These points can then be mapped relative to one another and to a map of Portland streets as shown below.

Notice that the points above clearly indicate that the truck was first moving east on Sumner road and then turned right to head south on Montana.

Although the analysis above is pretty straightforward to the human eye, the task of identifying this route when thousands of data points are in question is beyond the capability of a single person or even a small group of people.  It was therefore necessary to create a computer-based algorithm that could quickly assign each data point collected to a street and indeed a small segment of a street such as a block.  This assignment was accomplished by combining the proximity of the point to nearby street segments with a latency factor to settle disputes in the case of points near intersections or overpasses.  The result of this placement process is an outline of the route that was taken by the truck and the amount of time required to complete this route.  A small example is shown below.

The street names have been left off of the map above for clarity's sake.  The route taken by the truck is outlined in red.  The route starts on the 6500 block of N Denver Avenue and ends on the 400 block of NW 8th Avenue.  The truck travels on N Denver Avenue to Portland Boulevard.  Then the truck takes Portland Boulevard to Interstate Freeway 5 (I5).  Travel continues south on I5 until the truck takes the Broadway exit and goes over the Broadway Bridge.  The truck turns right onto NW Glisan Street and then left onto NW 8th Avenue.

A travel time associated with each street segment and the route taken between each delivery/pickup stop is identified.  When possible the commodity category of the freight being moved is documented as well.

PILOT STUDY RESULTS

The pilot study was extremely successful in capturing data points.  During a few weeks of observation well over 600,000 individual position points were collected.  The assignment of data points to street segment required nearly eight hours of computer processing time.  Additionally, much data was collected regarding the commodity categories carried by the trucks in the study.  It is unclear although, how to correlate this commodity information well with the route information.

During the conduct of the pilot study some unforeseen obstacles were encountered.  These obstacles included the fact that GPS receivers do not work well without clear view of the sky such as when a truck traverses the lower deck of a two-level bridge.  Additionally, scheduling the download of data from the units was difficult at times because the participating trucks were occasionally not available at the time scheduled for downloading.  Consequently, some data was lost.

The managers of the two carriers that agreed to participate in the pilot study were satisfied by the lack of damage to equipment and the unobtrusive nature of the data collection process.

ANALYSIS

The data collected during the pilot study can be analyzed in many different ways.  The techniques demonstrated below are just a few that have been identified as valuable and many more exist.

		Regional Analysis
			Carriers often group deliveries and pickups according to geographic areas and then assign some number of trucks to each area on each day.  One truck may cover NW Portland while another covers NE Portland.  However, these geographic assignments are sometimes violated because of other factors.  One example that is often encountered is the lack of equipment at a customer location.  If a customer has no loading dock and the freight is too heavy to handle manually, then a truck with a lift gate must be dispatched to make the delivery/pickup.  The picture below shows the routes of five trucks over a two week period.  There was clearly some overlap, but in general the technique seems to have worked fairly well.
		Pickup & Delivery Analysis
			This picture shows the route taken by a truck on a single day along with the pickup and delivery stops that were made.  Notice that in the picture below the trip between each set of two sequential stops is indicated with a different color.  This allows for an understanding of the route selected to get from one location to the next and the time required to traverse the selected route.
		Transportation Zone Analysis
			The city of Portland has been divided into many transportation zones.  The number of stops made within each of these zones can be a telling indicator of economic activity and high usage of the roadways.
		Street Use Analysis
			The usage of the streets within the area can be estimated by the density of trucks utilizing those streets.  This information can improve the allocation of resources by increasing the visibility of highly utilized streets.  Notice how heavily Interstate 5 and Columbia Blvd are used.

CONCLUSION

There can be little question that the careful maintenance and expansion of the roadway infrastructure within the Portland area is critical to economic development.  However, in the absence of information regarding the utilization of this infrastructure it is difficult to manage the limited resources available for maintenance and expansion.  The objective of this project is to collect information regarding the true movement of commodities within the Portland metropolitan area in support of better management of these resources.