The human toll from highway accidents is immense. It has been estimated that a total of 3,551,332 motor vehicle deaths have occurred in the United States from 1899 to 2012. Every year there are an estimated 1.2 million deaths worldwide in automobile collisions – over 30,000 in the U.S., over 240,000 in India, over 270,000 in China. Autonomous vehicles have the potential to dramatically reduce the carnage on the world’s highways.
Google has had a fleet of 10 or more driverless cars on the roads in Florida, Nevada, and California for a couple of years that by now have logged over a million driverless miles. The latest cars from mainstream vehicle manufacturers already include intelligent functions such as keeping in lane while adjusting speed to the vehicle in front, automated parking, and crash prevention.
According to a recent report form Navigant Research there is an industry consensus that the first vehicles with comprehensive self-driving features will be brought to market by 2020. The technology to support autonomous vehicles such as image processing and sensor fusion are now ready for production use, and supervised decision-making software is in trial on public roads. Navigant Research forecasts that 94.7 million autonomous-capable vehicles will be sold annually around the world by 2035.
A new paradigm for intelligent highways
Autonomous vehicles are going to be a major beneficiary of "intelligent highways". Intelligent highways refers to high quality 3D digital models of highways that will be able to provide reliable data about highway infrastructure and its geographic context to autonomous and partially autonomous vehicles.
To enable these highly reliable highway models will require a new paradigm that needs to be data-centric and real-time. From a data management perspective it will require highway infrastructure lifecycle management, which requires a very different approach to the survey/design/construct/maintain process that is currently used for highway projects in North America.
Over the last 25 years the process of surveying, designing, and building highways has been increasingly automated – examples include electronic data collection, GPS, basic 3D laser scanning, design automation such as computer aided design and drafting, and some construction automation. The primary focus has been on speeding up the project development process, but the end goal has remained 2D paper (or digital paper) construction documents.
According to Ron Singh of the Oregon Department of Transport (DoT) the current way of designing and building highways has reached the end of its lifetime and a complete transformation to a new paradigm is required. The way Ron refers to this is standing the current process on its head and I have blogged about it previously.
Fundamentally it involves intelligent 3D models, during- and post-construction surveys and implementing a process that relies on using existing engineering data rather that requiring a complete resurvey at the beginning of every project. It implies that there is a reliable geospatially-enabled asset database that is maintained throughout the operations and maintenance phase of the lifecycle. From an autonomous vehicle perspective it enables maintaining high quality 3D models of existing highways that autonomous vehicles can rely on to provide an accurate real-time geographic context for their on-board sensors and analytics.
Developing a 3D model of a state highway system
Recognizing that it would take generations to develop a digital highway model for the whole system, as a first step the Oregon DOT are scanning the entire highway system with mobile LiDAR scanners and capturing the highway as it exists today. In addition to georeferenced point clouds they are also capturing imagery using a variety of reality capture technologies such as oblique imagery. This makes it possible for an engineer to visualize any stretch of highway on and to measure it accurately in 3D without leaving the office. Ron’s goal is to capture the entire state highway system as a combination of point clouds and oblique imagery. On top of this they overlay all of their asset information from the DOT’s GIS system. That enables visualizing all of their surface asset information in context very accurately.
In addition as they create 3D engineered (BIM) models for new projects with post-construction surveys,they will have more and more pockets of reliable engineered information appearing within this statewide database. When a new project is proposed the designers can query the database and if a 3D engineered project had been previously completed in the area they would know that this came from an accurate post-construction survey that can berelied on for future engineering. In this case the database would provide 80-90% of what the designers need to design the new project. It would be augmented by field reconnaissance including mobile scanning to verify that nothing else had changed.
Ron expects that autonomous vehicles will be data hungry and there will be immense benefit from knowing about the highway infrastructure and its geographic context. For example, if an autonomous vehicle is driving down the road it needs to know the roadway alignment and where certain things are. If there is an emergency it needs to know whether the vehicle can cross this line and go in this direction or is there a cliff or an obstruction.
This information cannot be sensed by on-board sensors alone -there needs to be information about the highway infrastructure and its geographic context available in real-time. Smart cars also have to sense what is changing, where other things are on the road. For example, there are cooperative objects like other cars anduncooperative objects like a cow that has wandered onto the road. Ron’s vision is of a 3D intelligent highway model that is not just used during construction, but is also used to operate and maintain the highway system and that includes providing real-time reliable information to increasingly autonomous vehicles.
The benefit to humanity is immense – dramatically reducing the 1.2 million highway crash fatalities that occur every year.