Click for next page ( 36


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 35
CHAPTER 4 CONCLUSIONS AND RECOMMENDATIONS Based on the findings presented in Chapter 3 and discussed in Chapter 4, it is suggested that the traffic-actuated control analysis methodology developed in connection with this project is valid and satisfies the project objectives. The signal timing estimation model developed has demonstrated a clear superiority over the status quo. Of particular importance are the model features that make it sensitive to the traffic-actuated controller settings, the free queue phenomenon and the coordinated control mode. The combination ofthese features gives the model capabilities that are needed by the analyst and cannot be found together in any other technique. RECOMMENDATIONS ON FUTURE RESEARCH The proposed analytical model was tested extensively with simulation. Simulation results were augmented with field data to the extent at which project resources were available. Some additional field data collection would be beneficial, keeping in mind that the model is analytical in nature and therefore depends less on empirical data than a purely empirical model would. During the course of the research, a number of questions arose that could not be fully addressed within the scope of the project. Additional research in these areas would be very helpful in the further enhancement of the HCM methodology and would contribute to its future credibility. The most critical questions that affect the full implementation ofthe proposed model in the Year 2000 HCM are described as follows: Choice of Incremental Delay Models Two alternative incremental delay models were considered in this study, and a clear winner did not emerge. Model II was recommended over Model I because of its similarity to the existing HEM model structure. Further consideration should be given to which of these delay models is most appropriate. There is a possibility that each model would prove to be better suited to different conditions than the other. More extensive field data collection would be required to settle this issue. Shared Lane Permitted Left Turns l The empirically based treatment of shared lane left turns now included in HCM's Chapter 9 does not lend itself well to the iterative computational structure. Idiosyncracies ofthis model prevented the computations from converging to a unique solution in many cases. An analytical substitution was therefore necessary as an expediency. Of particular importance is the equilibrium between the shared lane and the through lanes, especially with traff~c-actuated control. The computational structure NCHRP Project 3-48 Final Report: Page 35

OCR for page 35
developed to implement the proposed analytical mode! would accommodate an iterative technique that would seek equilibrium here. This feature should be added to the Year 2000 HCM procedure for signalized intersection analysis. Retrofit to Pretimed Control equally to ~. . ~ The traffic-actuated signal timing analysis model described in Appendix C could apply dreamed control. The use ofthis model for pretested control would enhance the accuracy of pret~meo control analysis, but more importantly, it would support a better comparison between pretimed control and traffic-actuated control. While it would not be difficult to adapt the procedures to pretimed control, it would be necessary to abandon the simple description of the operation now offered in HCM's Chapter 9 in favor of ~ much more complex procedure. No iteration would be necessary with pretimed control except for the proposed enhancement to the shared lane left turn model, however the level of complexity would still be too great to implement in worksheets. The Highway Capacity Committee must accept the fact that future enhancements to the signalized intersection analysis procedures cannot continue to be based on the simplistic notion of a worksheet. RECOMMENDATIONS ON TElE ROLE OF SIMULATION The role of simulation as a surrogate for field data collection is an important issue. There are two separate purposes for simulation in this context: . Simulation models may be used to generate data as an alternative to field data collection in the development of empirical relationships, and Analytical models may be validated, or at least their credibility may be improved' by demon- strating their agreement with the simulation results. The first application involves a much bigger "leap of faith" than the second because it requires absolute confidence In the ability ofthe simulation mode} to replicate the real world. Simulation was not therefore used for this purpose in any of the NCHRP Project 3 -48 tasks. It was, however, used extensively for the second purpose, i.e., comparison of results with the analytical models developed under this project. Excellent agreement was obtained between the analytical and simulation results. What is less clear at this point is the role of simulation as an alternative to the analytical modeling process. This issue was introduced briefly In Chapter T. There is no question that simulation models such as NET SIM are able to address a much broader range of problems than analytical models. The main question at this point is how to resolve the differences that wall inevitably occur in a comparison of the results obtained from these two techniques. NCHRP Project 3-48 Final Report: Page 36

OCR for page 35
The ability to reconcile these differences is very important to the credibility of both techniques. The following suggestions are offered for enhancements to simulation models such as NET SIM to promote a more unified set of capacity and delay analysis techniques: i. The simulation model should provide a direct estimate of capacity for each approach. This would probably require an analytical component to be incorporated into the simulation model. 2. The resolution in the steady state headway in the NETSIM input data should be increased fromO.] seconds to O.O] seconds. 3. NET SIM should produce explicit outputs of the traffic-actuated signal timing plan. A postprocessor for the animated graphics output files had to be developed as a part of the project described in this report to coax this information from the NET SIM output. This feature should be internalized in NETSIM. 4. A feature should be incorporated into NET SIM that would T) recognize vehicles leaving a stop line and accelerating in the downstream link, and 2) transfer the resulting delay into the upstream link. This would produce signal delay estimates that are more compatible with the HEM model. 5. More control should be provided over the bunching characteristics of amvals at a signal. The current version of NETSIM creates this bunching implicitly by car following logic alone. This makes the bunching characteristics highly dependent on the link length. Consequently, it is very difficult to control the arrival distributions and make them similar to the distributions used in the analytical models. The principal product of this project is software dependent. The iterative nature of the analysis procedure does not lend itself to manual implementation. It has therefore been developed as a stand- alone application that will be very useful in practice regardless of any decisions concerning its adop- tion as an official part of the HCM Chapter 9 procedure. In its present form, it could be distributed and used effectively by traffic engineers irnrnediately. Its utility to the profession would be greatly enhanced, however, if it were adopted by the Capacity Committee. A strong effort should be made to encourage the Cornrnittee in this direction. In any event, if the NCHRP 3-48 Project Pane! accepts this work, it could and should be provided to McTrans and other software distribution centers as a finely operational product VCHRP Project 3-48 Final Report: Page 3 7