What autonomous driving can learn from trains and planes

What autonomous driving can learn from trains and planes

Technology News |
By Christoph Hammerschmidt

Although many adults still report experiencing a fear of flying, the majority of people implicitly trust that planes will get us from A to B in one piece. The same is true for trains and cars, which are both regarded as safe means of transportation. Indeed, few factors are as crucial in the transportation industry as trust. If people did not believe in the safety of a train, plane or car, no one would ever travel with these vehicles. However, the arrival of autonomous driving and increasingly connected cars casts doubts on the safety of the automobile. Is travelling by car still safe, even if the driver is replaced by software?

Interestingly, the introduction of the autopilot in commercial aviation did not seem to bother most passengers, and self-driving trains are standard in many European cities – people still trust in the safety of planes and trains. For the same to be true of autonomous cars, the system must earn the public’s trust.

Trust needs to be built over time

Public trust was not always a given in the transportation industries. It was built over time with the introduction of numerous safety measures. Confronted with deep suspicion following early fails in safety and reliability, rail operators, for instance, perfected new engines and rail systems that greatly reduced boiler explosions and derailments – earning the public faith in trains. In the automotive industry, as the number of cars on the roads grew from a handful to millions, car companies created safety mechanisms such as seat belts, crumple zones, and airbags. But most impressive is the airline industry, which has achieved unprecedented levels of safety. Indeed, 2017 was the safest year on record for flying.

Although none of the industries mentioned above have made the transition to full autonomy, they took important steps towards ensuring safety and allaying fears about their potential perils. Looking at these steps, the automotive industry can learn a lot for the acceptance of driverless cars. Admittedly, the autonomous car still has a long road ahead. The recent fatal crashes involving self-driving cars underscore the need to ensure their safety – to save lives, but also to win the hearts and minds of a skeptical public.

Until that happens, even though self-driving cars will likely deliver safer roads, lower costs, increased mobility, and reduced pollution, there is a risk that the wider public will simply not get into them. A look at the aviation industry shows how it succeeded in winning public trust by focusing on three key areas:

Frequent Maintenance and Checkups

Doing a check-up on your car once a year may seem inconvenient, but planes are monitored non-stop. Instead of waiting for the planes to land, onboard sensors transmit real-time data during the flight, in addition to the check-ups they get every time they touch down. Planes also undergo more in-depth inspections, such as monthly functionality checks and less frequent probes. Constant monitoring and frequent maintenance have been key to building confidence in airline safety.

Clearly, autonomous cars would require these kinds of stringent standards. To ensure safety and public buy-in, the industry must think of maintenance as a continuous, real-time process, just as the airlines do. Cars should have robust diagnostic tools running at all times with the ability to identify any potential malfunctions. This is particularly crucial as cars transform into “software on wheels”.

System Redundancies

If your car engine fails, you can pull over to the side of the road. In airplanes it is not so simple. That is why every airborne system – from hydraulics to flight management software – comes with a backup. Aerospace manufacturers spend billions installing these redundant systems to ensure that even if key systems fail, catastrophe is averted. Multiple engines and navigation systems ensure safe travel, even if one engine or navigation system fails.

What can the automotive industry learn from this? From a mechanical perspective, the handbrake can stop a car in case the primary brakes fail. But with the increasing amount of software code in connected cars, it is crucial that a car’s software systems also have robust software backup options in case a coding error leads to an operational malfunction.

Government Oversight

Aviation authorities perform rigorous inspections and certifications to promote airline safety and features. Furthermore, air traffic control systems monitor the movements of planes and respond to any incidents that may arise in-flight. As the mobility ecosystem becomes increasingly interconnected, traffic control for the roads may become an integral part of the solution for promoting autonomous transportation and traffic safety.

While regulators should be careful not to throttle innovation, regulation is an essential element for the autonomous future. If the developers of self-driving cars want the public to buy-in, they will need to advocate for a strong regulation that guarantees safety and reliability standards and monitors compliance.

Self-healing software for connected cars

To ensure that the “software on wheels” runs safely and error-free, problems must be eliminated as quickly as possible. Therefore, the ability to detect and repair software errors over-the-air (OTA) must become as universal as safety belts. These OTA updates are not a new concept. With the introduction of 3G networks and smartphones, they have become a widespread technology for remote updates. In the automotive industry, however, OTA updates have so far been limited to infotainment systems.

But before the OTA update become active, the problem must first be detected. Finding the software bugs that require an update in the vast amount of code in the complex car architectures is as difficult a task as fixing them. This requires a solution that scans the software for errors and detects them before they even become a problem. Software health pursues such an approach. The procedure monitors the behavior and resource requirements of the ECU software in order to proactively maintain the vehicle’s on-board systems and thus minimize downtimes and breakdowns.

Aurora Labs’ Software Health Solution uses machine learning algorithms that take over all three stages of troubleshooting to identify and repair software problems and seamlessly implement OTA updates. The solution does not require installation on the ECU or rebooting it. Using a backend solution, software errors and the risk of a possible ECU failure can be predicted. Then, the software heals itself by rolling back to the last safe version. Now, the OTA updates come into play: Efficient, clientless OTA updates ensure that all ECUs in the vehicle always remain up-to-date – without errors or downtime.

With this solution, car manufacturers and OEMs can guarantee the safety and quality that consumers expect. In addition, recalls and workshop visits are avoided as the software health solution detects problems in advance and repairs them even while the car is on the road. Enabling greater transparency, this will contribute to building public trust in autonomous driving. Widespread acceptance of self-driving cars will come when the industry adopts the robust backups, regulations, and oversight that are taken for granted in other transportation sectors.

About the author:

Zohar Fox is co-founder and CEO of Aurora Labs. He founded the company together with Ori Lederman in 2016 in Tel Aviv. Zohar Fox can look back at a 20-year leadership career in technology, product and sales of IoT, automotive, and banking systems.

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles