Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

EdgeAI Drone for Autonomous Construction Site Demonstrator

Emre Girgin, Arda Taha Candan, Coşkun Anıl Zaman

TL;DR

This work tackles the challenge of deploying robust, low-power AI on edge devices for autonomous construction-site operations. It introduces a UAV-based demonstrator that performs MCU-edge object detection, communicates via a 5G-enabled central coordinator, and coordinates multiple ground and aerial robots through ROS-based architecture and MQTT-based messaging. A novel dataset, the TUBITAK-EdgeDrone, supports MCU-friendly training, while a monocular depth-estimation pipeline leverages altitude and image scale with RANSAC to produce global obstacle coordinates. Field experiments validate real-time performance, demonstrate scalability advantages over conventional UAV setups, and highlight edge-AI's practical potential for safer, more efficient civil automation.

Abstract

The fields of autonomous systems and robotics are receiving considerable attention in civil applications such as construction, logistics, and firefighting. Nevertheless, the widespread adoption of these technologies is hindered by the necessity for robust processing units to run AI models. Edge-AI solutions offer considerable promise, enabling low-power, cost-effective robotics that can automate civil services, improve safety, and enhance sustainability. This paper presents a novel Edge-AI-enabled drone-based surveillance system for autonomous multi-robot operations at construction sites. Our system integrates a lightweight MCU-based object detection model within a custom-built UAV platform and a 5G-enabled multi-agent coordination infrastructure. We specifically target the real-time obstacle detection and dynamic path planning problem in construction environments, providing a comprehensive dataset specifically created for MCU-based edge applications. Field experiments demonstrate practical viability and identify optimal operational parameters, highlighting our approach's scalability and computational efficiency advantages compared to existing UAV solutions. The present and future roles of autonomous vehicles on construction sites are also discussed, as well as the effectiveness of edge-AI solutions. We share our dataset publicly at github.com/egirgin/storaige-b950

EdgeAI Drone for Autonomous Construction Site Demonstrator

TL;DR

This work tackles the challenge of deploying robust, low-power AI on edge devices for autonomous construction-site operations. It introduces a UAV-based demonstrator that performs MCU-edge object detection, communicates via a 5G-enabled central coordinator, and coordinates multiple ground and aerial robots through ROS-based architecture and MQTT-based messaging. A novel dataset, the TUBITAK-EdgeDrone, supports MCU-friendly training, while a monocular depth-estimation pipeline leverages altitude and image scale with RANSAC to produce global obstacle coordinates. Field experiments validate real-time performance, demonstrate scalability advantages over conventional UAV setups, and highlight edge-AI's practical potential for safer, more efficient civil automation.

Abstract

The fields of autonomous systems and robotics are receiving considerable attention in civil applications such as construction, logistics, and firefighting. Nevertheless, the widespread adoption of these technologies is hindered by the necessity for robust processing units to run AI models. Edge-AI solutions offer considerable promise, enabling low-power, cost-effective robotics that can automate civil services, improve safety, and enhance sustainability. This paper presents a novel Edge-AI-enabled drone-based surveillance system for autonomous multi-robot operations at construction sites. Our system integrates a lightweight MCU-based object detection model within a custom-built UAV platform and a 5G-enabled multi-agent coordination infrastructure. We specifically target the real-time obstacle detection and dynamic path planning problem in construction environments, providing a comprehensive dataset specifically created for MCU-based edge applications. Field experiments demonstrate practical viability and identify optimal operational parameters, highlighting our approach's scalability and computational efficiency advantages compared to existing UAV solutions. The present and future roles of autonomous vehicles on construction sites are also discussed, as well as the effectiveness of edge-AI solutions. We share our dataset publicly at github.com/egirgin/storaige-b950
Paper Structure (11 sections, 4 figures, 2 tables)

This paper contains 11 sections, 4 figures, 2 tables.

Table of Contents

  1. INTRODUCTION
  2. Related Work
  3. Methodology
  4. Architecture
  5. Central Coordinator Software
  6. Autonomous Drone Design
  7. AI on the Edge and Data Collection
  8. Experiments
  9. Object Detection and Pose Estimation
  10. Field Tests
  11. Conclusion Acknowledgement

Figures (4)

  • Figure 1: Our use-case demonstrates of autonomous loading in a construction site environment where the entire operation is monitored by edge-AI equipped UAV.
  • Figure 2: The overall architecture of the demonstrator is summarized. While the GCS coordinates all the vehicles, communication is facilitated by a 5G mobile base infrastructure. The UAV also contains an Edge-AI subsystem, which is responsible for detecting obstacles in the operational field.
  • Figure 3: The CAD drawing of the hexacopter, including the AI subsystem. The carbon-fiber layers situated between the landing gear houses the batteries, RPI, MCU board, 5G router, and camera.
  • Figure 4: OpenAirLab is an autonomous construction test site in Gebze, Turkey, comprising an area of 50 meters by 100 meters.