Recent developments in technology have definitely brought changes into our lives. Gaming is becoming the new approach of socialization as well as a new style of life for people based on their interaction with devices and applications.
Fabio Aiolli, Matteo Ciman, Michele Donini and Ombretta Gaggi from the Mathematics Department of the University of Padua, Italy have developed a new machine learning based technique to recognize and count stairsteps when a person climbs or descends stairs. They have presented the study at the 11th International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services.
The developed classification algorithm have been presented as part of ClimbTheWorld, “a smartphone serious game aiming at incentivize people to use stairs instead of elevator”. Through this game people should be stimulate to leave the sedentary life which all of us moved to in recent decades.
First of everything, what are the main advantages of ClimbTheWorld?
- it is able to recognize a single stairstep vs. a simple step, during the activity itself, providing an online, fine-grained classification
- It has no restrictions on the smartphone orientation: a new method is introduced to translate data received from the accelerometer to a fixed coordinate system,
- It uses smarthphone sensors instead of requiring to buy expensive tools
- energy consumption has a key role for development of the serious game. The effect is to avoid the constant recharge of smartphones.
The importance of lower energy consumption is a key point to understand the developed method: “Our experiments have shown that the proposed method performs better than the native Android solution in terms of consumed energy. The best trade-off between energy saving and precision is reached with KOMD classification algorithm combined with our method for data acquisition, using a sampling rate of 30Hz (in this case, the obtained precision is 85%, recall is equal to 86%)” – the researchers said. They also underlined: “To the best of our knowledge, this is the first work that deals with energy consumption issues in a classification problem. Finally, we have obtained a counting performance of 99.4% using this configuration”.
The researches highlighted the idea under the game through an example: let’s imagine the user climbing real world buildings, e. g., the Empire State Building or the Eiffel Tower, making stairs during the everyday life. Once started, the game records and analyzes data from the accelerometer and counts the number of stairsteps made by the user, even when the application goes in background.
The game proposes different difficulty levels: easier levels correspond to a lower number of stairsteps and provide a lower quality and number of photos from the top of the building
To constantly encourage and help the users a set bonuses has been proposed by the researchers “For example, in case your performance is higher than the one of the day before, you get a 10% increase on the total number of stairsteps made”.
Last but not least the user can also share results with friends on Facebook. According to the researchers this is a good way to engage people as well as enhance a more dynamic lifestyle.
Interested in ClimbTheWorld? Find here all the details of the research.
