The Institute of Transportation Studies at the University of California, Irvine, recently developed the California Vehicle Activity Database (CalVAD) for ARB. CalVAD merges Caltrans’ raw vehicle detection system (VDS) volume and occupancy data with Caltrans’ weigh in motion (WIM) data. The VDS data are collected approximately every half mile on urban California highways every 30 seconds, and the WIM data provide weights and measures for every truck at just over 100 detector stations scattered throughout the state on major truck routes. A line layer is required to properly merge these data sets together. In the absence of official GIS network available from Caltrans, CalVAD incorporates a simple highway network using the open, freely available OpenStreetMap data, leveraging the route relation tags in those data. In this way vehicle measurements from the two data sources were merged with a highway network layer to form a picture of roadway vehicle activity. Missing in the current version of CalVAD is information on non-highway vehicle movements. The original development of CalVAD incorporated a rudimentary form of the HPMS data set into CalVAD. Although the 2007 HPMS data are loaded into the CalVAD database, and can be queried along with other data, these data do not have any geographic referencing associated with each link. This means that straightforward bounding box and area queries that rely on the PostGIS extension to PostgreSQL can’t be performed. (It is noted that Caltrans once had a linear referencing system for these data, but their license was subsequently revoked by the commercial vendor.). The first task in this proposed effort will wire up the HPMS records such that geographic queries can grab the right link-level data. This task is described as the first task of the project, and will be performed without requiring or referring to any GIS or network layer, but rather will use the information in each HPMS record combined with known information about geometries (cities, counties, air basins, and so on). This will provide a coarse geocoding of the HPMS data, but probably not good enough to produce something like gridded estimates of vehicle activity. The second task, and the bulk of the project hours, will be devoted to assembling street level geographic data, and then linking this with the HPMS records. We propose two different options for this task for ARB to consider: 1) to use OpenStreetMap data once again, or 2) to turn instead to regional planning models. The final task of this project will be to identify new and different sources of information for arterial vehicle flows that may be incorporated into the CalVAD system. The HPMS data may be the best currently available, but estimates of average annual daily traffic on a link is a few steps removed from storing the raw measurements that went into those estimates. The goal of CalVAD is to produce a portrait of vehicle activities that is reasonably accurate in time and space, but that is firmly grounded in observed data. This final task will advance the goal of entering lower level data on arterial travel, and will also identify commercial sources of information, such as Google, and provide ARB with analysis on the potential of these information sources going forward. By anticipating new kinds of data and new ways of collecting data, we can keep CalVAD relevant and perhaps even grow it into a player that shapes these new and rapidly evolving technologies.

The Institute of Transportation Studies at the University of California, Irvine, has recently developed for ARB a tool that merges data from different data sources related to roadway vehicle activity. The predominant sources of data are the Weigh in Motion (WIM) system and the Vehicle Detection Sytem (VDS). These data sources produce huge volumes of data from detectors scattered throughout the state. The tool developed pulls in data from both of these sources and merges these sets based on time, geographic proximity, and statistical imputation techniques.  The purpose of this proposal is to complete the project. There is still a small amount of work remaining in the development of the web-based graphic user interface, some work left on the data fusion task, and finally the important training and documentation tasks. The work proposed here is within the scope of the original proposal, and the total budget requested does not exceed the amount remaining from the original contract.

The Institute of Transportation Studies at the University of California, Irvine, recently developed for ARB a tool that merges vehicle measurements with a highway network layer in order to form a picture of roadway vehicle activity. This system is called the California Vehicle Activity Database (CalVAD). CalVAD is an excellent tool as it stands, but it is missing any information on non-highway vehicle activity. This proposed project would add arterial traffic flows to CalVAD. The initial focus will be to add HPMS data to CalVAD that will enable straightforward bounding box and area queries that rely on the PostGIS extension to PostgreSQL to be performed. Although the 2007 HPMS data are loaded into the CalVAD database, and can be queried along with other data, these data do not have any geographic referencing associated with each link. The key technical hurdle will be addressing the lack of georeferencing in that data set, as well as anticipating possible successors to HPMS data that might become available. Arterial roads are where the majority of roadway vehicle activity takes place, but there is scant measurement of these flows. The primary barrier is the cost of measurement. Caltrans can justify putting loops into the state highways and installing weigh in motion stations at key locations, but local traffic authorities cannot justify installing count stations on every road. Even producing counts for the annual HPMS survey is limited to major arterials, with minor arterials and other roads allowed to be sampled every 3 years or more. And yet, without deploying any hardware, Google has managed to produce reasonably accurate estimates of link speeds on many of the major arterials in the state by relying on cell phone position data from mobile search queries. We want to anticipate this kind of new data, so as to allow CalVAD to grow and remain relevant in the future.