Probabilistic Model Based on the Effective Range and Vehicle Speed to Determine Bluetooth MAC Address Matches from Roadside Traffic Monitoring

摘要

The use of Bluetooth technology to determine vehicle travel times offers several benefits over traditional methods-most notably, an increased amount of data points. An understanding of the Bluetooth device discovery procedure provides insight on the impacts of vehicle speed and device effective range on the number of in-range scan intervals, allowing the data collection setup to be modified to improve the probability of device detection. It is determined that the number of in-range scan intervals increases as vehicle speed decreases and that vehicle speed has a greater influence when the device effective range is decreased. Our match probability is defined as the probability that a single vehicle containing a discoverable Bluetooth radio will be detected by two different scanning radios, not a sampling rate produced by the ratio of MAC detections to total number of passing vehicles. The probability of device detection escalates for roadways with lower functional classification, slower posted speed limits, and during periods of congestion. At speeds of 56 km/h (35 mi/h) the probability for a match exceeds 30%, whereas at faster speeds [105 km/h (65 mi/h)], the probability decreases to under 10% because of poor signal strength. Optimizing the data collection setup to increase the effective range between the target radio and roadside scanning radio will result in an improved probability of device detection for faster traffic streams.