2020 Fiscal Year Research-status Report
Stochastic Control-Theoretic Approach to Development of Simultaneously Cyber-Secure and Energy-Efficient Randomized Transmission Methods for Dependable IoT
| Project/Area Number |
20K14771
|
|---|
| Research Institution | National Institute of Informatics |
|---|
Principal Investigator |
CETINKAYA AHMET 国立情報学研究所, アーキテクチャ科学研究系, 特任研究員 (60851730)
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Keywords | Control systems / Jamming attacks / Multi-agent systems / IoT |
|---|
| Outline of Annual Research Achievements |
In the first term (2020) of this project, there were three main research achievements.
1) Security of control systems: With a multi-objective genetic algorithm, a new wireless transmission approach was developed for control systems. This new approach relies on heuristic search algorithms, but it was shown to have theoretical security guarantees against jamming attacks. Several case studies showed the superiority of the approach compared to other approaches.
2) Game-theoretic analysis of jamming attacks in multi-agent systems: A game-theoretic framework was developed to provide a complete analysis of jamming attacks on the communication links of multi-agent systems.
3) Optimal transmissions in IoT: An analysis of optimal transmission times against potential jamming attacks was conducted. The analysis revealed the scenarios where the randomizing the transmission times is the optimal approach.
|
| Current Status of Research Progress |
Current Status of Research Progress
1: Research has progressed more than it was originally planned.
Reason
One of the critical challenges of this project was to design transmission methods that are optimal/suboptimal with respect to several measures such as IoT system security, performance, and energy use. Collaborations with researchers from the field of software engineering enabled a breakthrough in the development of a completely new design approach. This approach is based on multi-objective genetic algorithms. It turned out that this heuristic approach also allows theoretical analysis by using probability theory and control theory. Such an analysis is quite rare in the literature.
|
| Strategy for Future Research Activity |
The next phase of this research project will handle four topics.
1) Cybersecurity analysis and design problems will be addressed for quantized and nonlinear systems. Quantization and nonlinearity are essential properties of realistic systems which are often seen in practice.
2) Effects of jamming attacks on model predictive controllers will be analyzed. While model predictive controllers are common, the literature on the security analysis on these controllers do not explain how malicious/nonmalicious packet losses affect the performance.
3) Security of discrete-time multi-agents systems will be investigated.
4) Regarding wireless transmissions in IoT, an important challenge in this project is to incorporate learning in designing efficient transmission methods. To this end, data-driven control approaches will be investigated. In addition, feasibility of the proposed transmission appraoches is currently being investigated with the help of a practical IoT testbed.
|
| Causes of Carryover |
Due to the unexpected Covid-19 pandemic, academic conferences were held virtually/online. As a result, in 2020, there were no travel expenses. Some of the budget that was originally planned to be used for travel expenses was replanned for the purchase of a computer and IoT testbed for conducting numerical experiments. In the next fiscal year, the incurring amount is planned to be used for the purchase of books, purchase of devices for the development of an IoT testbed, payment of journal article processing charges, and registration for academic conferences.
|
