Guest Post by Paul Kostek (first posted on CERM ® RISK INSIGHTS – reposted here with permission)
You’ve no doubt read about Google’s driverless car and the effort at the state level to gain permission for its use on public roads.
Driverless Car Risks
There are the obvious concerns and risks with operating a driverless car, though Google does have someone sitting behind the wheel, just in case. Of course, can a person just along for the ride respond fast enough to a problem? This becomes one of the risk areas needing to be addressed before full adoption of driverless vehicles. Integrated with Intelligent Transportation Systems (ITS) we’ll see vehicles and roads sharing data with each other to produce safer travel and relief of congestion.
We’re seeing more technology being transferred from the aerospace industry to cars, radar (for lane drift and braking), GPS, displays (increased data) and sonar for backing cars into parking spots.
How to Mitigate Risks
How will risk be addressed? If we follow the aerospace model, redundant systems, improved software and hardware design (application of DO-178C and DO-254) and communications will be help.
There’s an obvious cost factor that will come to play, how much redundancy will be required, what kind of warning will drivers be given. I’d expect agency’s at the state and federal level (NHTSA and NTSB) to look for improved technical performance and reliability, but what level of training will be required for operators, (will they still be drivers?), to properly respond to an emergency?
To answer these questions we’re not just identifying the risks, but also determining the appropriate response to failures, e.g. what happens if the steering computer fails? How will the vehicle respond and how will the operator?
Besides the technical issues with driverless vehicles. we’ll also need to address insurance and licensing issues. Are we prepared to send a car off with no licensed driver aboard. Who will be liable for an accident?
Too many questions and risk. Too few answers. What we do know is that interesting challenges will arise as we move to this new phase of transportation.
Bio:
Paul J. Kostek is a Principal of Air Direct Solutions, a systems engineering/project management consulting firm. He works with companies in defining system architecture, system requirements, interface definition, verification planning, risk management and software development standards. Paul received his BS from the University of Massachusetts, Dartmouth. Paul works in a range of industries including: aerospace, defense, medical device and e-commerce.
Paul is a long-time volunteer with several professional engineering societies including IEEE, AIAA, SAE, INCOSE and PMI. He also writes for the CERM Risk Insights emagazine.
Greg Hutchins says
Good day.
The challenges with autonomous vehicles now are many:
Bad weather.
Lidar accuracy.
Software object detection.
Public perceptions.
Suitable roadways
Data input
Algorithm accuracy
List goes on.
However the future still points to their use.
Paul Kostek says
Agree, we have moved from the general driverless cars are coming to identifying specifically what technology challenges we will need to address.