The effort required to simulate a network is directly proportional to the size and complexity of the network.” One of the limitations encountered with the INTEGRATION model was the inability to establish a run with the network loaded entirely with unguided vehicles. Hence, the base run for the ATIS investigations was carried out with a network loaded entirely with guided vehicles. The research by Bacon et al. also highlighted the difficulty that detailed simulation models have in determining origin-destination demand patterns. Hence, it was recommended that a travel demand model be applied to generate origin-destination matrices for future simulation exercises. These investigations revealed that for a minor incident, a 100 percent IVIS system could offset the adverse traffic effects. However the effect of the major incident was only partially offset by all vehicles in the system being guided vehicles. Bacon et al. did not comment on how the results might change if not all vehicles were guided. There is a generally accepted theory that a threshold exists for increasing benefits from the penetration of IVIS technologies. Glassco et al. State that “the value of information to the ITS user decreases as the proportion of the population receiving that information increases. The value of information to the overall system, however, generally increases as the proportion of the population receiving that information increases.” The exception to this is when a very large proportion of the population reacts to the information, in which case the system performance may decrease. Gardes and May found that more than half of the system benefits resulting from the presence of guided vehicles on a portion of the Santa Monica Freeway corridor was possible with the first five percent of guided vehicles.

An average travel time savings of 3.8 percent resulted from the presence of five percent guided vehicles,blueberry grow pot compared to 5.6 percent for 25 percent guided vehicles, and 6.2 percent for 100 percent guided vehicles. Additionally, Gardes and May found that the travel time benefits of route guidance were similar under incident and non-incident scenarios. Another interesting finding of the study by Bacon et al. study was the conclusion that the effect of “distorted” information on total trip time was almost as adverse as a major incident upon which motorists receive perfect information. Despite the difficulties encountered with applying the INTEGRATION model, Bacon et al. still considered it to be the most suitable model for such studies. Unfortunately, this same study made no attempt to assess the environmental impacts. Nevertheless, this project team intends to consider extending their work to evaluate energy and environmental impacts.Glassco et al. found that adaptive signal control strategies and traveler information services have positive impacts on user and system efficiency for both recurrent congestion and non-recurrent congestion scenarios. However, the modeling studies revealed greater benefits resulting from these ITS user services, under non-recurrent congestion conditions . One specific finding was that vehicles that bypass incidents in an inter-city corridor network due to a route guidance system can reduce travel delays by 50 percent or more. The potential benefits of the ITS user services modeled were found to be greater when the network experiences a greater deviation from expected conditions. Hence, these ITS technologies may be most beneficial for a highly variable transportation system. In a highly predictable system, traffic control systems and travelers tend to adapt their plans according to the anticipated congestion. This process can take place without the influence of ITS technologies such as traveler information services. One of the interesting key findings of the study by Glassco et al. is that “the value of traveler information is proportional to the value of the alternative opportunities available. When the time required to follow an alternate route or to shift mode is high, then the timeliness and accuracy of travel time prediction is more critical.”

This would implies that pre-trip information has greater potential for providing benefits than en-route information, as once a traveler has started a trip he is not likely to have the opportunity to change mode, and by definition the number of route choices decreases for any trip as the traveler gets closer to the destination. In addition, once the trip has begun, the choice has already been made to make the trip at that time. If information was provided prior to the trip, warning of high levels of congestion, travelers have the option of postponing or even canceling the trip. Glassco et al. concluded that perhaps one of the most positive results of their studies of ATIS user services was that a relatively small number of probe vehicles are needed to support adequate network surveillance. They found that an unguided probe population as low as one percent can provide over 50 percent of the benefit of full surveillance, and that the value of route guidance with low probe percentages can be increased by updating travel times more frequently than every ten minutes. It is more effective to use unguided vehicles as probes than guided vehicles, since if all guided vehicles are being routed around an incident, the route guidance system does not learn that congestion has dissipated when the incident is cleared. Van Aerde and Rakha present the methodology and results of a modeling study of the Orlando TravTek operational field test. This test involved the deployment of 100 vehicles equipped with in-vehicle information systems for a one-year period. The modeling study attempted to extrapolate from the available field data the expected performance of a TravTek type system for levels of market penetration ranging from one percent to 100 percent, in terms of measures of effectiveness such as vehicle stops, fuel consumption, vehicle emissions, and accident risk. The Orlando TravTek network consisted of 2,700 links and 90 origin-destination zones, with a traffic loading of approximately 65,000 vehicles. The INTEGRATION model was used to simulate recurring congestion scenarios and freeway incidents for the afternoon peak period, with a small number of simulations of increased and reduced traffic demand to represent future and off-peak conditions, respectively. Consistent with findings by Glassco et al. , travel time benefits were found to be greater for the higher levels of congestion. An interesting finding relating to the much debated issue of latent demand, was that the LMP can be increased to provide travel time savings that have the potential to offset much of the induced demand that may be generated.

Travel time reductions were found for all levels of market penetration analyzed; however, the most significant marginal benefits were achieved at lower market penetration levels. The maximum reduction in travel time was found to occur for a LMP of 100 percent, where the network travel time saving was 15 percent. In addition to travel time savings, the study considered travel distance savings and found that a 7 percent reduction was possible at an LMP of 100 percent. En-route traveler information can have two conflicting effects on travel distance: distance traveled can be reduced by route guidance information which minimizes navigational errors, and distance traveled may be increased when users receive real-time traffic information and subsequently divert around an incident. The fuel consumption benefits of the TravTek system were predicted to be proportional to the travel time savings, with respect to the various LMPs. The magnitude of potential fuel consumption benefits is related to the topology of the network. That is, if when an incident occurs there are alternative routes available with little or no increase in travel distance, the potential for energy savings is greater. This is similar for emission benefits; however, the relationship between vehicle emissions,square plastic pot and traffic conditions and distance traveled is much more complex than for fuel consumption. HC emissions were found to decrease for all LMPs, with a maximum of 16 percent being achieved at a LMP of 100 percent. CO emission increased up to three percent for LMP’s less than 10 percent, and then decreased up to seven percent for LMP’s up to 100 percent. Van Aerde and Rakha claimed that ATIS may tend to always cause an increase in NOx emissions—they predicted increases in NOx of up to five percent, with the only reduction being one percent at an LMP of 100 percent.FOTs have provided little direct evidence of emission impacts of ITS in general, but they do provide some insights to ETC and ATSC. The US DOT’s comprehensive review of data available from FOTs does not even classify emissions or other environmental impacts in their summary table. Other FOT reviews do extract some limited results specific to ETC and ATSC and other traffic flow control systems, but echo the DOT’s general conclusion that FOTs have provided limited insight into environmental impacts. Little and Wooster reported that the focus of most of the FOTs had been on “technical feasibility and, to a lesser extent, user response.” Further, they note that most of these tests address more conventional traffic management user services, such as dynamic signal coordination. ETC facilities are one of the types of ITS technologies for which FOTs have produced emission assessments. Large emission reductions can be achieved at individual electronic toll facilities; however, the overall benefit is dependent on the frequency of toll plazas. The reductions, although large, are highly localized and may be insignificant at the network level. ATSC has shown significant potential for providing energy and environmental benefits in more than one FOT. Modeling results support the FOT conclusions regarding ETC and ATSC and provide additional insight for traveler information and vehicle guidance. Results reported in US DOT show that the application of the ATMS user services provided a general increase in speed and a reduction in total delay, although they do not analyze emissions specifically. Maximum benefits were achieved with the simultaneous implementation of signal coordination on arterial streets and ramp metering on freeway facilities.

Glassco et al. found that ATSC strategies and traveler information services have positive impacts on user and system efficiency for both recurrent congestion and non-recurrent congestion scenarios. However, the modeling studies showed greater benefits resulting from these ITS user services under non-recurrent congestion conditions . Modeling results reveal mixed results for navigation and traveler information systems. Analysis of Orlando’s TravTek network indicated varying levels of travel time, travel distance, and emissions effects depending on levels of market penetration. HC emissions were found to decrease for all LMPs, with a maximum of 16 percent being achieved at a LMP of 100 percent. CO emission increased up to three percent for LMPs less than 10 percent, and then decreased up to seven percent for LMPs up to 100 percent. Van Aerde and Rakha claimed that ATIS may always tend to cause an increase in NOx emissions. They predicted increases in NOx of up to five percent, with the only reduction being one percent at an LMP of 100 percent.Many researchers are trying to estimate the market for various ITS technologies. Three methods that are popular for estimating the market for ITS technologies include: focus groups with potential market segment members, interviews, focus groups and surveys of ITS experts, and evaluations of the traveler response to ITS demonstration programs . Without actual data and experience, it becomes virtual impossible to accurately predict the market for single and multiple ITS products and services. In market and traveler behavior research, the more innovative the solution or the farther into the future an innovation will be introduced, the more researchers must rely upon survey research methods and participant responses to experimental situations to understand the types of impacts that are probable or even possible. Turrentine and Kurani conclude that an experimental situation will often elicit and engage the decision processes of its participants and reveal participant lifestyle goals.Based on their new technology experience, Turrentine and Kurani conclude that current Stated preference survey methods often produce inaccurate estimates of the demand for unknown and new technologies. Other methods such as interactive Stated response when presented in the context of a household’s actual activity space and modal split data can produce more accurate demand estimates. In market evaluations of electric vehicles , many Stated preference studies often estimate huge price penalties for limited-range vehicles . In general, these studies rely on data from hypothetical-choice experiments in which participants are presented with choice sets of the vehicle. Then, participants are asked to identify the one vehicle, from each of the choice sets, they would be willing to purchase.