Author: Tan Zhang, Suman Banerjee
TV white spaces could provide us with precious spectrum resource. It could enable many interesting applications. We need to look up databases before using the channel to avoid interference according to FCC rules.
However, the current commercial databases are not that accurate. According to their experiments, the databases have low under protection rate that is usually less than 1%, but it has wasted a lot of resource to protect TV and wireless microphones. 42% of the spectrum is wasted according to their measurement.
The authors proposed a system to fix the database using opportunistic measurements. The challenge of detecting primary signal comes from the very weak signal. The FCC rules indicates that the channel is unoccupied if the primary signal is weaker than -114dBm. This is hard to detect even on the most advanced spectrum analyzer.
To solve these problems, the author proposed the V-Scope system.
It includes a zoom-in pilot-tracking algorithm that could measure the power of TV signals up to -120dbm. At the same time, the author also proposed a model refinement procedure to augment the existing propagation model with their measurement.
To evaluate the V-Scope system the author carried out experiments in a 100 square-km area with over 1 million result collected. The experiment shows that opportunistic wardriving is useful. Users could reclaim at most 33% of the measured locations.
The paper pointed out the problem of current TV Whitespace database and provided a practical system to fix that. Experimental results indicate a significant improvement over the state of the art database.
Q: Wireless microphones may not transmit all the time. Opportunistic sensing may miss the signal.
A: Sampling at the same place for multiple times could solve this problem. Public transit will pass by the same place for many times per day.
Q: Why there are commercial databases?
A: FCC specified the standards, and the commercial databases make profit from providing services to CR devices.
Q: Is it legal to do so? Does FCC force these rules?
Q: Does FCC force these models?
A: Yes, the models are specified in the standards.
Q: Why do you use this model?
A: Many models are trying to predict the channel availability. Log based model is generally hold, so we need to tune the parameters only.
Q: Which databases did you studied?
A: Google and SpectrumBridge.