Micro-Weather Data Fusion Can Accelerate Fielding of Autonomous Transportation Systems
December 18, 2016
by Don Berchoff
This week the Unmanned Systems Institute Conference on unmanned aerial vehicles, and semi-autonomous and autonomous vehicles was held in San Diego, CA. The conference was more akin to a workshop, where industry technologists with car and unmanned aerial vehicle equipment manufacturers and sensor manufacturers; guidance and communications systems experts; investment bankers; test facility operators and other experts focused on the unmanned systems industry. The workshop covered unmanned systems technology, growth outlook and schedules, international and domestic uses, investment opportunities, application to commercialization, technology challenges and future uses. This was a high-powered group of innovation leaders and technologists, which included representatives from Tesla, Daimler, Boeing, Mercedes-Benz, BMW to name a few.
The industry milestone that drew the most discussion was the prediction that fully autonomous ground vehicles will operate by 2021, transporting the general public and workers in semi-controlled areas such as work and college campuses, large work sites, government installations and other such environments. I spent considerable time listening to sensor presentations and talking with robotics experts assessing the sensor and guidance systems that will guide the vehicles safely. My assessment is that while ambitious, the goal of deploying the vehicles in controlled environments by early next decade are attainable with the right roadside infrastructure supporting vehicle on-board sensors, even in snow, ice, heavy rain, and ponding water conditions.
Google, Apple, Mercedes Benz, Daimler and others are making huge investments to bring the technology to market as rapidly as possible. My experience and expertise tells me costs and risk to commercializing and operating these vehicles safely in cities is manageable. But my enthusiasm is guarded as it appears society—the general public, lawmakers and policy makers–may lag behind the industry in accepting the notion of self-driving vehicles at the accelerated timeframe discussed in the workshop. This is somewhat evident in the latest news out of California in the article link below.
Additionally, there are still technical hurdles to fielding a true autonomous vehicle without roadside guidance infrastructure by the mid-2020s, as some manufacturers believe. The sensor technologies tested on autonomous vehicles today will periodically degrade in certain weather conditions during circumstances where vehicles can operate today with a driver in relative safety. There will be use cases or areas of the country where vehicle-borne sensors will suffice, but without the investment in some roadside infrastructure, certain significant weather conditions will impact the widespread use of reliable self-driving vehicles.
But whether we are talking about semi-autonomous vehicles with driver override, autonomous vehicles with only onboard sensors, or autonomous vehicles with roadside infrastructure support, micro-weather sensing is paramount to a safe and productive ecosystem that will drive consumer confidence in embracing the technology. Derived data from vehicle computer and telematics systems, today’s onboard temperature sensors, and crowd-sourced human reporting is not accurate enough to ensure comprehensive micro-road segment situational awareness and road forecasts. A high fidelity, high-density road and weather sensing ecosystem, especially coupled with a road side guidance infrastructure, provides the best chance for autonomous vehicles to operate in any weather within the normal physical limits vehicles can operate today.
High-density road and weather sensing will also significantly improve road segment weather and road forecasts at a magnitude better than anything available today. And forecasts will be critical for safe route and journey planning transmitted directly to onboard systems, to mitigate not only what is impacting the vehicle at that moment, but in potentially avoiding high threat areas likely to occur hours in advance through proactive re-routing and planning.
In a nutshell, with the right mix of weather sensors, forecasts systems and roadside infrastructure, it is plausible for a self-driving vehicle to successfully operate safely and effectively in a Boston snow event in January, or a heavy rain thunderstorm in Miami in the next decade. Driving down weather risk will enable quicker FAA and DOT action in producing the regulation needed to operate in the most autonomous matter possible, and in bringing the general public along in welcoming and embracing this life-changing technology.
TruWeather Solutions is a thought leader focused on the nexus of unmanned transportation systems technology and weather and water solutions. Don Berchoff is a meteorologist and was the National Weather Service Science and Technology Director from 2008-2012.
CEO – TruWeather Solutions
Cell: (703) 915-7639