Weather Report with AV index
Today’s automobile needs a great deal of infrastructure to be useful. A major part of the infrastructure is the civil transportation system consisting of roads, traffic control, and static signage. More recently, an information layer has been added to this infrastructure which consists of dynamic signage, metering systems, cellular coverage (enabling mapping applications), and sensors (camera, fog, etc) for traffic optimization. To date, the focus has been to provide safe traction, traffic management, and safety information. Perception and decision-making tasks were left to the driver.
With Autonomous Vehicles (AVs), there is a very large challenge in accurate perception of the external environment, and even a larger challenge of in-context decision making. Perception is impacted by the fidelity of the sensors and their limitations. These limitations are further impacted by global effects such as quality of signage, weather/related road conditions and electromagnetic noise (from the environment or other AVs). After initial perception, the raw information must be interpreted by variables with semantic meaning (road, car, human,.etc.), and then the AV must interpret and then apply reasoning to these objects in the context of final path planning.
These tasks are exceedingly difficult, and there has been a desire to have the environment enable the AV function. Since lane following is such an important task, AV-enabled lane markings has been suggested as a viable solution to help AVs. In addition, until the perception systems can robustly discern the environment, high-definition fully labeled maps have been suggested as an answer. Thus, there is a cottage industry of suppliers that are attempting to deliver this information with semi-automated methods for labeling. Perhaps the biggest active area is the Connected Autonomous Vehicle (CAV) where RF technologies such as DSRC and 5G are used to communicate the static and dynamic environment to the AV.
As with any infrastructure investment(FCC article), there are open issues on source of funding, as well the chicken-and-egg problem of infrastructure investment and AV demand. The risks of this investment are further amplified because they may change radically based on the capabilities of the AV. The big system design issue with all of the infrastructure related solutions consists of efficiency of development and responsiveness to environmental changes (e.g., accidents, construction zones).
“Perhaps the most dynamic environmental change is the weather and associated deteriorating road conditions,” said William Mahoney, Director, Research Applications Laboratory at NCAR.
What infrastructure must be built to communicate weather and road conditions to the AV? How will sensors perform when ice and snow have accumulated on vehicle sensors either while parked or while operating?
As an example, today’s taxonomy of communication of weather developed by the National Weather Service (NWS) for the human population includes words like “light rain,” or “scattered clouds.” NWS has a mission to provide public weather forecasts, warnings, and advisories. Today, NWS has a policy to not provide tailored and detailed information to specific economic sectors, such as surface transportation. This capability is provided in the U.S. by a robust private sector weather enterprise which provides value-added information such as detailed weather and pavement conditions to state DOTs. This leads to a host of questions:
- Should NWS and/or private weather enterprise expand its services to cover AV requirements?
- Do we need an AV impact index for weather and road condition reports?
- Should NOAA’s Space Weather Prediction Center broadcast GPS disruptions caused by solar flares, and does this need to be extended to AVs?
- Will AVs need to be certified to operate in certain weather and road conditions, and how will that information be delivered to the vehicle to ensure the safe completion of a trip segment?
Overall, will an AV index be added to the heat index, wind chill index or boating index outlooks provided by your friendly weather forecaster ?