Weather Applications and Products Enabled Through Vehicle Infrastructure Integration (VII)

Appendix A

VII Weather Applications Workshop I
Boulder, Colorado
Wednesday, February 22, 2006
Meeting Notes

NCAR, with support from the Federal Highway Administration, is investigating the possibilities of deriving weather and road condition information from vehicle data elements. The benefit is enabling tactical and strategic response to weather related surface transportation hazards. The objective is to produce a feasibility and concept development report for VII.

A presentation by the Federal Highway Administration (FHWA) on the VII Initiative and Day 1 Applications – The Clarus Initiative - was given, with discussion on the following topics: motivations for the Clarus Initiative, VII's impact on overcoming gaps and decision making, an overview of VII program activities, VII use case development, the purpose of use cases, an overview on VII program activities, use case development schedule, roadway environmental condition information, winter maintenance, traffic management (potholes), and roadway environmental condition information (traveler notification with and without Clarus).

Additional comments:

• The end result of this research is not data but rather relevant information, i.e. taking raw data, assigning relevance and then converting it to useful information. Some public data sets will be available; Clarus is much more focused. Data can be obtained from publicly owned assets and we are also working with many private entities. The operational entity to manage VII and to design the architecture is to be determined. Security is a big issue.
• A clear vision on how VII data is to be used is needed; the process could easily take a direction not intended. Privacy principles are a consideration. If VII is perceived as enforcement information it will more than likely not be accepted by the public.
• A set of data exists which can be transformed into benefits for society – this is the crux of the surface transportation program. We have an obligation to protect goods and people going from point A to point B within a reasonable time range. To accomplish that, we need data; the Clarus system manages that data.

A FHWA architecture overview was given on VII with emphasis on: the program overview, SDNs and NAPs, supporting a variety of network services or applications, providing data transport in support of both public and private services, vehicle probe data, vehicle probe data collection, and vehicle probe data distribution.

Additional comments:

• Commercial vehicles are likely to become the earliest users of VII. Each commercial user makes use of data in a different way and that may or may not include a standardized data bus. The controller area network is a challenge.
• A test bed is to be implemented by the end of April with the framework to be determined. Warning of severe weather occurrences, such as tornadoes, could potentially be of high significance to the customer. The common thread is what transpires longitudinally along the roadway surface.
• Vehicle sensors can fill in surface data gaps and help fine tune products. A greater number of observations will help the weather prediction effort. Upper air information can be obtained from aircraft. Related techniques in terms of data utilization could be use for VII and the surface transportation community.
• Regarding secure technology, there is concern about web traffic. There is also concern about how many RC's are going through the server; 250,000 come from the DOT. The VII system delivers public and private data through the network of SDN service routes and private services through NAPs. Roadside equipment was first used in urban areas to foster safety at intersections.

Discussions took place concerning the type and placement of sensors in Michigan and whether or not they can be taken to the level of RFP. There is currently a small network with capability to grow but that is uncertain at the present time. The citizen driver should be able to decide what options to use and that would be controlled. Three new processes are involved: 1. safety, 2. mobility and 3. commercial processes. Observations generally fall into 3 categories:1. public domain data, 2. probe data from vehicles in range of RSE and 3. public safety messages. Private data is confidential – encrypted. Vehicle probe data is constructed by the OBEs and will contain the following data: periodic data – snapshots, event data – traction control, anti lock brakes

Cost is a consideration for car companies. This is not meant to be an "all or nothing" scenario; the plan is to proceed slowly.

Discussion on communication between vehicles and a fixed infrastructure. User does not mean "driver" but "application."
RSE broadcasts every 100 milliseconds – RSE public message priorities are

1. Local safety applications
2. TOC advisory messages
3. OEM diagnostics and safety notices
4. Commercial services
   

A presentation on acquisition and utilization of vehicle data elements was given and the following areas were discussed: Use cases, Use case scenarios, VII OBE Hardware Architecture, Present OEM Traffic (Incident, Weather) Information Collaboration, Collect and Distribute Probe Data, Provide Off-Board Navigation, OBE hardware architecture, human machine interface, and the mechanism for collecting.
For the proof of concept, vehicles that can support the system will be designed and chosen - a fleet of about 100 vehicles fully equipped with all on board equipment. Design experiments will commence in the beginning of 2008 and last about a year. It will be available to automakers by approximately 2011; however these are very rough timeframes. It will be a difficult interaction to package weather information and set it up.

A presentation was made on vehicles as mobile sensing platforms for critical weather data and discussed were task objectives and timeline, collaboration with NCAR, the Dulles toll road instrumented corridor, sensors, test vehicles, sensor placement, sensor maintenance and calibration, data collection, mobile data samples, fixed sensors, the Dulles Airport ASOS, NWS Doppler radar, road domain translation to radar reference frame via GIS, precipitation estimates on road segments, establishing ground truth, temperature data time series, heavy traffic occurrences, average temperature statistics, average variation by sensor type, mobile versus in situ temperatures, proposed data comparisons and a deliverable report.

A presentation on Weather & Road Condition Product Improvements Enabled by Vehicle Infrastructure Integration included the following: data fusion – road weather impact products, weather improvements enabled by VII, radar based precipitation identification, diagnosis of precipitation type, identification of foggy regions, improved high resolution modeling, defining atmospheric vertical profiles, boundary layer characterization, road condition improvements enabled by VII (some examples), winter maintenance operations, road condition reporting, and surface temperature gradients.

Additional comments:

• VII enables tactical and strategic decisions to be made on how we drive our vehicles. The scale is large to small. The challenges lie in how data is conveyed to the community. Placement of sensors and sensor type can affect the quality and accuracy of the data. The problem in accounting for biases is vehicle knowledge and privacy. The suggestion is to stay on the conservative side, work with more data and get a sense for the numbers. VII has filled in gaps from point to point. Many of the data needs to be moved downstream to the network.
• An observation is simply just that - an observation. It is probably a mistake to try and average observations. We may need to do some data volume reduction – compression of data. Each sensor and observation is processed in a specific way depending on how data is to be used, i.e. wind.
• There is more accuracy collecting data from urban areas. Rural vehicles are typically not equipped to make applications work in all environments. The accident record for rural versus urban was noted; there are more fatalities in rural areas – speed is a factor.
• Privately owned vehicles are needed for spread of data. The reality is that most of the commercial companies have sensor capability but using commercial vehicles as high quality sensors, identifying sources of data, sharing is not feasible. A lot of traffic activity goes on between midnight and 6 a.m. and gaps need to be filled.
• VII would have provided differences in temperature between Denver proper and the mountain areas. In mountain areas, the number of observations would be increased.

DOTs would love to have access to information from the vehicles. Many data will be duplicated, some weather related and some not. The possibility exists for getting information related to blow-over phenomena moving with lateral force, such as high wind. Tornado prediction should be added. There is a target audience and a large emotional impact on the public. The data can also be used elsewhere. Predicting paths of tornadoes can be quite a challenge. The usefulness of information to the driver and cost considerations are at the forefront.

A second meeting is being planned, possibly at Chrysler Corporation, Detroit. An effort will be made to keep the group small.