A Portable Intelligent Navigation System Based on Embedded Hybrid Computing for Visually Impaired People

This paper presents a portable intelligent navigation system for visually impaired persons that integrates environmental perception, voice interaction, and real‐time navigation, developed on an STM32 multicore embedded computing platform. The designed system acquires environmental information and users' commands via integrated modules, for example, camera, GPS, and microphone. Due to the limited data‐processing capabilities of the STM32 platform, lightweight tasks are executed locally. In contrast, computation‐intensive tasks, for example, image processing and path planning, are sent to the cloud for processing to enhance the system's real‐time responsiveness. Experimental results demonstrate that the designed system can effectively guide visually impaired people to the desired destinations, which can achieve an accuracy of up to in traffic element recognition, for example, traffic lights and crosswalks, an average response time of less than 5 s, and a navigation accuracy of m.