Model-Based Control for Optical Image Stabilization
(2025)Department of Automatic Control
- Abstract
- Cameras are an ever growing part of all humans’ everyday life, from smartphones to surveillance cameras. One thing they all have in common is the need for a stable and focused image, even in conditions with camera movement as a result of external factors. To counteract the issue of blurry images from vibration disturbances, a technique called Optical Image Stabilization (OIS) is commonly used. In this thesis the possibility of a new more advanced control algorithm for the OIS lens was investigated. More specifically, the research questions ask whether it would be suitable to implement a Model Predictive Controller (MPC) as an alternative solution to the industry standard PID controller, which is the current solution for the system used in... (More)
- Cameras are an ever growing part of all humans’ everyday life, from smartphones to surveillance cameras. One thing they all have in common is the need for a stable and focused image, even in conditions with camera movement as a result of external factors. To counteract the issue of blurry images from vibration disturbances, a technique called Optical Image Stabilization (OIS) is commonly used. In this thesis the possibility of a new more advanced control algorithm for the OIS lens was investigated. More specifically, the research questions ask whether it would be suitable to implement a Model Predictive Controller (MPC) as an alternative solution to the industry standard PID controller, which is the current solution for the system used in this study. The performance of the two controllers has been analyzed to determine if the alternative solution outperforms the current one and the feasibility of the implementation has been evaluated. As the performance of an MPC is heavily reliant on the quality of the model used, a large part of the thesis is focused on system identification and model structures. While the MPC was implemented successfully, the results show that it could not outperform the PID. The performance of the MPC showed potential, but was impaired by the effects of cross-axis coupling. The inferior performance was likely because of insufficient accuracy of the models, and unmodeled dynamics of the system. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9212157
- author
- Lindmark, Albin and Philipson, Sofia
- supervisor
- organization
- year
- 2025
- type
- H3 - Professional qualifications (4 Years - )
- subject
- report number
- TFRT-6290
- other publication id
- 0280-5316
- language
- English
- id
- 9212157
- date added to LUP
- 2025-09-18 14:07:20
- date last changed
- 2025-09-18 14:07:20
@misc{9212157,
abstract = {{Cameras are an ever growing part of all humans’ everyday life, from smartphones to surveillance cameras. One thing they all have in common is the need for a stable and focused image, even in conditions with camera movement as a result of external factors. To counteract the issue of blurry images from vibration disturbances, a technique called Optical Image Stabilization (OIS) is commonly used. In this thesis the possibility of a new more advanced control algorithm for the OIS lens was investigated. More specifically, the research questions ask whether it would be suitable to implement a Model Predictive Controller (MPC) as an alternative solution to the industry standard PID controller, which is the current solution for the system used in this study. The performance of the two controllers has been analyzed to determine if the alternative solution outperforms the current one and the feasibility of the implementation has been evaluated. As the performance of an MPC is heavily reliant on the quality of the model used, a large part of the thesis is focused on system identification and model structures. While the MPC was implemented successfully, the results show that it could not outperform the PID. The performance of the MPC showed potential, but was impaired by the effects of cross-axis coupling. The inferior performance was likely because of insufficient accuracy of the models, and unmodeled dynamics of the system.}},
author = {{Lindmark, Albin and Philipson, Sofia}},
language = {{eng}},
note = {{Student Paper}},
title = {{Model-Based Control for Optical Image Stabilization}},
year = {{2025}},
}