# Research on a Visually Assisted Efficient Blind-Guiding System and an Autonomous Shopping Guidance Robot Arm Adapted to the Complex Environment of Farmers’ Markets

## TL;DR

This paper introduces a robot system to help visually impaired people shop in crowded farmers' markets by combining visual and radio-frequency navigation with a robotic arm.

## Contribution

The novel RFTPAD and A*-FRN algorithms improve navigation accuracy and efficiency, while the IRAGS algorithm enhances product selection for visually impaired individuals.

## Key findings

- RFTPAD improves mapping accuracy by 23.9% compared to classic methods.
- A*-FRN reduces driving trajectory length by 23.3% compared to general navigation.
- The robotic arm increases product selection efficiency by 100% compared to manual searching.

## Abstract

It is great challenge for visually impaired (VI) people to shop in narrow and crowded farmers’ markets. However, there is no research related to guiding them in farmers’ markets worldwide. This paper proposes the Radio-Frequency–Visual Tag Positioning and Automatic Detection (RFTPAD) algorithm to quickly build a high-precision navigation map. It combines the advantages of visual beacons and radio-frequency signal beacons to accurately calculate the guide robot’s coordinates to correct its positioning error and simultaneously perform the task of mapping and detecting information. Furthermore, this paper proposes the A*-Fixed-Route Navigation (A*-FRN) algorithm, which controls the robot to navigate along fixed routes and prevents it from making frequent detours in crowded aisles. Finally, this study equips the guide robot with a flexible robotic arm and proposes the Intelligent-Robotic-Arm-Guided Shopping (IRAGS) algorithm to guide VI people to quickly select fresh products or guide merchants to pack and weigh products. Multiple experiments conducted in a 1600 m2 market demonstrate that compared with the classic mapping method, the accuracy of RFTPAD is improved by 23.9%. What is more, compared with the general navigation method, the driving trajectory length of A*-FRN is 23.3% less. Furthermore, the efficiency of guiding VI people to select products by a robotic arm is 100% higher than that through a finger to search and touch.