Loading...
Search
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
Digital Object List
Bookmark
Review List- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 58712 (05)
- University: Sharif University of Technology
- Department: Electrical Engineering
- Advisor(s): Nayebi, Mohammad Mahdi; Sheikhi, Abbas
- Abstract:
- Target tracking is one of the key challenges in radar applications, including battlefield scenarios, autonomous vehicles, and localization, is. This problem becomes more complex when there are multiple targets, and the possibility of false alarms or clutter. The target tracking algorithm in radar is called TWS (Track While Scan), which is responsible for tracking and estimating the trajectory of each target based on the set of received observations. The assignment of observations to tracks and the estimation of each track’s parameters constitute the two main stages of this algorithm. The goal of the first stage is to assign the appropriate observation from among the available ones in each scan to the desired track. The most complex issue in this stage is how to establish a correlation between tracks and nearby observations. Algorithms such as Global Nearest Neighbor (GNN), Joint Probabilistic Data Association (JPDA), and Multiple Hypothesis Tracking (MHT) are among the common assignment methods. In the second stage, filtering and estimating the parameters of each track enables the removal of measurement noise and the prediction of each target’s position in the next scan. This is done using filters such as the Kalman Filter (KF), Interacting Multiple Model (IMM), and Particle Filter (PF). In this research, an effort is made to develop an observation-to-target assignment algorithm using deep learning-based methods, and its performance is compared with other existing methods
- Keywords:
- Target Tracking Problem ; Association ; Estimation ; Deep Learning ; Radars ; Radar Target Recognition
-
محتواي کتاب
- view