TY - JOUR
T1 - An Adaptive Online Co-Search Method with Distributed Samples for Dynamic Target Tracking
AU - Li, Feng
AU - Zhou, Mengchu
AU - Ding, Yongsheng
N1 - Funding Information:
This work was supported in part by the Key Project of the National Natural Science Foundation of China under Grant 61134009, in part by the National Natural Science Foundation of China under Grant 61473078, and Grant 61473077.
Funding Information:
Manuscript received November 2, 2016; revised January 11, 2017; accepted January 30, 2017. Date of publication March 1, 2017; date of current version February 8, 2018. Manuscript received in final form February 8, 2017. This work was supported in part by the Key Project of the National Natural Science Foundation of China under Grant 61134009, in part by the National Natural Science Foundation of China under Grant 61473078, and Grant 61473077, in part by Cooperative research funds of the National Natural Science Funds Overseas and Hong Kong and Macao scholars under Grant 61428302, in part by the Program for Changjiang Scholars from the Ministry of Education (2015-2019), and in part by the International Collaborative Project of the Shanghai Committee of Science and Technology under Grant 16510711100. Recommended by Associate Editor Huijun Gao. (Corresponding author: Yongsheng Ding.) F. Li and Y. Ding are with the Engineering Research Center of Digitized Textile and Apparel Technology, Ministry of Education, and also with the College of Information Science and Technology, Donghua University, Shanghai 201620, China (e-mail: ysding@dhu.edu.cn).
Publisher Copyright:
© 2017 IEEE.
PY - 2018/3
Y1 - 2018/3
N2 - Dynamic optimal problems (DOPs) are often encountered in target search, emergency rescue, and object tracking. Motivated by the need to perform a search and rescue task, we clarify a DOP in a complex environment if a target unpredictably travels in an environment with general non-Gaussian distributed and time-varying noises. To solve this issue, we propose a recursive Bayesian estimation with a distributed sampling (RBEDS) model. Furthermore, two kinds of communication cooperative extensions, i.e., real-time communication and communication after finding the target, are analyzed. To balance between exploitation and exploration, an adaptive online co-search (AOCS) method, which consists of an online updating algorithm and a self-adaptive controller, is designed based on RBEDS. Simulation results demonstrates that searchers with AOCS can achieve a comparable search performance with a global sampling method, e.g., Markov Chain Monto Carlo estimation, by applying real-time communication. The local samples help searchers keep flexible and adaptive to the changes of the target. The proposed method with both communication and cooperation exhibits excellent performance when tracking a target. Another attractive result is that only a few searchers and local samples are demanded. The insensibility to the scale of samples makes the proposed method obtain a better solution with less computation cost than the existing methods.
AB - Dynamic optimal problems (DOPs) are often encountered in target search, emergency rescue, and object tracking. Motivated by the need to perform a search and rescue task, we clarify a DOP in a complex environment if a target unpredictably travels in an environment with general non-Gaussian distributed and time-varying noises. To solve this issue, we propose a recursive Bayesian estimation with a distributed sampling (RBEDS) model. Furthermore, two kinds of communication cooperative extensions, i.e., real-time communication and communication after finding the target, are analyzed. To balance between exploitation and exploration, an adaptive online co-search (AOCS) method, which consists of an online updating algorithm and a self-adaptive controller, is designed based on RBEDS. Simulation results demonstrates that searchers with AOCS can achieve a comparable search performance with a global sampling method, e.g., Markov Chain Monto Carlo estimation, by applying real-time communication. The local samples help searchers keep flexible and adaptive to the changes of the target. The proposed method with both communication and cooperation exhibits excellent performance when tracking a target. Another attractive result is that only a few searchers and local samples are demanded. The insensibility to the scale of samples makes the proposed method obtain a better solution with less computation cost than the existing methods.
KW - Dynamic target tracking
KW - distributed sample
KW - multiagent
KW - target search
UR - http://www.scopus.com/inward/record.url?scp=85016515651&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85016515651&partnerID=8YFLogxK
U2 - 10.1109/TCST.2017.2669154
DO - 10.1109/TCST.2017.2669154
M3 - Article
AN - SCOPUS:85016515651
SN - 1063-6536
VL - 26
SP - 439
EP - 451
JO - IEEE Transactions on Control Systems Technology
JF - IEEE Transactions on Control Systems Technology
IS - 2
M1 - 7867053
ER -