3D tactile mapping robotic system and relief generation for visual impairment

Abstract

This paper presents the development of an autonomous mobile robot designed to generate indoor maps (Occupancy Grids) and automatically convert them into three-dimensional models (.STL) with tactile relief. The primary goal is to provide visually impaired individuals with a haptic spatial pre-orientation tool, enabling them to understand the architectural layout of an environment prior to navigation. The system implements a distributed hardware architecture where an NVIDIA Jetson Orin Nano processes SLAM and computer vision algorithms under ROS 2 (Jazzy Jalisco), while a Raspberry Pi Pico microcontroller handles the low-level control of a LiDAR sensor. This paper details the mechanical design, the implementation of simulated odometry nodes in C++ with native hardware access via libgpiod, and the creation of an offline web application for live map rendering without external software dependencies. Finally, the 3D extrusion algorithm based on morphological dilation (OpenCV) and the Marching Cubes method is described, validating the design under Universal Design principles.