Weather Applications and Products Enabled Through Vehicle Infrastructure Integration (VII)
1. Executive Summary
1.1 Objectives
Vehicle Infrastructure Integration (VII) involves the two-way wireless transmission of data from vehicle-to-vehicle and vehicle-to-infrastructure utilizing Dedicated Short Range Communications (DSRC). VII will enable the development of applications designed to improve safety and increase mobility and efficiency along the nation's roadways. The overarching goal of the VII initiative from a weather perspective is for the weather enterprise (defined as all the public and private organizations that collect, process, and generate weather products) to utilize vehicle data to improve weather and road condition products and to provide those products to transportation system decision makers, including travelers.
The utilization of data from mobile platforms is not new in the weather community, as ship-based observations have been used for more than a century. Data from aircraft have been used successfully for nearly a decade, and the number of parameters available from aircraft is expanding from primarily wind and temperature to humidity, turbulence, and icing. The utilization of data from vehicles poses significant technical challenges, particularly with respect to data quality; nevertheless, VII represents a technology that will significantly increase the density of weather observations in the Atmospheric Boundary Layer (ABL).
This report assesses the feasibility of using VII-enabled data to enhance road weather products. This is accomplished by providing a fundamental understanding of the weather-related vehicle data elements currently available on production vehicles, as well as anticipated data elements. Additionally, the potential contribution of these vehicle data to the diagnosis and prediction of weather-related hazards is examined. The report identifies technical issues and challenges related to the use of vehicle data, and it outlines a comprehensive list of research topics that will be essential for effective use of weather-related vehicle data. Finally, the report summarizes the feasibility of using VII data for road weather product development and improvement, and provides recommendations that will help ensure successful exploitation of vehicle probe data in weather applications.
1.2 Approach
The National Center for Atmospheric Research (NCAR) held two VII Weather Application Workshops to bring together experts from the automotive industry, VII Consortium, VII architecture and probe message processes development team, and the weather community. The overarching goal of these two meetings was to discuss and gather input related to the proposed VII architecture, probe message processes, potential data elements, possible road weather products, and the technical and scientific challenges associated with the use of these data.
The authors of this report attended several other conferences and symposia, which provided the opportunity to interact with different groups working on VII-related projects. In addition, the forums were used to update the VII community and others on the progress being made regarding the development of this report.
An extensive literature review was conducted as part of the report development process. The primary goal of the review was to acquire a deeper understanding of the envisioned VII architecture and probe message processes, as well as identifying potential weather-related vehicle data elements.
1.3 VII Deployment
The three principal elements that make up the VII system architecture are the On Board Equipment (OBE), Roadside Equipment (RSE), and the VII Network. Each vehicle that is a part of the VII system will be equipped with an OBE. The OBE contains the On Board Unit, which is responsible for transmitting and receiving data. Vehicle-to-vehicle and vehicle-to-infrastructure communication will be supported. The RSE, which will be located along the roadside, also will be capable of transmitting and receiving data. Public data received by the RSE will be passed along to the VII network and made accessible to data subscribers.
Two critical components of the VII deployment and implementation plan include RSE deployment and probe message processes. The strategies used to site RSEs, gather vehicle data, and transmit these data to the VII network are very important in terms of improving and generating road weather products.
The currently envisioned plan for siting RSEs includes criteria for both urban and rural deployment. Most importantly, the minimum spacing between RSEs will be 10 minutes at 60 mph for rural regions and 2 minutes at 20 mph for urban areas. It is likely that the RSE coverage resulting from these criteria will be adequate for application development in urban areas, but it may lead to the loss of data in rural regions. Data loss in urban domains will likely be offset by the shear number of vehicles transmitting data to RSEs. Moreover, ample RSE coverage will potentially exist across the eastern U.S., while the western states will likely experience significant gaps in RSE spacing.
The feasibility of using VII-enabled data in weather-related applications will be affected by two key aspects of the VII probe message processes: onboard buffer storage limitations (currently 30 snapshots) and snapshot hierarchy. Snapshots, which contain vehicle data elements valid at a specific point in time, are generated in one of three ways. Periodic snapshots are produced at prescribed intervals based on the speed of a vehicle, event-triggered snapshots are generated when the status of selected systems change or a predetermined threshold is met (e.g. ABS status change from "off" to "engaged"), and start/stop snapshots occur when a vehicle comes to a stop or begins to move. Note that all snapshots (whether event-driven, start/stop or periodic) contain all available weather elements. Event-triggered and start/stop snapshots have priority over periodic snapshots. This is an important aspect because under certain situations periodic snapshots could be deleted from the onboard buffer in favor of event triggered and start/stop snapshots. A set of periodic snapshots is more likely to contain correlated data that represent environmental and road conditions across a broader domain as compared to event-triggered and start/stop snapshots. Therefore, the deletion of periodic snapshots because of snapshot hierarchy could lead to the loss of important data.
Other aspects of the probe message processes that will limit the availability of vehicle probe data relate to privacy concerns. They include not allowing any identifying information associated with the vehicle or the vehicle's operator to be transmitted as part of a snapshot, restricting snapshot generation for a certain distance after a vehicle is started, and deleting all snapshots when the vehicle is turned off.
1.4 Prospective Data
A number of data elements that could aid in road weather product improvements have been identified. These elements range from direct measurements of the environment (e.g. temperature) to elements that could be used to infer weather and road conditions (e.g. ABS). However, a significant amount of research will be required to fully understand how to effectively use vehicle-based data, as the characteristics of the data will vary greatly between vehicle manufacturers, vehicle models of the same manufacturer, and sensor types and models. It is unlikely that any single vehicle-based data element will be able to stand alone as truth, as there will be too many uncertainties about its quality and/or representativeness. Vehicle data will need to be processed in a statistical manner to address data outliers and to raise the overall confidence in data quality. The weather community has experience combining multiple disparate datasets to derive products. Vehicle data will have to be treated in a similar manner. Even with those caveats and concerns, vehicle data will result in the generation of improved weather and road condition analysis and prediction products because of the large volume of data, distribution of observations, and frequent update rates.
1.5 Data Efficacy
Data from vehicles operating in the Detroit Metropolitan region were acquired from DaimlerChrysler. These data contain information on vehicle location, observation time, wiper state, barometric pressure, and air temperature. Vehicle data elements were compared to NEXRAD Doppler weather radar data and observations from surrounding stationary weather stations. These case studies were promising, as a good correlation between vehicle data elements, radar data, and surrounding surface observations existed. In one case, vehicle data responded to an outflow boundary that was generated by a nearby thunderstorm. These case studies support the idea that vehicle data could play an instrumental role in enhancing road weather applications and products.
1.6 VII-enabled Improvements
A number of weather and road products will be improved with vehicle data including improvements in both the diagnoses and prediction of weather and road condition hazards. Examples of improvements are supplied in Table 1.1, but they only provide a sampling of what may be possible with these new datasets.
VII technology will not only enable vehicles to communicate probe data to external systems, but it will enable safety and mobility related products to be delivered to vehicles. When drivers enter their vehicles, they are usually cut off from normal information sources such as television and the Internet, and until cell phone technology became dominant, phone service was unavailable. Wireless communication technologies are becoming more reliable, more widespread, and less expensive, which provides an enormous opportunity for the automotive, consumer electronics, and telecommunication industries. The widespread adoption by the public of wireless vehicle technology is just around the corner. In-vehicle weather and road condition products that are likely to be of the most interest to drivers are presented in Table 1.2.
1.7 Data Processing
The amount of data that could potentially flow through the VII infrastructure could be immense. It is likely that many prospective users will not be capable of handling the vast quantities of data that are expected. Additionally, it is unlikely that some users will be able to contend with the complexities associated with the data, such as data quality, representativeness, and format. Applications (middleware) will need to be implemented that will facilitate the use of VII data. Without such a function, the feasibility of utilizing vehicle probe data will be lower and there will be substantially more risk in its use.
One solution for addressing this issue is to utilize a Weather Data Translator (WDT) to preprocess weather-related vehicle data before they are distributed to data subscribers. Raw data would still be made available to users that wish to obtain unprocessed data. The WDT would be made up of three primary components: filtering, quality checking and translation. The proposed WDT would acquire, parse, and process vehicle probe messages. Prior to a quality checking process, data would be filtered to extract data that are known to be unrepresentative.
Standard quality checking algorithms (e.g., outlier, format, bounds, and spatial tests) as well as quality checking algorithms specifically developed for VII data would be included as part of the WDT. Finally, the WDT would also demonstrate the ability to translate VII data into statistical samples that represent chosen parameters over a selected region and time.
1.8 Data Fusion
The lack of consistently accurate road weather hazard products is the result of several factors including a limited surface observation network, an inadequate understanding of the physical mechanisms responsible for some road weather hazards, and a need for improvements in weather and road condition modeling, particularly with respect to clouds and precipitation. VII will facilitate improvements in all of these areas; however, due to the complexities and uncertainties associated with VII-enabled data, several challenges and deficiencies must be overcome. Therefore, it is important that advanced data fusion techniques be developed and implemented to maximize the benefits of VII data for road weather applications.
1.9 Research Requirements
VII holds considerable promise in supporting the development of weather-related products for the surface transportation community. However, it is evident the use of VII-enabled data for product development and enhancement will be beset with challenges. In order to make effective use of mobile data for weather-related applications, it will be necessary to invest in research to understand issues associated with the use of current and anticipated data elements. VII-related research and development topics that are necessary to support the development and improvement of weather-related products include, but are not limited to:
- Data Fusion Techniques
- Data Quality and Accuracy
- Quality Checking
- Human Factors
- Numerical Modeling Using VII-enabled data
- Data Processing
The research needs are discussed more thoroughly in the main body of this report.
1.10 Conclusions
The authors believe that it will be feasible to utilize VII-enabled vehicle probe data in the generation of weather and road condition products and that these new datasets will improve roadway safety, mobility and efficiency. However, to ensure the success of the VII initiative from a weather perspective, several recommendations are provided. They include: (1) having experts from the meteorological community take an active role in helping to guide selected aspects of the VII program such as the proposed strategies for RSE deployment and probe message processes; (2) developing a Weather Data Translator (WDT) to facilitate the use of weather-related vehicle probe data; (3) Investing in research that will support effective use of current and anticipated weather-related vehicle data elements; (4) refraining from attempting to use weather-related VII data elements as stand alone truth; (5) conducting extensive research on the vehicle data elements of interest in an effort to ensure the proper use of those data including, if possible, collaborating with multiple OEMs that design and implement the sensors or devices from which the data originate; and (6) initially targeting basic applications and products that can be improved or constructed with rudimentary vehicle data elements.
