DESIGN AND RAPID PROTOTYPING FOR COST-EFFECTIVE SLAM-BASED SEMI-AUTONOMOUS GROUND VEHICLES

Abstract

This thesis explores the integration of additive manufacturing and LiDAR-based navigation to manufacture a semi-autonomous ground vehicle (SAGV) capable of operating in diverse environments. The research scope aims at scalable and cost-effective manufacturing of a SAGV product. By utilizing additive manufacturing, specifically Fused Deposition Modeling (FDM), critical structural components such as mounts, chassis modifications, and protective casings were rapidly prototyped. Thermoplastic polyurethanes were evaluated for their high-impact durability and stress-fatigue performance, ensuring the vehicle’s reliability during testing and use in high-impact scenarios.LiDAR was integrated into a compact ground vehicle platform. Through ROS (Robot Operating System), LiDAR data was processed in real time for obstacle detection, path planning, and Simultaneous Localization and Mapping (SLAM). The autonomy stack incorporated algorithms such as A* and Dijkstra for pathfinding, in addition to real-time collision-avoidance strategies. Extensive testing involved both simulation-based trials, allowing for rapid iteration on software parameters, and physical experiments that validated hardware resilience and system responsiveness. Testing showed that additive manufacturing reduced design lead time and supported rapid redesign of structural parts, while the LiDAR stack provided repeatable obstacle detection and map generation in GPS-denied trials. The thesis identifies possible applications in logistics, sports training, and reconnaissance, and discusses how the platform could be adapted to each use case. Future work will include exploring additional sensors, advanced SLAM capabilities, and 3D point cloud generation.