Abstract:

Many products were designed for visually impaired people to provide safe navigation. Most of the proposed products are based on obstacle detection without according attention to their nature. Nevertheless, the obstacle brand and especially its material is of a great importance for the information of the visually impaired person in his navigation. This paper presents a new solution based on only one ultrasonic sensor to recognize obstacle material and precise its rigidity. This kind of information can help in efficient obstacle recognition based on ultrasonic technology. The material rigidity is divided into two classes namely hard and soft class. We propose in the present paper a feature based model of each of these two classes. The proposed material rigidity recognition approach is then based on the analysis of spatial-frequency representation of ultrasonic wave, using Haralick’s texture features. Feature selection step is considered to select the most discriminating features. The data-mining tool WEKA has been used to evaluate the performance of our proposed system. The obtained results show the efficiency of the proposed method to recognize the two obstacle material rigidity classes. An authentic Iot dataset, is achieved, using a unique ultrasonic sensor, since no previous work has considered such a challenge and so no dataset of ultrasonic signal, acquired for material rigidity recognition, occurs in literature.