The Design And Realization Of The Blind Aided System Based On Tactile Vibration Signal

Assistive technology for the visually impaired and the blind is an important area of research.Due to the strong interest of researchers from different disciplines,research in this area has received increasing attention.Solving the problem of blind travel has a very important social impact on our growing ageing and blind population.Most blind-assistance technologies are used to help them navigate and avoid obstacles.Their focus is to help blind people identify objects and use language or beeps to convey information to users.However,few studies have transformed ground information into user's intuitive haptics.Tactile sensation is the most sensitive way for blind people to replace vision.Therefore,a blind assistance system based on tactile vibration signals came into being.Discusses the development of vibration technology closely related to this article and the blind auxiliary equipment based on the vibration signal,analyzes the scientific research progress in the field of blind assistance,based on the analysis and research,it proposes to collect the ground image and the collected image generation network to generate the corresponding spectrum The overall design idea of feeling vibration through contact with the vibrator,and designing a blind assistance system based on tactile vibration signals.The algorithm principle,training obstacles and methods for solving training obstacles of adversarial generative networks are analyzed.Based on the analysis results,a generative network using DCGAN as the basic framework of the network is designed.In the training of the network,two sets of generators and A discriminator,and an auxiliary classification module of ACGAN is added to the discriminator,and an end-to-end generation network of a ground image-vibration spectrum with a dual discriminator is established.Considering the interaction between the user and the machine,two sets of blind assistance devices based on the tactile vibration signals were designed and manufactured based on the Hikey970 chip guide stick and the Android mobile phone-based handheld system.The designed device can activate the image acquisition system for camera work.The acquired images are trained under the adversarial generation network architecture.The spectrum is converted into WAV format audio signals by Griffin operation.The audio signals are transmitted to a power amplifier for amplification.The amplified audio The signal is converted into a vibration signal by the vibrator.In order to test the performance of the guide stick and the Android-based handheld system,the two systems were tested for the authenticity and accuracy of vibration signals and real-world simulation experiments.The results show that the vibration signals generated by the two systems are sufficiently distinguishable that the user can identify the ground type.And the participants can quickly identify the category of the ground they are in by less than 3 attempts in the real scene.