The vision of driverless intelligent vehicles on an Automated Highway System is compelling, but seems always to be decades away. However, benefits can be had today from incremental technology improvements, as demonstrated by Con-way's initiative to equip 1,300 trucks with driver alert and truck control technologies.
Full Article Below -
In 1962, researchers at Ohio State University built what was probably the first automated vehicle with steering, braking, and speed controlled not by a human, but by a built-in computer (the electronics filled the entire trunk, back seat, and most of the front passenger seat!). In the ensuing 50 years, there have been an astonishing array of technologies (e.g. location-based services), standards (e.g. DSRC), conferences, consortia and industry groups (e.g. Intelligent Transportation Systems Society), projects (e.g. EUREKA Prometheus Project), and contests/trials (e.g. DARPA’s Grand Challenge and VisLab’s Intercontinental Autonomous Challenge) aimed at fulfilling the vision of fully autonomous vehicles and an intelligent highway system. In this vision, cars and trucks take people and goods from origin to destinations without any human driver. They self-organize into 'platoons' (bunches of 10-20 vehicles) driving at highway speeds with only a meter or so separating each vehicle—made possible by the tight integration of their systems and the instant response of computer controls. This would dramatically improve vehicle efficiency and mileage, increase the capacity of our highway system approximately 5X (see "Coping with the Transportation Crunch”), and practically eliminate accidents, saving many tens of thousands of lives each year.
How to Build a Driverless Vehicle
Broadly speaking, there are two approaches to creating driverless vehicles: A) build completely autonomous vehicles, which are not reliant on any special infrastructure or on 2-way communication with other vehicles, B) build smart highways and put interoperable communications in all participating vehicles. Approach A is attractive because it would not require wholesale upgrade of the transportation infrastructure or national vehicle fleet, but it is a more difficult engineering challenge. Approach B, while easier to design, has the chicken and egg problem of who is going to buy those special cars and trucks before the special highways are built, and which politician will approve funding for a very expensive highway that none of today’s car can take advantage of.
Thankfully, there is an approach C, which is the gradual introduction of increasingly sophisticated driver-assist features, taking us, in steps, closer to the fully autonomous vehicle.
Sounds great... when can I get one? Well, not today... and probably not for many years (perhaps decades) to come. The engineering and economic challenges are considerable (see sidebar). However, there are incremental technologies available today that can yield value without having to clear high interoperability or economic hurdles. There are many examples. One recent commercial implementation is Con-way Freight’s (LTL division of Con-way) purchase of 1,300 new Freightliner Cascadia tractors equipped with rollover stability, front collision warning with adaptive cruise control, and lane departure warning systems.
Studies have shown that front-end collisions, lane changes or departures, and rollovers are the most common causes of commercial motor vehicle crashes. So it is no accident that Con-way selected these three technologies. The Forward Collision Warning uses adaptive cruise control to maintain a safe following distance and forward-looking radar to detect a potential collision and alert the driver with audio and visual alarms. Lane Departure Warning monitors the truck’s position relative to lane markers and sounds an alarm similar to the sound a vehicle makes when it travels over a highway 'rumble strip,' alerting the driver to make a correction. Roll Stability Control senses driving conditions consistent with a vehicle about to roll over, such as hard cornering or change of direction. It then alerts the driver and automatically decreases the engine torque.
This was not an insignificant investment—over $5M—adding over 5% to the cost of the tractors. We assume that Con-way did their due diligence and concluded that this was a good investment. Accidents are expensive, not only in terms of damaged property and insurance costs (medical, disability, and property) but also in human terms (injuries and fatalities) and reputational costs.
While Con-way selected these technologies primarily for their safety benefits, they contain some of the core elements which will ultimately be required for the driverless vehicle—technologies that sense when the vehicle is getting too close to another vehicle or is about to roll over, and correction technology that adapts the driving of the vehicle (e.g. accelerating, braking, and reducing engine power). While it is a long way from the fully autonomous driverless vehicle, Con-way and others have shown that there is no need to wait for the ultimate driverless vehicle. The benefits of incremental technology, derived from driverless vehicle research, can be had today.
To view other articles from this issue of the brief, click here.