where the output does not match reality.

Authors in the community:

• 

  B

  Bruno Alexandre Oliveira Santos

  ist182053

Supervisors of this institution:

• 

  P

  Paulo Jorge Pires Ferreira

  ist12958

electrical-engineering-electronic-engineering-information-engineering - Electrical engineering, electronic engineering, information engineering

eng - English

Instituto Superior Técnico