Cornell University
Decentralized coordination of multi-robot networks for active target tracking
Abstract
dc:description.abstractThis paper introduces a decentralized framework for optimizing the coordination of robot networks to track multiple moving targets in applications like security and surveillance. The problem of network optimization is proven to be NP-hard, highlighting the need for efficient solutions. The proposed frame- work presents two novel decentralized coordination methods: the group-based algorithm and the bundle-based algorithm. These methods aim to achieve adaptive and conflict free target assignments, with the bundle-based algorithm providing more effective coordination and guaranteeing a 12 approximation in the worst-case scenario. Simulation results demonstrate that the proposed approaches outperform existing algorithms, achieving performance close to the optimal solution in significantly less time. Compared to EER control and PD control, the group-based assignment and control (GBAC) and the bundle-based assignment and control (BBAC) demonstrate superior performance due to their adaptive target assignments achieved through network coordination. Among the two methods, BBAC shows higher average target tracking rate (ATTR) and lower average robot traveling distance (ARTD), resulting in improved track- ing efficiency. Physical experiments using a network of ground robots tracking human targets further validate the practicality of the proposed approach in real- world scenarios.
Degree
thesis:*- Name thesis:degree_name
- M.S., Mechanical Engineering
- Level thesis:degree_level
- Master of Science
- Discipline thesis:degree_discipline
- Mechanical Engineering
- Grantor
- Cornell University
- Year dc:date.issued
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Jang, Kyung Rak
- Chair dc:contributor.chair
-
- Ferrari, Silvia
- Committee member dc:contributor.committeemember
-
- Ju, Wendy
Rights
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Bibid
- 16219150
- Dc Identifier Other
-
ProQuest Submission ID: 11845
ProQuest Publication ID: 30569077