Work

Hybrid A*-Guided Model Predictive Path Integral Control for Robust Navigation in Rough Terrains

Year: 2025

Type: article

Abstract: Navigating rough terrains requires a robust path planning algorithm that accounts for the physical properties of the environment to maintain stability and ensure safety. This article proposes the Hybrid A*-guided Model Predictive Path Integral (MPPI) algorithm augmented with traversability estimation to address the challenges of path planning on uneven terrains. The traversability estimation process quantifies surface characteristics, such as slope and roughness to identify traversable regions. Using this information, the Hybrid A* algorithm computes paths that minimize surface irregularities and prioritize regions with lower gradients, thereby enhancing stability and reducing dynamic disturbances. These computed paths are then used to define the mean control input for the MPPI algorithm, which performs localized optimization while adhering to the terrain-aware trajectory. By integrating terrain-aware guidance through the Hybrid A* algorithm with the MPPI, the proposed methodology automates the selection of the appropriate mean control input and enhances control performance by explicitly incorporating terrain properties into the planning process. Experimental results demonstrate the ability of the algorithm to navigate complex terrains with reduced roll and pitch motions, contributing to improved stability and performance.

Source: Mathematics

DOI: 10.3390/math13050810

Authors: Joonyeol Yang, Minhyeong Kang, Seulchan Lee, Sanghyun Kim

Institution: Kyung Hee University

Language: English

Topic: Robotic Path Planning Algorithms

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $3,137

Funder: National Research Foundation of Korea