Affiliations: [a] Engineering Department, Melco Power Systems Co., Ltd. Hyogo-ku, Kobe, Japan | [b] Graduate School of Engineering, Tottori University, Tottori-city, Tottori, Japan
Abstract: We have been proposing the automatic human tracking system to overcome limitations of existing video surveillance systems. This paper proposes two bypass methods to keep tracking a targeted person by the system constructed by using mobile agent technologies even if unexpected problems are occurred. The proposed methods are called as “Recalculation Bypass Method” and “Additional Calculation Bypass Method”. These bypass methods utilize the algorithm which is elaborated in the paper, “An Algorithm to Determine Neighbor Nodes” [7]. The mobile agents are operating in automatic human tracking system and the bypass methods are utilized by the mobile agents. It is a very important ability for the mobile agent not to lose tracking in automatic human tracking systems. The bypass methods contribute to the realization of robust system and resolve two classifications of problem to ensure the tracking ability of mobile agents. The two classifications of problem are logical and physical problems. In the case of logical problems, they occur inside the system and they are called as “broken server”. While in the case of physical problems, they occur outside the system and they are called as “pathway being modified”. This system is able to track the targeted person continuously by utilizing the methods even if unexpected problems are occurred. We confirmed the effectiveness of the proposed bypass methods by the experiments that used more than one broken servers and pathway being modified.
