.png){kind=icon}
.png){kind=icon}
Return
A Personalized Social Knowledge Base Framework for Self-Learning Mobile Service Robots
.png){kind=icon}
PRE
AI
Y
Q
R
M
Y
D
W
S
L
DOI：10.1007/s12369-026-01408-9.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Natural language instructions often exhibit seque-ntial constraints rather than being simply goal-oriented. With the development of artificial intelligence technology, service robots have been used in various scenarios for its advantages in enabling natural and smooth human-robot interaction. In this work, the proposed knowledge base framework is composed of five modules: basic knowledge base, knowledge inference module, knowledge query module, knowledge analysis module and knowledge update module. Firstly, a basic standard knowledge base is established according to the statistics of human conventions, and the knowledge inference module is used to reason about the relationships between nodes and the properties of nodes. Then, the robot queries rules for social behaviors through the knowledge query module and generates social behaviors through the knowledge analysis module. Finally, the knowledge base is updated by the knowledge update module according to the user feedback. This enables the robot to make social behaviors more in line with the preferences of social subjects during the next interaction. In order to better realize the inference of knowledge base nodes and edges, we propose clustering the person nodes in the knowledge base. We formulate question sentences in specific scenarios and collect dialogue information from characters. The dialogue information is encoded into sentence vectors based on the Bert model. At the same time, in order to improve the clustering effect and efficiency, we perform PCA dimensionality reduction processing on the sentence vectors, and then cluster the character nodes through K-means. Simulations and real-world experiments are carried out, and experimental results reveal that the proposed social knowledge base framework enables robots to generate personalized social behaviors, and compared with the results without the proposed framework, the satisfaction of users with the robot behaviors in this work is increased by 24%. In this study, we demonstrate this capability through a mobile robot equipped with a range camera, focusing on personalized distance and speed regulation as representative social behaviors.
Keywords:
Knowledge base
Service robots
Cluster
Personalized socializing
Journal
IF:
3.7
Papers:
1.4K
Citations:
5.6K
Organization
Cited Papers