{"id":14666,"date":"2019-07-22T13:59:36","date_gmt":"2019-07-22T11:59:36","guid":{"rendered":"https:\/\/www.cosinuss.com\/?p=14666"},"modified":"2020-08-17T10:25:45","modified_gmt":"2020-08-17T08:25:45","slug":"activity-recognition-of-running-and-walking","status":"publish","type":"post","link":"https:\/\/www.cosinuss.com\/en\/2019\/07\/22\/activity-recognition-of-running-and-walking\/","title":{"rendered":"Activity recognition of running and walking"},"content":{"rendered":"

Learn more about your own training than just the number of steps you have taken? Move in an optimal way without being exposed to too much strain? Prevent injuries and recognize diseases?
\nCosinuss\u00b0 One is now able to recognize activities and distinguish between running and walking, so that you can see whether you have completed a marathon or just did a sightseeing day. The movements are analysed and can be used for health or fitness related applications. <\/strong><\/p>\n

Why is it important that the type of movement is recognized correctly?<\/strong><\/h2>\n

The examination of the human movement is called \u201agait analysis\u2018. Gait is the way the user walks or runs. For health purposes, gait assessment helps prevent injuries to professional athletes, detect Parkinson’s disease, or assist a patient’s recovery.
\nFor fitness purposes, it is useful to determine which distance has been covered or how long has been trained and with which intensity. These parameters are important for controlling one’s own training, because only in this way the best possible result can be achieved.<\/p>\n

Why is the cosinuss\u00ba One particularly suitable for activity recognition?<\/strong><\/h2>\n

Research has shown that the head is the steadiest part of the body during body movement, thanks to the balance of the trunk and neck. This makes it robust and less affected by movement than other parts of the body. The accelerations achieved in the ear are therefore more regular and weaker. In addition, the sensor can be worn for longer periods of time without slipping or restricting movement.<\/p>\n

Our study conducted to test activity recognition<\/strong><\/h2>\n

We performed a study in the Performance and Prevention Lab of the TUM<\/a> with a total of 18 subjects aged between 18 and 50 years. They all had to walk and run at different speeds to get as many variations as possible. Each participant had a different gait pattern when walking and running. As a result we obtained a comprehensive database that can now be applied to any type of subjects.<\/p>\n

Based on this database, we created a machine learning model that classifies if the user is walking or running. The results obtained show a high degree of accuracy and precision in the recognition of activities. Thus, the cosinuss\u00b0 One<\/a> is able to work as an everyday monitoring system for health and fitness purposes.<\/p>\n<\/div>

<\/div><\/div><\/div><\/div><\/div>\n
Share this post! \/ Teile diesen Beitrag!<\/div>