Office of Operations
21st Century Operations Using 21st Century Technologies
Table of Contents

Research, Development, and Application of Methods to Update Freight Analysis Framework Out-of-Scope Commodity Flow Data and Truck Payload Factors

Chapter 11. Review of Alternative Truck Payload Factors Methods

A number of methods to compute Payload factors were reviewed as shown in table 33. The methods, and their sources will be discussed in the subsections that follow.

Table 33. Current and alternative methods.
Truck Payload Factors (TPF) Method Notable Features Comments
Freight Analysis Framework (FAF) 2 method Different payload factors by Truck Size and Weight (TSW) Region. TSW regions by State of registration, not State of operation.
FAF 3 and 4 Method Payload factors, by 3 dimensions from Vehicle Inventory and Use Survey (VIUS). Also includes distance variable. 3 dimensions are not part of sampling frame. While shares are stated precisely, those shares ignore standard errors. This method increases standard error. Distance variable does not consider all Standard Classification of Transported Goods (SCTG) information and may not be properly applied. Even as computed, payload factors are fairly constant for miles > 100.
VIUS Revisited Ton-miles and miles by SCTG2 for each sample record. Distribution could reflect proper expansion sampling. Payload factors as sum of Ton-miles divided by sum of Miles. Payloads can be computed for each record by SCTG based on reported data. Expanded based on sampling frame.
CA-VIUS Reporting uses California Statewide Freight Forecasting Model Commodity Groups. More recent survey than 2002 U.S. VIUS. Compare findings with VIUS revisited.
Weigh-in-motion (WIM) alone Inferred empty weight. Payload uses this inferred empty weight subtracted from average observed weight. Does not include commodity detail. Inferred empty weight only.
WIM with Loop Inferences Add payload by inferred body type. Infers body type based on reported data. Can only do payload by body type. Possibly compare with, and modify, FAF, payload factors by SCTG.
WIM with Timestamp Video Add payload by confirmed body type. Can only do payload by body type. Possibly compare with and modify FAF payload factors by SCTG.
WIM with Enhanced Electronic Clearance/Electronic Logging Devices (ELD) Add payload by confirmed entry weight, commodity type, body type. Aspirational only, if possible could confirm empty weight, body type, commodity type (current information does not support SCTG2 codes. Current ELD/Electronic Clearances only report restricted commodities.).
Possible VIUS Replacement, VIUS Pilot Consistent with above. Integrated with Freight Performance Measurement system and National Performance Management Research Dataset. Performance Management System Rules. VIUS Pilot and Replacement suspended because no suitable methods could be identified.
Canadian VIUS Subject to Canadian operating rules by selected drivers; commodity detail added by driver by electronic box. Canadian economy and operational rules may not be directly applicable to U.S./FAF. Considered under Replacement of VIUS.
(Source: Federal Highway Administration.)

Freight Analysis Framework 2 Method

FAF1 was based on the commercial TRANSEARCH database which reports trucks, and thus payload factors were not needed. FAF2 used a method that apparently varied the payloads by SCTG commodity by TSW region. These TSW regions were applied to the information from VIUS. However, the State, as reported in VIUS, is the State of registration, which may not be the TSW State, the State in which a truck is operating, which would be required for variation by TSW region.

Freight Analysis Framework 3 and 4 Method

The current Freight Analysis Framework Version 3 (FAF3)/FAF4 methods have been discussed previously in chapter 2. The current method uses:

  • Distance ranges which show little variation.
  • Are only for the principal commodity carried and not for each individual commodity carried.
  • May have not been computed using the correct expansion factors as applied to the survey records.
  • Have standard errors that will differ by SCTG2 commodity, and that standard error will increase for each mathematical computation that is used. The payload factors in the current Freight Analysis Framework Version 4 (FAF4) use at least 3 computations.

2002 Vehicle Inventory and Use Survey Revisited

From the 2002 VIUS micro data, tons and ton-miles by SCTG2 commodity were calculated for each record, using the mileage expansion factor. With this calculation, it is possible to compute the standard deviations, counts, means, etc., and compute the standard error. The payload factors by SCTG2 could be computed by only one mathematical operation as ton-miles divided by miles for each SCTG2 commodity. In addition to minimizing the increase in errors, basing the payload factor on tons and miles, where changes in tons and miles can be validated against other sources, is a better option than using body type and truck configuration. Truck configuration and body type data are not easily available or on a consistent basis thereby preventing the ability to update payload factors.

California Vehicle Inventory and Use Survey

California Vehicle Inventory and Use Survey (CA-VIUS) microdata are not yet available, but the payloads by its commodities, which were aggregations of SCTG2 commodities, were provided to the project team and are shown in table 10. It is noted that CA-VIUS sampling frame was trucks registered in California that travel on California roads, and trucks registered in other States (according to the International Registration Plan (IRP)) that also travel on California roads. While CA-VIUS is more recent than 2002 U.S. VIUS, the calculation of payloads for CA-VIUS commodities was according to the truck sizes as used in the California Statewide Travel Demand model.

Weigh-in-Motion Alone

A number of researchers have proposed determining payloads factors using truck weigh-in-motion (WIM) data, which is required to be submitted by each State Department of Transportation. That research highlighted a number of issues. First the WIM data never reports the cargo of the trucks but only the weight distributions of observed trucks. The most frequently observed weights are often assumed to be the average empty load and the average full loads. This allows the inference of the cargo payload as the difference between the inferred fully loaded trucks and the inferred empty trucks. Trucks are differentiated by the number and spacing of axles which is used to infer truck type. WIM is an indirect set of observations which does not provide any method to determine the commodity of the cargo being transported. Relying on percentages of use by different commodities as in the VIUS will help determine the commodity distribution but doing so will ignore the errors that are associated with these percentages in the VIUS.

Weigh-in-Motion with Loop Inferences

The inferences of truck size, which is based on the number and spacing of axles, can be supplemented by an inference of the truck body type. The magnetic or other signature of the truck is detected by WIMs or supplemented with loop detectors, and that is used to infer the truck body type based on the signature detected. While these inferences have shown promise (Hernandez 2016), they do not include any information about the contents of the truck, i.e., the commodities being transported. Using the VIUS percentages by body type and size typically ignores the error associated with these percentages.

Weigh-in-Motion with Timestamp Video

The inferences of truck size and body type can be confirmed with video images. A problem is reconciling the observed video image with the inferred truck and body size. This requires that the video image and the WIM observations be of the same truck, perhaps by timestamping both observations. This method has been proposed and shows promise, but its practical widespread application has not been demonstrated. Additionally, the video cannot observe the contents of the truck and using percentages from 2002 VIUS will be problematic if the errors associated with those percentages are ignored.

Weigh-in-Motion with Enhanced Electronic Clearance/Electronic Logging Devices

Enhanced Electronic Clearances, which requires trucks with this required technology, can report the contents of the truck. Typically, only restricted commodities are reported, and to be useful in calculating payload factors by SCTG2 commodity, reporting would have to be expanded and all commodities reported using the SCTG2 codes. While Electronic Clearances are associated with a truck, ELD drivers records the Hours of Service by truck drivers. While these ELDs report only the restricted commodities that are permitted for a given driver, these could be expanded to report the SCTG2 commodities. Even if these methods could report the commodity being carried, it would still be necessary to associate a truck with a driver. If these issues could be overcome, and the SCTG2 commodity could be determined, to be useful in computing payload factors the bias in ELD usage also need to be addressed.

Vehicle Inventory and Use Survey Replacement

The Vehicle Inventory and Use Survey (VIUS) was the principal data source on the physical and operational characteristics of the United States truck population from 1963 through 2002. The survey was discontinued prior to the 2007 survey due to budget constraints. Since that time, State departments of transportation (DOT), metropolitan planning organizations (MPO), as well as many Federal agencies have had no other alternative than to use the outdated 2002 VIUS data. The need to update the payload factors that could be computed from VIUS was addressed by FHWA in the VIUS Replacement project, but no suitable methods could be identified. In sum, these efforts determined that the best course of action would be to pursue the traditional VIUS survey model.

Canadian Vehicle Inventory and Use Survey

Canada has deployed units in certain trucks that assist in collecting information similar to what was reported in VIUS. The commodity being carried must still be entered by the driver into a unit installed in certain trucks. Canada also uses the SCTG system and commodity information is being collected according to this system. However, the differences in operations and legal regulations, such as weight limits, cost controls, or Hour of Service rules, will probably mean that payload factors that can be computed based on Canadian data are not necessarily directly transferable to the U.S. as required by the FAF.