Using acceleration measurements for activity recognition: an effective learning algorithm for constructing neural classifiers

摘要

Thanks to the rapid adoption of smartphones and other advanced consumer electronics, the use of accelerometers has indeed accelerated in the last several years. At the cost of less than $1 US each, these sensors can enable a wide range of applications, from intelligent user interfaces to automated wellness monitoring.rnThis paper studies how accelerometers can be used in activity recognition. In the past, many scientists and engineers have addressed this problem using various approaches; the paper is built on top of them.