.png){kind=icon}
返回
Global path guided model predictive path integral control: Applications to GPU-parallelizable robot simulation systems
.png){kind=icon}
PRE
AI
DOI：10.1016/j.compeleceng.2024.109645.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
In this paper, a global path guided model predictive path integral (Global-MPPI) control algorithmic framework is proposed, which integrates the model predictive path integral (MPPI) control with global path planning algorithms. The method is designed to address the challenges of robot path planning and execution in large-scale spaces with obstacles. It is applicable in both 2D and 3D spaces and can be implemented for both mobile robots and manipulators. To enhance practicality, we present an obstacle size expansion method aimed at minimizing collision risks. Validation of the algorithm is conducted using the advanced parallel physics simulator, IsaacGym, highlighting the effectiveness of the method in a randomly populated obstacle space. Multiple global path planning algorithms have been tested using Global-MPPI framework and the key factors affecting the effectiveness of the algorithms are analyzed. Furthermore, Global-MPPI is compared with recent algorithms solving similar problems. It can substantially increase the success rate when the right parameters are set.
Keyword:
Global path planning
Model predictive path integral (MPPI) control
Collision avoidance
期刊
C
IF:
4.9
论文数:
6.7K
被引数:
1.3W
机构
引用论文
.png){kind=icon}
Information-Theoretic Model Predictive Control: Theory and Applications to Autonomous Driving信息论模型预测控制: 理论和在自动驾驶中的应用
OAAI
OAAI分享收藏分享收藏Research on obstacle avoidance of multi-AUV cluster formation based on virtual structure and artificial potential field method基于虚拟结构和人工势场法的多AUV集群编队避障研究
PREAI
分享收藏分享收藏
查看更多内容