BioDKM: Bio-inspired domain knowledge modeling method for humanoid delivery robots’ planning |
| |
Authors: | Wanpeng Zhang Tianjiang HuJing Chen Lincheng Shen |
| |
Affiliation: | College of Mechatronic Engineering and Automation, National University of Defense Technology, Changsha, Hunan 410073, PR China |
| |
Abstract: | A bio-inspired human domain knowledge modeling method, BioDKM, is proposed and developed to make delivery robots think more humanly and act more effectively. This presented method focused on feasible fusion between artificial intelligent and bionics in the field of tasks planning or scheduling in delivery robots. BioDKM is designed and implemented with several components, in terms of human knowledge, workflow (WF), hierarchical task network (HTN), and planner. In detail, WF is utilized as the human domain knowledge modeling tool, because of its convenient applications, friendly user interface and explicit representation. Moreover, WF can effectively complement conventional HTN planning with great convenience to formalize human domain knowledge. Translation from WF to HTN is also considered and established to make task planning smooth. Finally, examples and simulations are carried out to validate the effectiveness of this proposed bio-inspired domain knowledge modeling method. |
| |
Keywords: | Humanoid delivery robot BioDKM Workflow (WF) Hierarchical task network (HTN) planning |
本文献已被 ScienceDirect 等数据库收录! |
|