High prEcision and Robust Blind guidance System – HERBS

According to the World Health Organization (WHO), the number of visually impaired people is expected to triple by 2050, reaching 588M in the world. Today, in France, this population represents 1.8 million people. They remain poorly integrated in our society, in part due to mobility constraints. In fact, there are very limited solutions in guiding the user towards a specific desired location, especially in an unknown environment. The people's assistance is a feasible solution, but it does not improve the idea of autonomous guidance based on the knowledge of the person's location. Commonly used positioning solutions, like GPS and equivalents, offer good performance for outdoor mobility, but indoor performance remains insufficient for safely guiding a visually impaired person. Our project will unlock visually impaired people real time guidance by enabling cost effective precise and reliable indoor localization. To achieve that we propose advanced concepts on indoor localization techniques targeting high-accuracy indoor localization, including blind and visually impaired people in the evaluating process, while leveraging the unprecedented availability of data using the combination of new Artificial Intelligence techniques to achieve high-accuracy localization. The project will also investigate how to tackle the data hungry fingerprint-based localization problem. In our approach is to consider Non Line Of Sight environment and propose a localization solution by investigating all levels from radio channel to Artificial Intelligence algorithms including modulation waveform and using at each stage real visually impaired guidance case scenario defined and validated by a demonstrator.