Highway Performance Monitoring System

Highway Performance Monitoring System – Motorcycle Vehicle Miles of Travel 2007

The comment period for the Highway Performance Monitoring System draft ended June 30, 2007. Below: the draft and the list of states/links that submitted comments. Some states did not mention motorcycle VMT.

Draft: http://dmses.dot.gov/docimages/pdf99/436214_web.pdf

[motorcycle VMT related]

Page 38:

4.3.1 Immediate Implementation

The mandatory Reporting Year is indicated in parenthesis. Unless otherwise noted, states will be able beginning reporting these data to FHWA in 2008. States already submitting the new/changed data, are expect to continue to do so. All other states are expected to submit data in the revised format as soon as it is available, and not wait till the required reporting year (2010).

4.3.1.1 SAFETY

Motorcycles (2008 Reporting Year)

• States to include motorcycle travel data collected in calendar year 2006, as part of their June 2007 HPMS submittal. The reporting of these data in 2007 would be optional, but starting in 2008, the reporting of these data in the area-wide summary table would be mandatory.

Why Motorcycle Travel is Currently Not Being Reported

Data Quality/Limited Availability (7): 54%

Not Collected (4): 31%

State Policy (2): 15%

Page 97:

A. Motorcycle Data

The quality of travel data for motorcycles has been an ongoing area of concern. In HPMS, the motorcycle travel data comes from the area wide summary table and is coded as percent travel by vehicles by functional system. There are a number of issues with the motorcycle data. In the area wide summary table, reporting the motorcycle data is optional and several States have chosen not to code it. A greater concern is the fact that collecting motorcycle travel data is not a priority in most States. Since motorcycles are much less common than other types of vehicle travel, a proportionately larger effort would be needed to detect the same percent change. Motorcycle data is important because the number of serious and fatal crashes associated with motorcycles continues to increase nationally.

Page 98:

A. Motorcycles

Some States may have problems providing accurate classification count data for all functional systems. The higher functional systems, especially those on the State highway systems should be fairly well covered, but on lower order systems these data may be more difficult to provide. Since these data are being used more frequently by FHWA and NHTSA as input into performance measures and to provide incentive rewards to states, the quality of this information becomes more critical. Of particular interest, because of the significant increase in motorcycle fatalities, is motorcycle travel. The limited motorcycle travel data affects the credibility and accuracy of analyses and hampers the ability to produce precise motorcycle statistics. This problem is compounded by the low priority placed on obtaining accurate motorcycle travel by many states. Within FHWA, the area wide summary table is the primary source of travel statistics by vehicle type. Part of the reassessment should be to determine how states are populating this table and evaluate whether it would be beneficial to collect additional metadata on the procedures the states are using.

Page 101:

A. Motorcycle

The options for revising the motorcycle data depends heavily on the quality, consistency, timeliness, reporting, and collection issues raised in the previous section. An addition source of motorcycle travel data that should be explored is the National Household Travel Survey (NHTS). The survey samples households by telephone from all regions across the country. The data include demographic characteristics of households, people, vehicles, and detailed information on travel for all purposes by all modes. It may be worth considering a special motorcycle “add-on” to the 2008x NHTS.

The quality of motorcycle travel data will likely continue to suffer until States start to make the collection of these data a priority. FHWA’s Office of Safety and the NHTSA feel that collecting these data from all States, regardless of the initial quality, is very important. Therefore, they are recommending that that States begin submitting these data as soon as possible. It has been suggested that States include motorcycle travel data collected in calendar year 2006, as part of their June 2007 HPMS submittal. The reporting of these data in 2007 would be optional, but starting in 2008, the reporting of these data in the area wide summary table would be mandatory.

State DOT Comments to HPMS:

California Comments

PDF 5 pages: http://dmses.dot.gov/docimages/pdf101/474775_web.pdf

Page No. 4-10

Proposed Change

Reporting of Motorcycle data is optional in 2007 but will be mandatory in 2008

Probable Impacts

Existing reporting practices may suffice in that the state is able to complete the summary data. Contracts for vehicle classification counts are let in order to gather some of the off-system traffic data. This should be sufficient to gather enough information to estimate the motorcycle VMT by function class. But will the FHWA develop any sort of minimum sampling requirements as to vehicle classification counts in order to attain some confidence level for motorcycle VMT? If so, more class counts could be needed on system and off.

The on-system classification counts are done at control count stations. Many of these are weight-in-motion (WIM) sites. The WIM sites were designed to accurately capture truck travel but not motorcycles. Motorcycles have a short wheel spacing and the detectors on WIM sites are typically spaced at 16 feet hence the ‘trap’ is too long. More advanced detection devices are on the market that could more accurately capture motorcycles but they are expensive and it would take several years to put them in place.

Florida Comments

PDF 3 pages: http://dmses.dot.gov/docimages/p101/474376.pdf

Page 3 Traffic Data

Motorcycle data: We currently collect and report motorcycle data, but we are concerned about the usefulness of this data. Sensors don’t register motorcycles consistently, motorcycles have different usage patterns during the week (more on weekends, which are avoided for most traffic data collection), and different seasonal patterns (more during good weather and during special events, such as Daytona Bike Week).

Georgia Comments

PDF 8 pages: http://dmses.dot.gov/docimages/pdf101/474854_web.pdf

4. Safety

4.1 Mandatory reporting of motorcycle data starting in 2008 submittal

GDOT currently reports motorcycle data in HPMS. However, we are concerned about recent comments related to collecting motorcycle data on weekends. We have weekend motorcycle data from ATR’s but would not be in favor of collecting portable vehicle class counts on weekends. This would be very costly and substantially impact our schedule for portable vehicle class collection operations.

Hawaii Comments:

PDF 11 pages: http://dmses.dot.gov/docimages/p101/474743.pdf

Page 6

Hawaii DOT/Highways Planning Survey Section

SAFETY -Specifically, the requirement to include motorcycle travel data as part of HPMS is particularly onerous. In Hawaii, motorcycles make up just two percent of the vehicles registered, and would not warrant or make a good measure towards examining the condition or performance of the state's highways given the inordinate amount of resources that would be required to collect quality data. Much of the actual travel may also occur outside of periods and facilities traditionally regularly monitored like during weekends and on local roads. States with large percentages and mileages of local roads are at a great disadvantage in this case.

Furthermore, because of cost, minimal license requirements, no insurance, no helmet laws and ready availability for the millions of tourists (e.g. rentals on many blocks in Waikiki) who are unfamiliar with both the vehicle and the streets, <50cc scooters are frequently involved in fatal accidents, and these are recorded as motorcycle crashes (making up 25% of the fatalities reported last year). What we are getting at here is that it is not a matter of collecting more and better motorcycle travel data to find out what the safety problem is and how to fix it. There may also be other better sources or more appropriate ways to obtain data for the real problem. We would like to avoid this being an example of Horwood's 6th Law of Data Processing, which states:

"If you have the right data, you have the wrong problem and vice versa."

We suggest this be examined in greater depth by a focus peer group with a full cost/benefit analysis performed before making this a HPMS data requirement. We understand the driving force behind this requirement are the safety data needs within SAFETEA-LU; however, under the last reassessment, a similar situation was encountered where ITS data elements were added to HPMS due to the then focus ofTEA-21 only to be found not necessary/redundant and to be removed in this reassessment. We would not like to repeat something like this; moreover, we believe in staying true to the traditional mission of HPMS and focus on core data that provide true measures of the condition and performance of our highways.

Moreover, safety has not historically been a core function of the HPMS, and motorcycle travel would only be a small component within the safety arena and not representative of that performance at the system level. For safety, a more appropriate measure would be the accident data that was removed as a result of the previous HPMS reassessment. We recommend that motorcycle travel data be considered as a part of and within the framework of the short and long term studies for safety in general, particularly with coordinating with other safety databases/management systems and the investigation of seeking other information sources other than HPMS.

Illinois Comments

PDF 4 pages: http://dmses.dot.gov/docimages/p89/458145.pdf

FHWA is mandating that motorcycle VMT be reported with the 2007 HPMS data.

Previously motorcycle data has been optional and Illinois has not submitted data for motorcycles.

We have four issues with the mandating of motorcycle data for 2007:

• Lack of lead time in order to integrate motorcycle collection into the existing Traffic

Monitoring Program

• Different travel patterns for motorcycles as opposed to vehicles which form the basis for the Traffic Monitoring program

• Equipment, safety, and cost difficulties in collecting accurate motorcycle AADT

• Dangers in using simple estimates based on limited data for motorcycle VMT

Lack of lead time in order to integrate motorcycle collection into the existing Traffic

Monitoring Program

The Illinois Department of Transportation (IDOT) has been using Consultant contracts for the collection of short-term traffic counts throughout different districts in the state. As IDOT’s staffing levels continue to decline, expanded use of outside Consultants contracts will continue to grow. The last set of contracts that were awarded cover traffic data collection for 2007-2008. These contracts specify types of counts, locations, vehicle classification schemes, and equipment to use. Specific classification data requirements for motorcycles are not included in those contracts as it has not been a part of IDOT’s Traffic Monitoring Program. Even if all the other issues with motorcycle data collection could be resolved, there would not be time in 2007 to establish contracts for motorcycle specific data.

IDOT continues to evaluate new equipment and technology, focusing on safer nonintrusive collection equipment for the future needs within the Traffic Monitoring Program. New non-intrusive devices (TIRTLs – traffic infra-red traffic loggers) have been installed in continuous count locations and contracts are in place for the replacement of the HiStar NC-97 with the new NC-200 counters. If motorcycle data collection becomes a priority item in the Traffic Monitoring Program, IDOT will have to re-evaluate the long-term direction that the Program is heading to make sure that equipment being purchase today, and planned for the future will fulfill the new mandated FHWA requirements.

Different travel patterns for motorcycles as opposed to vehicles which form the basis for the Traffic Monitoring program

Motorcycles have significantly different travel patterns than passenger vehicles or trucks, so in order to develop reasonably accurate motorcycle VMT, a new and larger traffic monitoring program would need to be initiated to create day of week, and monthly seasonal factors that are regionally based specifically for motorcycles.

There are several intuitive problems with collecting short-term motorcycle counts:

• Motorcycle travel season in northern Illinois is significantly shorter than what is possible in Southern Illinois due to the seasons.

• Motorcycle travel is more heavily concentrated during weekends, but the current traffic program counts during Monday-Thursday.

• Motorcycle travel is extremely seasonal with very limited travel during winter, increasing in spring and fall, with the highest travel in the summer months.

• Weather has a significant role in motorcycle traffic. A rainy day will lower the motorcycle travel to a much greater degree than vehicular travel.

The differences in the traffic patterns will require a sizable investment in equipment, resources, and time in order to collect the background information that will need to be on file to produce accurate motorcycle VMT information throughout the state.

Equipment, safety, and cost difficulties in collecting accurate motorcycle AADT

Motorcycle detection is more difficult depending on the types of portable and permanent equipment used.

Portable or short term counts -

Manual counts can be effective in counting motorcycles, but the labor cost involved, including counting on weekends makes that option cost prohibitive.

Using multiple road tubes is one method to collect specific motorcycle data for short term counts. On lower AADT routes, this method could achieve the motorcycle classification data. However, IDOT eliminated the use of road tubes for classification back in 1999 as we moved to collecting 3-category classification based on vehicle length. The main reasons IDOT stopped using road tubes for classification were:

• Inability of multiple road tubes to stay in place and produce classification data on higher AADT routes.

• Safety factors with the amount of time it takes staff to be exposed to traffic while installing and securing multiple road tubes on the pavement.

• The amount of misclassification or default classification that comes from road tube counters.

• The need to utilize equipment that could provide additional vehicle classification data which was IDOT’s needs and eventually a new requirement in FHWA’s TMG.

Road tube classification will produce errors when motorcycles are clustered together in different wheel paths which occurs when motorcycle ride in groups.

For the last 7 years, IDOT has been using NuMetrics HiStar traffic counters that are magnetic sensors placed in the center of the through lanes. This technology has provided over 10,000 vehicle classification counts every year. Illinois has complete truck coverage on all state routes and are expanding the truck collection off of the State routes. The truck classification program far exceeds the Federal TMG requirements. The current classification scheme is based on three categories by length (0-21 passenger vehicles, 22-39 single-unit trucks, 40+ multi-unit trucks). Motorcycles give out a low magnetic signature and depending on how close the motorcycle comes to the counter, the motorcycle may or may not be counted. If a motorcyclist was going slow enough to see the small counter in the center of the lane, their natural reaction would be avoid any contact with the counter and to move away which would decrease the chance of it be counted. Since IDOT’s traffic program counts throughout the year on Monday-Thursday, the total amount of motorcycles missed by the HiStars is a very minor amount compared to the overall AADT. However, if the counter was used on a weekend during higher motorcycle traffic and used specifically for motorcycle totals, the results would not be of any statistical value. More controlled tests of the new NC-200 traffic counter, specifically for motorcycles would be needed to be conducted.

Permanent counters -

IDOT’s continuous count network is based on volume locations with magnetic inductive loops, and classification locations with a loop-piezo-loop configuration. The piezo sensor used is 6 foot long and is installed from the outside edge of the lane towards the center of the lane. The current configuration will not detect motorcycles that ride in the inside wheel path. If an 11 or 12 foot piezo was used, all motorcycles would be detected at the expense of double counting other vehicles that are straddling the centerline.

IDOT has installed the non-intrusive TIRTL devices at two continuous count locations during 2006. Classification of motorcycle data at those locations has been very good. As the TIRTL non-intrusive device is expanded through the continuous count network, base motorcycle data would be collected and eventually could be used for motorcycle seasonal factoring. The complete TIRTL set-up with equipment runs $30,000 and IDOT has limited resources on the traffic staff to conduct a widespread and quick upgrade. In addition, the 85 continuous count stations established throughout the last twenty years were selected based on the functional class, geography, and traffic volumes for total traffic. These sites may not provide the statistical accuracy for the type of systems that motorcycles travel.

More research would be needed before determining if additional continuous count sites would be required to develop a reasonable motorcycle sample.

Dangers in using simple estimates based on limited data for motorcycle VMT

FHWA did say if a motorcycle count program was not in place, an estimate can be used.

As mentioned by FHWA in memos on this subject, motorcycle travel has become of more significant importance due to the increase in fatalities. Currently FHWA has the fatalities numbers and the motorcycle registration numbers but not the VMT. We are in agreement that fatality rates based on VMT are the best measure of exposure risks for crashes.

However, any estimate would be based on very limited information or possible nationwide averages. An estimate not based on statistically significant count data could easily produce fatalities rates that show motorcycle traffic is safer or possibly considerably more dangerous than currently assumed based on the overall assumptions used in the estimate.

Any estimate that would be submitted to FHWA via the HPMS submittal would become public domain information and be considered fact because it was published. Motorcycle helmet proponents, opponents, and safety interest groups would utilize the number in lobbying efforts for their cause.

At a later point in time when the Transportation Departments throughout the country can come up with better motorcycle data based on complete actual data, the differences between the original estimates and actual data could be significant. Currently the 1 percent difference between estimated VMT by the Travel Trend publication based on the state’s continuous counter and the eventual HPMS VMT calculation seems to cause FHWA problems even with only the slight difference. With the small percentage of the VMT that is responsible by motorcycle travel, imagine years from now when the State or FHWA converts from estimates to actual data, and the published fatality rates (using Motorcycle VMT) could change by plus or minus 30% or more.

Conclusion

Based on the short time frame that FHWA has given to the States to modify their Traffic program to collect motorcycle VMT, and the equipment and motorcycle travel patterns issues mentioned previously, IDOT will not be able to provide motorcycle VMT in 2007 based on any significant level of accuracy. While an estimate could be determined based on the few TIRTL locations in the state that are collecting continuous classification data including motorcycles, or by using other assumptions, IDOT will not put a VMT number that can not be backup up with any certainty in the public domain.

We would request that FHWA gives the States adequate time to research the implication of adding the motorcycle requirement to their traffic program. Specifically here at IDOT, we would need time to test currently owned and other traffic equipment for motorcycle classification accuracy, determine the minimum statistically located count sites to determine traffic pattern for daily, monthly, and regional adjustments, and determine the overall cost and resources needs for IDOT to implement such a program.

Minnesota Comments

PDF 3 pages: http://dmses.dot.gov/docimages/p101/474662.pdf

Page 2

2. Safety

Mn/DOT has significant concerns about the collection of data on motorcycle volumes. Mn/DOT realizes the importance of data in relation to the growth in motorcycle fatalities; however, collection of accurate data requires development of seasonal factors, collection of data in additional locations and on weekends, and additional resources for appropriate equipment and personnel. Additional implementation time is needed to provide useable data with higher reliability.

Montana Comments

Montana DOT Comments:

PDF 14 pages: http://dmses.dot.gov/docimages/pdf101/474926_web.pdf

Page 2 --Traffic:

Off-system data collection in general is very difficult due to limited staff resources, equipment and weather (our data collection season is generally limited to May through September) especially in a more rural and geographically expansive state such as Montana. Adding additional items to be collected will be a significant challenge. We question why more off-system data is even needed given the level of attention we need to continue to give to higher level federal aid roads.

We understand the need for motorcycle data. However, automated data collection methods are the only way in which we know how to adequately capture it. Our automatic traffic classifiers are not designed to capture motorcycle data correctly or consistently.

While there are methods we could employ to capture this data, it would take years to “retool” our systems and the cost would be overwhelming. The June 30, 2007 FHWA memorandum requiring the submittal of 2007 motorcycle data is troubling to us. This data is known to be of poor quality and we fear it may be used inappropriately. Having said that, knowing the critical nature of the need for this data, MDT is committed to researching methods in which we can collect and report quality data. But this will take time. We recommend FHWA rescind the requirement for reporting motorcycle data in June of 2008.

In summary, we request that FHWA extend the reporting time requirements for traffic data collection.

Page 3-- Safety:

We have been submitting motorcycle travel data under the area wide summary table, however recent analysis and verification has resulted in serious doubts about the quality of that data (Please see attachment 2 - motorcycle data collection issue). In addition, because such a high percentage of total system mileage is off-system, emphasis on, and even slight incremental increases of off-system data collection represents a significant burden to the state. We have enough difficulty trying to meet our data needs on the higher level systems.

Attachment 1:

Motorcycle Classification Issues

Why are motorcycles difficult to classify automatically?

The physical size of the motorcycle compared to the roadway that it travels in means that motorcycles: • Can be anywhere in the lane of travel. • Can be ridden “side by side” in a single lane if more than one motorcycle is present. (See attached drawing and a more detailed explanation below).

Motorcycles are seasonal vehicles with their primary usage occurring in the summer months. During this time, many motorcycle-oriented events are conducted throughout Montana and neighboring states.

• The big rallies in Sturgis, SD, and Billings (just to name 2) draw large amounts of participants from all over the country. • Local events held in Montana sponsored by local clubs or dealerships for charity, promotional, or other causes. • Many of the local events take advantage of more scenic less congested “back roads” where automatic classification equipment does not exist. • A large majority of participants in all these activities travel in groups.

Why does the size of the motorcycle make it difficult to classify with our systems?

Typical motorcycle dimensions have a wheel base of less than six feet, and a width between 18 to 30 inches (not counting the handle bars). There are, of course exceptions to this general description: • “Hogs” with extra large saddle bags or front faring. • “Choppers”, which have extended front forks, can make the wheel base greater than 6 feet. • Motorcycles with side cars. • Motorcycles pulling trailers. • “Tricycles”—motorcycles with a rear axle having two wheels.

To accurately classify any vehicle within the FHWA scheme F classification definitions, three things must be known:

• The beginning and end of a vehicle. (Identifying a single vehicle). • The number of axles contained within the limits of the vehicle. • The spacing between each axle.

Referring to the attached drawing, a typical classification site is laid out to detect vehicles that are: • In single file. • Are a minimum of 30 feet apart. • Occupy at least 50% of the lane width.

As shown in the drawing, the piezo sensor, which determines the number of axles a vehicle has, extends from the shoulder toward the center of the lane covering one half of the lane only, and is designed to detect the wheels (and thus the axles) on the right-hand side of the vehicles. Typical Weigh-In- Motion (WIM) sites utilize piezos that span the entire lane width so all wheels on a given axle are detected at the same time in order to get the total axle weight all at once. To detect an axle, a wheel (or wheels) must physically strike the sensor. The loops in the classifier system, which utilize magnetic fields for detection, indicate the “presence” of a vehicle, the beginning and end of the vehicle, and the speed of the vehicle, which is used for calculating the space between axles. Note that for a WIM system, a single loop is used to determine the “presence”, beginning, and end of a vehicle, while the two piezos are used to determine the number of axles and the speed of the vehicle.

In order for a loop to detect a vehicle:

• Its magnetic field must be sufficiently “disturbed” to cause a disruption in the flow of the field. • Motorcycles pulling trailers. • “Tricycles”—motorcycles with a rear axle having two wheels.

To accurately classify any vehicle within the FHWA scheme F classification definitions, three things must be known:

• The beginning and end of a vehicle. (Identifying a single vehicle). • The number of axles contained within the limits of the vehicle. • The spacing between each axle.

Referring to the attached drawing, a typical classification site is laid out to detect vehicles that are:

• In single file.

• Are a minimum of 30 feet apart.

• Occupy at least 50% of the lane width.

The loops in the classifier system, which utilize magnetic fields for detection, indicate the “presence” of a vehicle, the beginning and end of the vehicle, and the speed of the vehicle, which is used for calculating the space between axles. Note that for a WIM system, a single loop is used to determine the “presence”, beginning, and end of a vehicle, while the two piezos are used to determine the number of axles and the speed of the vehicle.

In order for a loop to detect a vehicle:

• Its magnetic field must be sufficiently “disturbed” to cause a disruption in the flow of the field. • Disruption is typically caused by passing a metallic mass through the field. • A large mass (like engine blocks, axles, frames, etc.) passing through a significant portion of the field will achieve detection. • A small mass (small wheels, small trailers) will not disrupt the field enough to cause detection to occur.

For a system to correctly classify a vehicle, all the sensors must be activated in the proper order. Any sensor that is not activated when a vehicle passes by will cause an error in the system, and thus an error in classification.

Given safety factors, legal and other issues, most vehicles are operated in the center of the lane, typically defined by wheel paths that are visual or physical (ruts), or both. Because of their width, most vehicles occupy at least 50% of the lane width and travel in single file.

Motorcycles, when traveling in the roadway, may legally occupy any part of the lane that the operator chooses to use. Because of their width, they typically occupy less than 30% of the lane, and typically run near the center line or near the shoulder stripe.

Motorcycles:

• Running near the center of the lane will miss the piezo sensor in a typical classification system, causing an error, and will not be classified. • Running near the center line or near the shoulder stripe will not put enough metal mass into the loop field to cause the disruption necessary for detection. This causes an error in either system, and classification will fail.

Why not change the “layout” of the loops to increase the size of the detection field to pick up motorcycles that travel at the edge of the lane?

As noted in the attached drawing, loops are typically centered in the lane, and for maintenance and operational purposes, are placed directly across from each other in adjacent lanes.

Loops:

• Generate a cylindrical magnetic field from each of their “legs” (sides and ends). • Are sized and placed in a lane so that the bulk of the magnetic field will “cover” the central part of the lane, but not overlap the adjacent lane.

A rectangular shape for the loop is used so that the resultant magnetic detection field is also rectangular in shape. This “rectangular” field is formed by cylindrical shaped fields generated in each “leg” of the loop. The cylindrical fields generated by a loop have, in theory, a diameter equivalent to the distance between parallel “legs” in the loop. Thus, a 5 foot by 7 foot rectangular loop would generate a theoretical magnetic field with its outside edges forming a rectangle that measures 10 feet by 14 feet. The interior edges of the cylinders all meet at the center of the loop, while the outside edges form the boundaries of the magnetic field at the dimensions listed above. Again, this is the theoretical size and shape of the field, but there are many exterior factors that can affect the field.

Some of the factors that affect the size and shape of a magnetic field are: • Metallic reinforcement material in the pavement (typically in concrete--rebar, metal netting, metal dowel pens, etc.) • Minerals that are part of the aggregate used in the making asphalt, concrete, and the road base. • Depth of the loop within the pavement. (Think of a barrel floating in a lake that slowly fills with water and begins to sink. As the barrel sinks, its visible width decreases.)

As might be expected, the strongest, or densest, part of the magnetic field lies closest to the physical boundaries of the loop wires. Vehicles traveling legally in a lane will cross some physical part of the loop, and place the bulk of their metallic mass within the stronger part of the field and create the necessary disruption to cause detection. Motorcycles that pass through the physical boundaries of a loop may also disturb it enough to cause detection, but as noted before, motorcycles that pass through the edge of the detection field may not disturb it enough for detection.

However, vehicles passing through the edge of the detection field can put sufficient metal mass into the field to cause detection. This is why loops are sized so their fields do not exceed the boundaries of the lanes. Vehicles that drive legally in the lane, but are hugging the center line, can cause detection to occur in the adjacent lane if that adjacent field is sufficiently disturbed.

Physically increasing the size of the loop to occupy more of the lane width will: • Cause a proportional increase in the magnetic field that will overlap adjacent lanes. • Provide enough overlapped field area to cause vehicles driving legally in a lane to be incorrectly detected in the adjacent lanes, even if they are not hugging the center line. • Potentially cause electronic “crosstalk” between loops that are directly opposite each other in adjacent lanes

Staggering the expanded loops by moving them physically apart so they are not opposite each other in adjacent lanes will:

• Cure the electronic “crosstalk” between loops that are directly opposite each other. • Still cause a field to be generated that will overlap the adjacent lane, resulting in errors being generated by the detection of vehicles that pass in that adjacent lane. • Provide enough overlapped field area to cause vehicles driving legally in a lane to be incorrectly detected in the adjacent lanes, even if they are not hugging the center line.

Why do groups or “packs” of motorcycles cause classification problems with our systems?

Because of their small size, motorcycles that are operated in groups tend to ride abreast of each other (something other vehicles can’t legally do) as well as closer to each other in the direction of travel (tailgating distance).

Given the physical size of the loop-piezo configuration:

• It is possible for multiple motorcycles to be within the detection area of the sensors at the same time. • WIM or classification systems are not capable of separating multiple vehicles (including motorcycles) that are detected at the same time in a single lane of travel. • The systems see simultaneous detection of multiple vehicles in a single lane as one vehicle only. With motorcycles this means incorrect axle counts and axle spacing.

The results are no classification of any of the motorcycles.

Are other methods or systems available to correctly classify motorcycles and will they work?

There are other technologies available for classification of vehicles that don’t use loop-piezo sensors to detect and classify vehicles. Some of these technologies are: • Road tube (short term counters). • Infra-red and laser sensor detectors. • Acoustic sensor detectors. • Digital radar detectors. • Video classification systems. • Video tape. • Manual observations.

Given the three requirements needed to classify any vehicle within FHWA scheme F definitions:

• The beginning and end of a vehicle. (Identifying a single vehicle). • The number of axles contained within the limits of the vehicle. • The spacing between each axle.

Road tube, infra-red/laser, and acoustic sensors are capable of satisfying the three requirements, just like loop-piezo systems. They are also subject to one of the same flaws: they cannot distinguish an individual vehicle when multiple vehicles pass through the sensor field at the same time in the same lane of travel. They do a better job of identifying single vehicles passing through their sensor fields because they do not rely on metallic mass to determine the “presence” of a vehicle. Road tube and infra-red/laser sensors detect the wheels (and thus the axles) of a vehicle. Acoustic sensors determine the location of the axles by identifying the noise the tires make as they travel on the pavement. All of these detectors cover the entire lane with their sensors, so the area of the lane occupied by an individual motorcycle does not necessarily cause a problem.

Digital radar and video classification systems cannot determine the number of axles a vehicle has. They can only determine the beginning and end of a vehicle, and they also cannot distinguish an individual vehicle when multiple vehicles pass through the detection zone at the same time in the same lane of travel.

Video tape produces the most accurate recording of the vehicles traveling in a lane, but requires someone to manually classify vehicles by watching the video tape. While capable of recording long periods of time (using surveillance-type equipment), this method, potentially 100% accurate, is extremely labor intensive. Reviewing multiple hours of video tape is still subject to human error due to fatigue or inattention.

Manual classification has the same pros and cons as video tape with the added problem of being a “one time” observation. Any potential error in observation cannot be reviewed, as could be done with a video tape. An additional problem is the duration of a manual count, which is again subject to human capabilities. Typically 4 hours or less, the statistical value of the manual count results in determining actual numbers of motorcycle traveling the road is highly questionable.

A Quick Summary

Classification systems are designed to produce accurate class data for all classes of vehicles. Motorcycles, due to their extremely low percentage of the overall volume of vehicles traveling, do not significantly impact the accuracy of a system. There is currently no vehicle classification system available (that I am aware of) that targets motorcycles only. Trying to adapt current available systems to target motorcycle classification still requires the ability of the system to identify individual motorcycles. As long as motorcycles are not restricted to single-file operation in a lane of travel, accurate individual identification will not be possible with the current technologies used.

South Dakota Comments

PDF 3 pages: http://dmses.dot.gov/docimages/pdf100/455092_web.pdf

4.3 Data Items

Motorcycles

The South Dakota Department of Transportation currently collects motorcycle data from our Automatic Classification Sites, so reporting the data in the summary data table by 2008 will not be a problem. We would like to point out that there are equipment issues in collecting motorcycle data from road tube counters and Weigh in Motion sites. We have found that there is a problem with counting motorcycles traveling side by side or close behind each other. As the size of motorcycles increases, the difficulty of distinguishing motorcycles from automobiles using automated collection methods also increases.

Texas Comments

PDF: 11 pages

HYPERLINK "http://dmses.dot.gov/docimages/pdf101/47http://dmses.dot.gov/docimages/pdf101/474172_web.pdf

Page 5:

4. New Safety-related Data Items

4.1 Motorcycle travel data

TxDOT started collecting motorcycle data I April 2007 with our Automatic Classification Sites. We would like to point out that there are equipment issues in collecting motorcycle data from road tube counters and Weigh in Motion sites. We have found that there is a problem with counting motorcycles traveling side by side or close behind each other. As the size of motorcycles increases, the difficulty of distinguishing motorcycles from automobiles using automated collection methods also increase. We are converting continuous volume (ATR) sites to classification every year, but this is costly and time consuming.

Texas has the capability to collect limited motorcycle data using accumulative count recorders and manual count data collection. At this time, we restrict these programs to Monday through Thursday. Assuming that the contractors can handle an increased workload, adding data collection for Friday through Sunday would result in a several million dollar increase each year for existing data collection contracts. TxDOT will need to explore using non-intrusive equipment to supplement our existing infrastructure.

NOTES:

California: http://dmses.dot.gov/docimages/pdf101/474775_web.pdf

Georgia:

http://dmses.dot.gov/docimages/pdf101/474854_web.pdf

http://dmses.dot.gov/docimages/pdf101/474703_web.pdf

Hawaii: http://dmses.dot.gov/docimages/p101/474743.pdf

Illinois: http://dmses.dot.gov/docimages/p89/458145.pdf

Florida: http://dmses.dot.gov/docimages/p101/474376.pdf

Michigan:

http://dmses.dot.gov/docimages/p101/473279.pdf

http://dmses.dot.gov/docimages/pdf101/474693_web.pdf

Minnesota: http://dmses.dot.gov/docimages/pdf101/474662_web.pdf