Skip to content

Research

Here are some brief descriptions of my current projects, each are in various stages of development (conceptualization, writing, data collection) and are carried out with different collaborators—technologists, information scientists, economists, and computer scientists. If you want more information or pre-prints related to these projects, please email me.

 

Decision and Risk Analysis of Spectrum Usage
The motivation for doing this research is that spectrum sharing is an investment decision for spectrum users. Their choices are made based on internal factors such as incentives and limitations, as well as external factors such as spectrum usage risks and mitigation strategies that are embedded in each sharing method. Consequently, an explicit understanding of risks and mitigation strategies as well as factors that impact spectrum users’ decisions is essential and will assist spectrum users in selecting the most appropriate spectrum usage model given their situation. Additionally, it will help policy makers, operators, and the spectrum market create interventions in order to obtain favorable outcomes.

 

Spectrum Occupancy Model for Machine to Machine / Device to Device Communication and Enforcement
Machine to Machine (Device to Device) communication is the future of wireless network. How to allow devices to share contents in the opportunistic network is the key question. The spectrum occupancy model tries to answer following questions: (1) what is the probability that devices can have a communication link in certain vicinity and time? (2) what is the probability that devices can have intented content under opportunistic network? (3) what is the probability that service level agreement can be met? etc. Spectrum occupany model not only provides a baseline for M2M (D2D) communication, but also assists ex ante enforcement.

 

Spectrum Trading
Spectrum trading is an effective way to allocate spectrum since it is bilaterally negotiated. However, it is difficult for primary users to accurately predict service demand a priori, especially for an extended period of time. Therefore, primary users may adopt a conservative strategy, whereby they over-reserve channels for their internal user and only make a minimum number of channels available to secondary users. An aggressive primary user, on the other hand, may opt to increase the number of leasable channels at the risk of service degradation. Neither of these strategies is effective. What is needed is an incentive-based, risk-mitigating framework, whereby primary users are encouraged to increase the leasable channels, while limiting the risk of service degradation. Drs. Znati, Weiss and I proposed the Coopetive Spectrum Trading framework to achieve this goal by two schemes: (1) revocable leasing protects primary users’ operation, and (2) the penalty of spectrum revocation assures secondary users’ investment.

 

Spectrum Assignement with Improved Database Assisted Approach
Spectrum assignment is not quantity limited, but interference limited. In order to promote coexistence of radio devices, spectrum assignment should convert from transmitter-power constraints into receiver interference limit. It is because receivers are the ones that differentiate desired data from interference. Database is already deployed in spectrum assignment in TV White Space (TVWS), however, it only displays information about PUs' usage. In other words, SUs do not know other SUs' working frequencies, which may
lead to tragedy of commons. Moreover, with many SUs operate in the same frequency bands, even each of them following the transmitter power constraints, the sum of the signal level may interfere PUs' systems. A improve database assisted approach can optimize spectrum assignment based on interference limit, which maximize the coexistence of radio devices while minimizing risks.

 

••••

Past Projects

Spectrum Sensing Security
From the security perspective, malicious users can increase spectrum sharing risks in sensing based DSA by two ways. On one hand, by always reporting primary users present, they can deter secondary users’ transmissions. On the other hand, by always reporting primary users absent, secondary users will bring harmful interference to primary systems. Reputation-based schemes are a common methodology to handle falsification attacks. Traditionally, sensing nodes with false reports are punished by a reduction to their reputation score. However, it may lead to punishment of benign nodes, since wireless channel irregularities such as fading and multipath also lead to false reports. Therefore, identifying the source of false report is important to identify true malicious users.

 

Spectrum Rights and Enforcement
How to define spectrum rights determines the allowable spectrum sharing methods. Dr. Weiss and I proposed defining spectrum rights as interference rights (similar to pollution rights). An interference right gives a user the right to interfere with another user up to a specified level. Under an interference rights regime, existing license holders can write interference rights on their licenses, which can be traded, combined, or exchanged with other users. Enforcement is another attempt that targets minimizing spectrum users’ uncertainties and risks. In enforcement related research, Drs. Weiss, Lehr, our colleagues and I provided a broader framework for evaluating enforcement under various modes of DSA in order to ensure license holders’ rights under different spectrum usage models, fairness in spectrum sharing among users, primary and secondary users’ conformance with contracts, detection of “free riders,” and fight for the “hit-and-run” problem.