Identifying Everyday Objects with a Smartphone Knock

Abstract

We use smartphones and their apps for almost every daily activity. For instance, to purchase a bottle of water online, a user has to unlock the smartphone, find the right ecommerce app, search the name of the water product, and finally place an order. This procedure requires manual, often cumbersome, input of a user, but could be significantly simplified if the smartphone can identify an object and automatically process this routine. We present Knocker, an object identification technique that only uses commercial off-the-shelf smartphones. The basic idea of Knocker is to leverage a unique set of responses that occur when a user knocks on an object with a smartphone, which consist of the generated sound from the knock and the changes in accelerometer and gyroscope values. Knocker employs a machine learning classifier to identify an object from the knock responses. A user study was conducted to evaluate the feasibility of Knocker with 14 objects in both quiet and noisy environments. The result shows that Knocker identifies objects with up to 99.7% accuracy