The National Highway Safety Traffic Administration attributes 94 percent of all crashes to human error, mostly associated with recognition and decision errors. Research and testing done so far on autonomous vehicles point toward a much safer world. But making vehicles smart enough to navigate an incredibly complex world is not happening overnight.
Fleet owners have arguably the most to gain from driverless vehicles. Trucks have been involved in 222 truck crashes per 100 million vehicle miles traveled since 2000. Plagued by human error including speeding, distracted driving, visibility problems, failure to yield right of way, and fatigue- or alcohol-related impairment.
Large trucks were involved in more than 415,000 police-reported crashes in 2015 alone, the last year comprehensive data was available. Non-fatal crashes cost companies on average more than $91,000 each. Fatalities can increase the average cost to $3.6 million per crash.
Some studies predict that the use of autonomous vehicles could bring that figure down from 222 to just 8 crashes per 100 million vehicle miles. Proponents of advanced driver assistance systems (ADAS) estimate that the technology could reduce accident fatalities by 44 percent. ADAS technology includes forward collision warning (FCW), headway monitoring and warning (HMW), lane departure warning, automatic emergency braking, and adaptive cruise control.
The return on investment (ROI) for safety technology is high. When you add up the direct, indirect, and intangible costs that are incurred when a collision avoidance safety system is not in place it’s easy to see why it costs less to invest in long term safety technology.
Significant ADAS Progress Has Been Made So Far
OEMs have been pushing hard in this area. Partnering with major transport/delivery companies, they have logged more than 1,000 miles of error-free platooning with connected Class 8 trucks. The platooned tractor trailers save time and fuel. The ADAS software that manages the platoon maintains a close, constant distance, and automatically maintains the gap between the vehicles by controlling the speed, acceleration, and braking.
Technology from Starsky Robotics enables companies to manage 10-30 highly autonomous trucks at once from the comfort of a command center. Last year Starsky-equipped trucks, that were piloted autonomously 85 percent of the trip time, carried 5,000 pounds of freight for 180 miles. The company is already earning revenue by operating vehicles in Florida, with plans to spread to Michigan and Nevada in the near future.
Many members of the general public imagine autonomous vehicles flying down the highway. But, among the first places this technology is adopted may be in trucks that travel on public roads and do so very slowly. Waste companies are showing interest in autonomous refuse trucks that allow an operator/helper to walk alongside the machine as it navigates itself along the collection route.
As the narrator in the video notes, autonomous vehicles “never get distracted by emotions or stress.” And thanks to an array of safety devices including radar-based object-detection sensors and cameras, they have a far better surrounding view than vehicles driven by humans.
The Industry Must Sell the Future to the Public
No matter how safe autonomous vehicles will probably be, the industry itself will have to work to convince the public that they should not be afraid of long platoons of driverless tractor trailers moving down the freeways frighteningly close to each other and moving at high speed. The move to vastly improved roadway safety will require sophisticated public relations work as much as it will sophisticated technology.
Creating vehicles capable of responding instantly and appropriately to the millions of variations that occur during even the simplest drive to the store is a challenge of gargantuan proportions. Industry experts these days find themselves having to constantly adjust their estimates of when the technology capable of doing this will be ready.
Deep Neural Networks to the Rescue?
As recently as 2010, autonomous-vehicle software developers at companies, including Google, Aurora, and Waymo, had to do all their coding line-by-line, a frustratingly slow process. The recent advent of a type of algorithm called a deep neural network has vastly speeded up the process. The algorithms themselves are able to learn how to complete tasks on their own, everything from detecting pedestrians to predicting what will happen on the road in front of the vehicle—and make plans for a way to keep moving forward without incident.
There may be some hiccups along the way, but there are companies that are actively working on a comprehensive safety solution to make the roadways safer. Take Bendix Commercial Vehicle Systems, LLC for example. In 2017 at the North American Commercial Vehicle Show they announced new system enhancements for their Wingman® Fusion™ driver assistance system. This new system will offer highway departure braking, multi-lane automatic emergency braking and strengthened collision mitigation and braking capabilities to more effectively reduce vehicle speeds in hazardous situations. In addition, Bendix™ announced the new Blindspotter 2.0 Side Object Detection System. The Blindspotter 2.0 will be integrated with the Wingman® Fusion™ system to help drivers further mitigate blind spot accidents.