多代理人智慧家庭系統之設計與實作

The design and implementation for multiagent based smart home systems

曾溫祥(Wen-Shang Tseng)

指導教授 ： 陳文輝

國立臺北科技大學/機電學院/自動化科技研究所/碩士(2005年)

https://doi.org/10.6841/NTUT.2005.00176

摘要

本研究旨在提出一個以多代理人為基礎，以家庭網路為代理人平台，組成一個智慧家庭系統。此系統包含有兩類代理人，分別為一般化智慧家電代理人(GIANT)和房間代理人，前者用來管理家電，後者用來管理GIANT。在設計上，結合有SCP控制網路協定的家電，可以在低速電力線通訊網路中相互通訊，另外，經由SCP-UPnP橋接器，使SCP家電可以成為UPnP網路成員，進而形成家庭網路。GIANT藉由灰色理論、模糊理論的演算，將原始資料轉換成使用者行為模式，並透過倒傳遞類神經網路，可以習得使用者行為模式，和提供主動協助使用者開啟或關閉家電的能力。此外，系統採用FIPA- ACL代理人通訊語言，作為代理人資訊交換的依據。經實測結果，可以確認家庭網路的可用性；經模擬結果，可以確認GIANT的主動協助能力和適應性能力。

關鍵字

多代理人 ； SCP ； UPnP ； 灰色理論 ； 模糊理論 ； 類神經網路

並列摘要

The thesis proposed a multiagent based smart home system by applying home networks as the agent platform. The multiagent system consists of several General Intelligent Appliance ageNTs (GIANTs) which manage home appliances and Room Agents (RAs) which manage GIANTs in the same room. In the design vision, a control network technology called Simple Control Protocol (SCP) was used to create SCP-enabled home appliances that can communicate with each other over low-speed power line communication networks. Through SCP-to-UPnP bridge, the SCP-enabled appliances can participate in Universal Plug and Play (UPnP) networks to form fully connected home networks. The agent, GIANT, can transfer raw data to behavior patterns of the user by using the proposed hybrid grey-fuzzy algorithm. GIANT can also learn the behavior patterns and proactively assist them to turn on/off home appliances via well-trained Back-Propagation Neural Networks. The proposed multiagent system also uses the FIPA-ACL as the agent communication language to provide agents with a mean of exchanging information. The experiential results show that the home networks are usable. GIANT is confirmed to provide the proactivity ability and adaptation from the simulation results.

並列關鍵字

Multiagent ； SCP ； UPnP ； Grey-Fuzzy Algorithm ； Neural Network

參考文獻

[5] S.R. Hedberg, “After desktop computing: a progress report on smart environments research,” Intelligent Systems and Their Applications, IEEE, vol. 15, issue 5, Sept.-Oct. 2000, pp. 7-9.

[12] S.K. Das, D.J. Cook, A. Bhattacharya, E.O. Heierman III and T.Y. Lin, “The role of prediction algorithms in the MavHome smart home architecture,” Wireless Communications, IEEE, vol. 9, issue 6, Dec 2002, pp. 77-84.

[13] D.J. Cook, M. Youngblood, E.O. Heierman III, K. Gopalratnam, S. Rao, A. Litvin and F. Khawaja , “MavHome: an agent-based smart home,” Proceedings of the 1th IEEE International Conference on Pervasive Computing and Communications, March 2003, pp. 521-524.

[16] “AIIA Lab,” Available : http://kaukoai.iis.sinica.edu.tw/.

[19] A. Pentland, “Healthwear: medical technology becomes wearable,” Computer, vol 37, issue 5, May 2004, pp. 42-49.