Aeris Blog

A Practical Approach to Autonomous Vehicle Safety

October 24, 2018

In 2013, The National Highway Traffic Safety Administration defined five levels of vehicle automation, and manufactures today are beginning to shift from level one and two partial-automation features to level three and four driverless features in upcoming generations of self-driving cars. Safety is of the utmost importance as manufactures gradually employ more IoT-rich elements to phase out the role of human monitoring in these vehicles.

According to a survey from Morning Consult, fifty percent of respondents believed that self-driving cars are less trustworthy than human drivers. These driver perceptions reflect an understanding of how people’s lives are on the line as vehicle manufactures introduce higher levels of vehicle autonomy to fast-moving streets. Thus, there is much more at stake than the success or failure of a business opportunity when it comes to extending vehicle autonomy: both public safety and the trust of consumers are serious factors that make safety an imperative consideration throughout the development phase of fully-autonomous vehicles.

car 777

Passing the Safety Test

In order for autonomous vehicle safety tests to be effective, self-driving cars need to be confronted with real-life scenarios where the level of accuracy in representing, understanding, and responding to unpredictable roadway surroundings can be measured. Nonetheless, introducing autonomous vehicles to the highly spontaneous streets and highways is a big step that risks the success of the self-driving car due to the potential for catastrophic accidents and roadway deaths.

To address these safety concerns and limitations within the current technology, key developers of self-driving cars, like Waymo and Driver.ai, are employing focused, trial-based tactics in their new pilot programs that test the capabilities of level three and level four automation for these vehicles. All current and future developers of self-driving cars can set similar core restrictions and limitations to ensure the safe, meaningful, and productive implementation of their pilot programs and pave the way for a more actualized iteration of the autonomous vehicle in the future. Following are three core features of self-driving vehicle pilot programs

  1.  Keep Things Under Control

A major key to implementing a safe and successful self-driving car pilot program is to set up a controlled environment. For example, cars in the Driver.ai pilot program offer a shuttle service that travels strictly between a business park and a retail center. The stops are designated and the vehicle follows a routine course. Setting these controls help focus the self-driving car on the tasks of lane centering, braking, and signaling without having to juggle spontaneous route diversions.

Furthermore, pilot programs should strategize when and where they take place in order to create optimal zones free of road construction that could create overly complex obstacles, and that have some spontaneous pedestrian and vehicle activity, but not so much that all other safety precautions of these program become compromised. Strategizing to implement the program during lunch hour in a small business district, for example, rather than a busy route with many intersections can lower the risk of accident while still testing safety features in a real-life scenario.

car 444

  1. To See and Be Seen as a Self-Driving Car

Visibility is another important component in safe self-driving car testing. Other roadway users need to know from a distance that the vehicle is autonomous or “being tested”. Strong labeling, brightly colored markings, and display screens on the outside of the vehicle are great ways for the vehicle to self-identity and set clear expectations to surrounding road users. Furthermore, implementing pilot programs on clear, sunny days will ensure that no weather complications impact the functionality and processing capacities of the self-driving cars.

Self-driving cars also should be self-identifying to signal to other roadway users that a unique type of vehicle is nearby. Like busses, light rails trams, and those marked as “student driver”, self-driving cars have a specific mode of operating that is distinct from typical consumer vehicles. Introducing the self-driving car to the market in this distinctive way can help road users become accustomed to these vehicles as they progress towards a fully autonomous future of sharing the same road with all other vehicles.  

  1. Back-up Measures

When testing level three and four automation, vehicles interpret their surrounding and respond with the proper driving motions without the help of a human driver. Nonetheless, as the functionality and safety capacities of these vehicles still are being tested, having a backup human driver seated near the internal emergency driving functions is a final measure to ensure that any potential crisis that may arise due to an interpretive failure of the car can be diverted and rescued by the attuned senses of a human backup driver.  

Aeris Puts Safety First

When it comes putting higher levels of self-driving cars into operation, safety is the priority. It’s important for autonomous vehicle developers to set focused parameters for their pilot programs in order to practically test the progressing technology of their self-driving cars in a method that prioritizes the safety of all roadways users. By testing higher levels of automation in focused, small steps, these programs bring self-driving cars closer to trustworthy applications in the real world while prioritizing the importance of safety.  

Aeris understands that safety come first. Learn more about how Aeris connects with the self-driving car industry to implement practical, safe steps towards future of IoT’s integration with cars and roadways.

 

Press Inquiries

Kevin Petschow
Sr. Director of Global Public Relations
kevin.petschow@aeris.net
+1 (312) 985-9182