John J. Guiry, Pepijn van de Ven, and John Nelson
ambient assisted living, unobtrusive mobility monitoring, smartphone, gravity vector estimation
Recent advancements in smartphone technology have showcased the viability of such devices to the field of human mobility monitoring. At the time of writing, it is commonplace to find smartphones containing sensors such as accelerometers, magnetometers, gyros, barometers and GPS. The widespread prevalence and acceptance of smartphones in society makes their usage as accurate mobility monitors even more appealing. However, one great challenge posed by smartphones is that their location and orientation is not normally known. This information is extensively used by state-of-the-art algorithms for physical activity monitoring. Moreover, in spite of their powerful processors, smartphones often need to prioritise other tasks than those necessary for obtaining timely sensor information. In this paper, the authors design, implement, test and validate a mobility monitor algorithm across a range of Android based smartphones. A trial with N=6 subjects was incorporated into the study, to investigate activities including sitting, standing, cycling walking, jogging & running. Provisional results appear promising, with average accuracies of 88.8% produced by the real-time mobility monitor, when using a custom classifier. Methods were also deployed which allow existing fixed position based algorithms to function in an orientation independent manner.
Important Links:
Go Back
