4. Conclusions

The "Before ATCS Installation" data collection can be summarized as a successful data collection project. The UTL team collected all relevant data that were planned to be collected. However, recent developments in the Bear Hollow area tend to spoil the validity of several traffic metrics collected during the "before" data collection project. Two traffic signals have been planned for installation in the next year. In addition, one intersection will be completely redesigned and moved from its current location. These changes in traffic signal infrastructure will have an impact on the validity of the evaluation of ATCS in Park City. Most of the data metrics qualifying traffic control performance for the whole system will be invalid. For example, for the data collection after the ATCS is installed, recordings from the travel time runs are likely to report a higher number of stops, total delay, and travel times due to additional stops at new signalized intersections. The expected low-grade data that will be affected by installations of new traffic signals are:

• Journey time runs
• System-wide stops
• Overall delay
• Running speed

The collected data that can still be used to compare before and after performances are:

• Turning movement measurements
• Sat flow rates
• Intersection-based control delay

Another problem is that additional "before" data cannot be collected after all the new signalized intersections are in place. The process of installing new traffic signals will coincide with the process of installing the ATCS. Therefore, by the time all new traffic signals are installed, SCATS will also be installed and there will be no opportunity for another "before" data collection.

Considering the shortcomings of the existing approach that was influenced by outside factors, we suggest:

• Make use of limited good "before" data – this approach limits evaluation to compare only individual intersections as opposed to comparisons of the performances of the entire system before and after SCATS is installed
• Focus on a "with/without" evaluation instead of a "before/after" approach – this approach depends on the political decision to turn SCATS off for few weeks
• Build the VISSIM network according to the model – this is the most rigorous approach that will enable comparison. It also gives UDOT the opportunity to explore various traffic signal scenarios and future expansions of the system (not only in Park City, but on any other network), such as:
  • "before/after" approach
  • system expansion
  • "freak" events

References

• NCHRP - National Cooperative Highway Research Program. 'Quantifying Congestion'. Volume 1 - Final Report. NCHRP Report 398. Transportation Research Board, 1997, Washington, D.C.
• Garber, Nicholas J., Hoel, Lester A., Traffic & Highway Engineering, 2001