Terrain-adaptive motion planning for autonomous mobile cranes with load sway suppression
Year: 2026
Type: article
Abstract: Mobile cranes are essential construction equipment for lifting and transporting materials. However, their robotisation presents significant challenges due to uneven terrain and complex payload dynamics during walking operations. This paper presents a terrain-adaptive motion planner for autonomous mobile cranes (AMCs) that navigates construction site topography whilst suppressing load sway. It integrates a vehicle posture estimator and an analytical load sway prediction model, implemented on a GPU-based parallel computation architecture. Validation was conducted through high-fidelity simulations in NVIDIA Isaac Sim. The posture estimation module demonstrated accurate predictions across elevation grid maps of varying resolutions, whilst the sway prediction model achieved accurate forecasts up to 3 s. The motion planner successfully navigated varying terrain and dynamic obstacles while maintaining sway angles below the safety threshold and achieving real-time performance. The results demonstrate the feasibility of integrating terrain adaptation with sway suppression for AMC operations in unstructured construction sites. • Developed terrain-adaptive motion planner for autonomous mobile cranes. • Integrated load sway suppression into motion planning framework. • Proposed predictive load sway model for mobile cranes on uneven terrain. • Implemented GPU-accelerated motion planning for enhanced efficiency. • Validated and investigated the motion planner with static and dynamic obstacles.
Source: Automation in Construction
Authors: Zhuomin Zhou, Yu Bai, Elahe Abdi
Institutions: Monash University, University of Hong Kong
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $5,490
Funder: Monash University
OpenAlex ID: W7154584773