Tech trend disengagements in autonomous vehicles
As a quality professional, you understand the process of perfecting a product or process. For many years, the automotive industry has been working to create a safe and reliable autonomous vehicle. Although many companies are making great strides, it remains to be seen which one, if any, will be the first to come up with a commercially viable product.
A long road ahead for autonomous vehicles
Autonomous vehicles still have a ways to go before they reach consumers as safety remains an urgent priority. To operate self-driving vehicles, fundamental technologies are available, but integration with automotive control systems is still a critical issue.
For instance, each developer is trying to deploy its own operating system to integrate with industry-standard automotive software. Think of transmission timing software as a good example. It can be deadly if a transmission slips when a driver disengages auto-pilot and takes control or vice versa.
If an autonomous vehicle were to lurch forward or suddenly decelerate, it can easily cause a driver to lose control. Any driver, no matter how experienced, occasionally over-corrects steering, which can be treacherous in hazardous driving conditions. Imagine driving around a curve during a thunderstorm. You can barely see your lane, and your steering controls may already be compensating for lost traction. Jumping between gears can easily cause an accidental skid.
Why driver disengagements matter
Along those lines, analyzing driver disengagement statistics is one way to gauge progress. In 2016, there were 2,578 driver disengagements reported to California’s Department of Motor Vehicles.
Some developers, like Waymo, a Google offshoot, highlighted fewer driver disengagements over last year. Data show that Waymo autonomous vehicles had 75 percent fewer disengagements in 2016, but this statistic does not tell the whole story.
You know how important it is to place any data in context. The question is: How do you define a disengagement? With good intentions, California is attempting to regulate autonomous vehicle safety, but the state does not provide guidelines on what data developers must divulge.
For instance, Tesla Motors’ report showed no specific number of driver disengagements because testing either took place out of state or on a private course. Instead, Tesla’s publicly disclosed data show “takeover events” without placing any incident in clear context. Their report vaguely classifies causes for disengagements as “invalid outputs,” and some data points do not even specify a road class.
The dichotomy between Waymo and Tesla, according to their own public filings, suggests that autonomous vehicle developers want to divulge as little data as possible to meet state regulations, likely because they want to avoid providing valuable information to competitors. The first developer to break through the driver disengagement issue will prove the value of its quality management processes.
How EQMS can play a role
Enterprise quality management software can give you a single source of truth. Developing autonomous vehicles shows why it is so important to eliminate data silos when bringing new products to market. Without context and valid correlations among data, you will not be able to form meaningful conclusions.
Autonomous vehicles already create an incredible amount of real-time data. As technology evolves to incorporate artificial intelligence, autonomous vehicles will generate even more data. If you want to be able to draw any conclusions from massive amounts of data, you need a platform like an EQMS to pinpoint opportunities to raise quality.
As the market for autonomous vehicles matures, more quality issues are sure to surface. Subscribe to this blog for updates on quality issues facing autonomous vehicle manufacturers.