Citation: Wang, J.; Ruan, X.; Huang, J.
HDPP: High-Dimensional Dynamic
Path Planning Based on Multi-Scale
Positioning and Waypoint Refinement.
Appl. Sci. 2022, 12, 4695. https://
doi.org/10.3390/app12094695
Academic Editor: Alessandro
Gasparetto
Received: 11 February 2022
Accepted: 27 April 2022
Published: 6 May 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
HDPP: High-Dimensional Dynamic Path Planning Based on
Multi-Scale Positioning and Waypoint Refinement
Jingyao Wang, Xiaogang Ruan and Jing Huang *
Beijing Key Laboratory of Computational Intelligence and Intelligence System, Faculty of Information Technology,
Beijing University of Technology, Beijing, 100124, China; jingyao_wang0728@163.com (J.W.); adrxg@bjut.edu.cn (X.R.)
* Correspondence: huangjing@bjut.edu.cn
Abstract:
Algorithms such as RRT (Rapidly exploring random tree), A* and their variants have been
widely used in the field of robot path planning. A lot of work has shown that these detectors are unable
to carry out effective and stable results for moving objects in high-dimensional space, which generate
a large number of multi-dimensional corner points. Although some filtering mechanisms (such as
splines and valuation functions) reduce the calculation scale, the chance of collision is increased,
which is fatal to robots. In order to generate fewer but more effective and stable feature points, we
propose a novel multi-scale positioning method to plan the motion of the high-dimensional target.
First, a multi-scale feature extraction and refinement scheme for waypoint navigation and positioning
is proposed to find the corner points that are more important to the planning, and gradually eliminate
the unnecessary redundant points. Then, in order to obtain a stable planning effect, we balance the
gradient of corner point classification detection to avoid over-optimizing some of them during the
training phase. In addition, considering the maintenance cost of the robot in actual operation, we pay
attention to the mechanism of anti-collision in the model design. Our approach can achieve a complete
obstacle avoidance rate for high-dimensional space simulation and physical manipulators, and also
work well in low-dimensional space for path planning. The experimental results demonstrate the
superiority of our approach through a comparison with state-of-the-art models.
Keywords:
path planning; high-dimensional; machine learning; multi-scale positioning; waypoint
refinement; ROS
1. Introduction
Path planning is a basic task of computer vision and robot kinematics. It is necessary
to predict the target position and gait at the same time. With the great development of
deep learning, path planning is successfully integrated into more and more real-world
application systems, such as autonomous driving, human-machine collaboration, and
remote control [
1
,
2
]. Among them, motion planning in high-dimensional space becomes a
hot topic, which is a key for realizing the slogan of “machine serving society” (for example,
palletizing in delivery warehouses, underwater exploration, aerial cruise) [
3
,
4
]. However,
today’s high-dimensional spatial planning can only accomplish simple motions in specific
environments with few obstacles, and this planning process is lengthy. Hence, it is necessary
to explore an efficient path planning approach, especially for high-dimensional space, to
fill the gap, which is a prominent and interdisciplinary research area.
However, it is a challenging task to deal with high-dimensional data for path planning.
Firstly, most current works are rarely studied in high-dimensional environments, and the
vulnerability of general path detectors has been revealed by planning in more complex
environments [
5
,
6
]. For example, although the Walk To algorithm [
7
] can be directly used in
any dimension, it is completely unable to deal with obstacles and other situations. Bugs [
8
]
need to rotate clockwise (or counterclockwise) around the obstacle, and it is difficult to
define the direction in high-dimensional space. The ant colony algorithm [
9
] will bring
Appl. Sci. 2022, 12, 4695. https://doi.org/10.3390/app12094695 https://www.mdpi.com/journal/applsci
