Article Per Year (5 Year)
p-Index From 2021 - 2026
0.408
P-Index
This Author published in this journals
All Journal International Journal of Robotics and Control Systems
Sarath Kodagoda
University of Technology Sydney
Control & Systems Engineering Electrical & Electronics Engineering

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search
Journal : International Journal of Robotics and Control Systems

 1 
Monte Carlo Simulations on 2D LRF Based People Tracking using Interactive Multiple Model Probabilistic Data Association Filter Tracker Zulkarnain Zainudin; Sarath Kodagoda
International Journal of Robotics and Control Systems Vol 3, No 1 (2023)
Publisher : Association for Scientific Computing Electronics and Engineering (ASCEE)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31763/ijrcs.v3i1.896

Abstract

Consistency of tracking filter such as Interactive Multiple Model Probabilistic Data Association Filter (IMMPDAF) is the most important factor in targets tracking. Inaccurate tracking capability will lead to poor tracking performance when dealing with multiple people's interactions and occlusions. In order to validate the consistency, Normalized Estimation Error Squared (NEES) and Normalized Innovation Squared (NIS) were evaluated and tested using Monte Carlo experiments for 50 runs. These simulations has proven that the tracker is conditionally consistent on targets tracking despite the fact that it has difficulties on handling occlusions and maneuvering people. NEES requires ground truth of tracking data and predicted data, whereas NIS requires observation and predicted data for Monte Carlo simulations. In NEES simulations, the result emphasizes that state estimation errors of IMMPDAF tracker are inconsistent with filter-calculated covariances especially when dealing with sudden turns in zig-zag motion where quite a large number of points fall outside 95\% probability region. In NIS simulations, IMMPDAF tracker is confirmed to have difficulties to handle multiple targets with a short period of occlusion although a small number of points falls outside of 95\% probability region. Filter tracker is considered mismatched when dealing with zig-zag motion; however, it deemed to be optimistic when dealing with occlusions. As a result, the IMMPDAF tracker has limited capability in monitoring sharp turns under occlusion conditions, although it is acceptable when dealing with occlusions only.
Gaussian Processes-BayesFilters with Non-Parametric Data Optimization for Efficient 2D LiDAR Based People Tracking Zulkarnain Zainudin; Sarath Kodagoda
International Journal of Robotics and Control Systems Vol 3, No 2 (2023)
Publisher : Association for Scientific Computing Electronics and Engineering (ASCEE)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31763/ijrcs.v3i2.901

Abstract

A model for expressing and describing human motion patterns must be able to improve tracking accuracy. However, Conventional Bayesian Filters such as Kalman Filter (KF) and Particle Filter (PF) are vulnerable to failure when dealing with highly maneuverable targets and long-term occlusions. Gaussian Processes (GP) is then used to adapt human motion patterns and integrate the model with Bayesian Filters. In GP, all samples in training phase need to be included and periodically, new samples will be added into training samples whenever it is available. Larger amount of data will increase the computational time to produce the learned GP models due to data redundancies. As a result, Mutual Information (MI) based technique with Mahalanobis Distance (MD) is developed to keep only the informative data. This method is used to process data which is collected by a robot equipped with a LiDAR. Experiments have demonstrated that reducing data does not raise Average Root Mean Square Error (ARMSE) considerably. EKF, PF, GP-EKF and GP-PF are utilised as a tool for tracking people and all techniques have been analyzed in order to distinguish which method is more efficient. The performance of GP-EKF and GP-PF are then compared to EKF and PF where it proved that GP-BayesFilters performs better than Conventional Bayesian Filters. The proposed approach has reduced data points up to more than 90\% while keeping the ARMSE within acceptable limits. This data optimization technique will save computational time especially when deal with periodically accumulative data sets. Comparing on four tracking methods, both GP-PF and GP-EKF have achieved higher tracking performance when dealing with  highly maneuverable targets and occlusions.
Co-Authors Zulkarnain Zainudin Zulkarnain Zainudin