ADAPT: An Autonomous Forklift for Construction Site Operation

TL;DR

The paper addresses the inefficiencies and safety risks of construction-site material handling by introducing ADAPT, an autonomous off-road forklift designed for unstructured environments. It combines a factor-graph based joint localization and pallet mapping with depth-based pallet pose estimation, a 2.5D traversability map, perception and planning/control pipelines, and a Hybrid A* navigation strategy with visual-servoing docking. A novel fork contact measurement using pressure feedback enhances manipulation robustness, and extensive real-world testing against a seasoned operator demonstrates near-human performance with minimal intervention. The work advances autonomous construction logistics by enabling safer, more efficient material flow on sites lacking infrastructure, and it outlines future work on dynamic planning, semantic mapping, and hardware upgrades to further close the gap to fully autonomous operation.

Abstract

Efficient material logistics play a critical role in controlling costs and schedules in the construction industry. However, manual material handling remains prone to inefficiencies, delays, and safety risks. Autonomous forklifts offer a promising solution to streamline on-site logistics, reducing reliance on human operators and mitigating labor shortages. This paper presents the development and evaluation of ADAPT (Autonomous Dynamic All-terrain Pallet Transporter), a fully autonomous off-road forklift designed for construction environments. Unlike structured warehouse settings, construction sites pose significant challenges, including dynamic obstacles, unstructured terrain, and varying weather conditions. To address these challenges, our system integrates AI-driven perception techniques with traditional approaches for decision making, planning, and control, enabling reliable operation in complex environments. We validate the system through extensive real-world testing, comparing its continuous performance against an experienced human operator across various weather conditions. Our findings demonstrate that autonomous outdoor forklifts can operate near human-level performance, offering a viable path toward safer and more efficient construction logistics.

Figures (31)

• Figure 1: The autonomous off-road forklift ADAPT (Autonomous Dynamic All-terrain Pallet Transporter) in its designated working environment --- an unstructured construction site. Image credit: © AIT/tm-photography.
• Figure 2: System architecture of the proposed autonomous forklift, illustrating the hierarchical organization of key components essential for autonomous operation. The diagram depicts high-level information pathways and data flow between integrated sensors and specialized processing modules throughout the system. Relevant section references are provided alongside each major component, directing readers to corresponding detailed descriptions within the manuscript.
• Figure 3: Key hardware components of ADAPT. (a) 3D central joint for maneuverability and structural flexibility, (b) Object detection system, (c) Obstacle avoidance and terrain mapping sensor, and (d) Close-range object detection for precise load handling.
• Figure 4: Actuated and unactuated(*) joints for vehicle base movement and fork positioning.
• Figure 5: Overview of the main hardware components, including core processing devices, actuators, as well as proprioceptive and exteroceptive sensors.