Bluetooth - Evaluation and results Torbjørn Haugen Ass. Professor Erlend Aakre Phd Student NTNU Traffic Engineering Research Centre Erlend Aakre & Torbjørn Haugen, NTNU
Content Background Field tests Trondheim 2011 - Swarco Oslo 2012 AADI (TrafficNow, DeepBlue) Trondheim 2013 (BLIP, Swarco) Oslo 2013 (Digital Innovation) 4 different bluetooth suppliers Results so far Number of registrations and travel times Challenges Recommendations and best practices
Background Norway has a long history of measuring travel times based on electronic toll collection tags (AutoPASS). This is a well proved technology and the data quality is high. However, the installation costs are relatively high because we have to mount antennas over each lane. www.reisetider.no Bluetooth has the recent years become an interesting technology for traffic and travel time registration. The products are easy to install, and they have low cost compared to other technologies. The last years we have looked into bluetooth as an alternative technology, and equipment from different suppliers have been tested. The purpose of the field tests has been to study whether bluetooth can be an alternative to ETC-tags, and if so where and how.
Test Area in Trondheim 2-lane Highway 4-lane Highway 6-lane Highway 2-lane + buslane Tags and bluetooth Portable bluetooth
Test area in Oslo 4-lane Highway 4-lane + buslane Tags and bluetooth
Test site / Equipment
Trondheim 2011 Swarco Bluetooth vs AutoPASS Number of travel times Bluetooth AutoPASS
Trondheim 2011 Swarco Bluetooth vs AutoPASS Number of travel times Bluetooth AutoPASS
Detection rate Trondheim AutoPASS 80% Bluetooth 25-60% BLIP Track better than other Oslo AutoPASS 60% Bluetooth 25% (AADI) Copenhagen (info from BLIP) BLIP Track 27%
Oslo 2012 - Travel Time AutoPASS Mode based Bluetooth Mode based Blootooth AADI (TrafficNow)
Trondheim 2013 BLIP vs AutoPASS Number of travel times Number of travel times AutoPASS Bluetooth Travel Time
Trondheim 2013 - Travel Time from BLIP Bluetooth Median based
Trondheim 2013 - Travel Time from BLIP
Test with 9 units in the same vehicle Sensor didn t detect every unit 0-8 seconds between time stamps
Results so far.. Bluetooth has less detection rate than AutoPASS, but still enough for travel time information on main roads Some bluetooth sensors are better than others (1.5-4 times higher detection rate) Portable solutions varies when it comes to battery capacity and solution No significant difference in detection rate between fixed and portable equipment from the same supplier Angle of antenna doesn t influence the detection rate much Some bluetooth units (i.e. iphone) is seldom detected
Results so far.. Problems with combining sensors from different suppliers because of different format and encryption. Suppliers offer their own software solutions, and both filtering/calculation algorithms varies i.e. results are not comparable. This is a challenge especially for statistical purpose and evaluation.
Further tests We want to test the different sensors in an urban environment with cars, busses, bicycles, pedestrians, offices, etc
Best Practice Bluetooth is an alternative technology (instead of AutoPASS) for travel time registration on main roads Detection rate varies between different sensors, and sensor type seems more important than installation and adjustment of the sensor There is also a need for a standard format on bluetooth-id so sensors from different suppliers can be combined It is important to know how travel time are calculated and there is a need for a standard method. Norway will probably use bluetooth for travel time registration, but only as a sensor in the existing travel time information system i.e. we will use the same algorithm as for AutoPASS We will also start a project to evaluate if/how use of different sensors and algorithms influence calculation of average travel time
Questions? You can also contact us: Torbjørn Haugen, torbjorn.haugen@ntnu.no, +47 907 77 865 Ass. Professor, NTNU Traffic Engineering Research Centre Erlend Aakre, erlend.aakre@ntnu.no, +47 932 48 754 Phd Student, NTNU Traffic Engineering Research Centre