Compact Conditional Joint Decision and Estimation for Joint Tracking and Identification With Performance Evaluation

• Main Authors: Wen Cao, Meng Hui, Lin Bai, Bobin Yao
• Format: Article
• Language: English
• Published: IEEE 2018-01-01
• Series: IEEE Access
• Subjects: Joint tracking and identification; compact conditional joint decision and estimation; joint performance evaluation
• Online Access: https://ieeexplore.ieee.org/document/8240601/

This paper presents a novel approach for the joint tracking and identification (JTI) problem. JTI involves interdependent tracking and identification, and thus solving them jointly is preferable. The recently proposed joint decision and estimation (JDE) framework provides a good solution for such problems involving coupled decision and estimation. To solve the JTI problem, this paper proposes a compact conditional JDE (CCJDE) method within the JDE framework. First, we propose a new CCJDE risk, which unifies the traditional decision and estimation risks in a concise form. Based on this, we present the optimal joint solution with analytical form. Second, inspired by the interacted parameters in CCJDE, we propose a new CCJDE scheme with time-varying parameters, which further utilizes the interdependence between decision and estimation. Third, we apply CCJDE to practical JTI problems. An applicable multiple model CCJDE algorithm is proposed for JTI. For performance evaluation, we propose a new joint performance metric (JPM), which unifies the tracking error and the identification error. Finally, two illustrative examples verify the superiority of the proposed CCJDE method. CCJDE outperforms the traditional two-step strategies in JPM. For multisensor data JTI, however, CCJDE can further utilize all information from heterogeneous sensor data. Besides, the effectiveness of the proposed JPM and the time-varying parameters in CCJDE are also demonstrated.