Better Software May Make the Difference in Real-World SDV Rollouts
One reason that Self-Driving Vehicles are appearing today is the convergence of several lines of technology. Among these are computer-controlled automotive components (there’s been a computer in your car since the 1980’s), Lidar-based sensors, V2V communications, and the availability of internet-based weather and satellite data.
All of this is necessary because information about other vehicles, pedestrians, roadside obstacles, weather, traffic congestion, maps and detours, all of this must be available, must be extremely up to date, and – here’s the rub – must be processed immediately.
Where will the software come from to handle this massive task?
It’s the Code, Stupid
According to a recent article by Edward Taylor, Reuter’s banking correspondent in Frankfurt, German car makers are already being affected by this limitation in their SDV development efforts.
Comparing the SDV development effort with the development of software for the military aircraft industry, Taylor points out that –
“Lockheed Martin’s $400 billion F-35 Joint Strike Fighter program, which began in 2001, is still being held up by software glitches, despite the company having more than 500 software engineers”.
Of course, the development of military software is shrouded in secrecy, while today’s open source revolution allows programmers to collaborate and improve code constantly. Lockheed’s 500 software engineers pale in significance to the millions of gifted programmers out there in the Internet. So can the car makers play nice and –
Share and Share Alike?
Can the software bottleneck be solved by information sharing? Just how open is the software development effort for SDVs?
One way of looking at this is to examine the plans of one of the current leaders in SDVs. According to Matthew Debord of Business Insider, quoting Tesla chief Elon Musk,
… Tesla thinks of itself as a Silicon Valley software company, as much as a car maker. He said that Tesla set out to build a very sophisticated computer on wheels.
“Most cars don’t improve over time,” he said. “But the Model S does get faster, smarter, and better as time passes. The car gets better as you sleep. When you wake up, it’s like driving a new car.”
So it would appear that the software embedded in the SDV is a key (and therefore lucrative) part of the package. Musk goes on to say that Tesla’s vehicles already have all of the components necessary to drive the car, but software (and data) updates, done automatically, will keep it on the road.
Keeping the Cards Close to the Chest
Bottom line – SDV software is a proprietary, valuable and closely-held asset. But this closed-source development model may slow down the progress of our SDV dream machines.