This project is mainly supported by the NSFC 61403325, and RGC-GRF grants.


Location awareness is critical to many indoor applications, e.g., way-finding in a metro station, industrial unmanned ground vehicle (UGV) navigation in a warehouse, and location-based services (LBS) in retail. As reported by MarketsandMarkets recently, the global indoor location market is projected to grow to USD 23.13 billion by 2021. While GPS has effectively solved the ubiquitous localization problem in most outdoor scenarios, a seamless indoor localization solution remains challenging that is capable of providing a comparative user experience to what GPS has done outdoors. It is expected to be accurate, responsive, lightweight, scalable, robust, low-cost and ubiquitous. An expected indoor localization solution should seek a reasonable balance between the system performance and cost so as to achieve widespread adoptions in large-scale environments. The Visible Light Communication (VLC) based solution has been recently in favor with both the academia and industry as a very promising technology to fill this gap.

Our Vision

We see VLC as an appropriate communication technology to augment localization performance in indoor scenarios while keeping the associated cost to a minimum, considering the flourish of LED lighting fixtures occupying the lighting markets. Benefiting from the research experiences in robotics for years, we also accommodate multi-sensor fusion, Simultaneous Localization and Mapping (SLAM), and machine learning techniques in our research. Our vision is towards a low-cost and light-weight VLC-centric indoor localization solution for either mobile robots or human individuals with guaranteed latency, accuracy, and scalability.

Research Area

We have been engaged in the study of VLC-based localization almost from scratch since the year of 2013. To be specific, we have completed the hardware platform assembly, light source radiation pattern modeling, mixed visible light signals decomposition, probabilistic light intensity distribution modeling, path planning within the proposed model, and localization validation in a realistic environment. As for the on-going research, we have iterated the hardware, the low-level physical and mac layer design of VLC-compatible lights, so as to provide an easy-to-use VLC-based localization service in a scalable manner. Moreover, we are now leveraging more sensing modalities aside from VLC, such as RF signals of opportunity (WiFi, Bluetooth, etc.), geomagnetic fingerprints, pedestrian dead reckoning (PDR), and a few to name. With the aid of either the probabilistic or graph -based sensor fusion framework, heterogeneous sensing information will be combined in a mathematically sound way to yield a promising gain in localization performance without increasing significantly the budget burden. For further details, please refer to our publications below.




Technical Reports


This dataset is taken in an environment with Visible Light Communication (VLC) light beacons, for the purpose of low-cost localization using VLC.

Author Kejie Qiu, Fangyi Zhang
File Type rosbag
Topic /energy Raw light intensity signal
/map Reference map (for visualization)
/tf Transforms as groundtruth

There are 70+ separate rosbags in the zip. The total length is over one hour.