Author(s)

Yuheng Ding

Corresponding Author

Yuheng Ding

ABSTRACT

The need for a fast and accurate indoor positioning solution is increasing rapidly since people are spending more time indoors, mostly for shopping and entertaining. Being aware of their own positions is not only vital to consumers’ safety, but also necessary to improve their shopping experience. In this paper, an alternative way for indoor positioning is proposed, which is based on convolutional neural network, particularly an algorithm called Single Shot Multi-Box Detector (SSD). Different from other indoor positioning technologies, this image-based technique does not depend on special detecting devices or communication networks. Testing shows that positioning with SSD in indoor spaces has great advantages (both convenience and accuracy) over other methods, indicating its potential to be applied in real-life situations.

KEYWORDS

Indoor positioning method, convolutional neural network, Single Shot Multi-Box Detector (SSD)