In a recent Automotive Megatrends webinar, Wind River and Altera examined the future of autonomous vehicles. Marques McCammon, Senior Director, Auto Solutions at Wind River, presented a list of factors that must be considered as the development of autonomous vehicle advances.
People and safety
“The way we approach the problem, the question, the challenge or however you choose to look at it, revolves around people, their personalities and skill sets,” he explained. “Whether that’s from the standpoint of technology development, or from the end use, we like to make sure we consider all aspects of the relationship with people.”
McCammon believes that autonomous driving could completely change the relationship between car and person. He said, “For the automotive industry, the autonomous vehicle is a convergence of the consumer electronics and computing industry with the conventional elements of the automotive industry. Factors like cyber security, high performance computing, consumer electronics – before, these were not really a factor for the automotive industry.”
In non-autonomous vehicles, the consumer is responsible for “the fundamental controls and safety.” The fact that control and safety is in the hand of computers poses significant challenges for the automotive industry – ones that McCammon thinks must be continually addressed as development progresses.
The process and challenges
The next factor that comes into play is “the process,” continued McCammon. Essentially, he described the “pathway from design all the way through to the implementation and use of the technology.” In the automotive industry, this course presents a “challenge that is considerably higher” than that of the aerospace industry. This is primarily due to the fact that there are “hundreds of engine control units (ECUs) per car,” which makes computing and coding extremely complex.
On top of that, “we want these vehicles to operate in an environment where they are surrounded by 200 million other vehicles that may or may not be controlled by individuals. It presents a significant challenge both from an engineering design standpoint, and from a processing a power standpoint,” McCammon noted.
Technology and cost
The third and final area of McCammon’s presentation revolved around the technology. As developments are made in the ‘process,’ technology evolves simultaneously.
“Software moves at a fast pace, and as new software is introduced its value increases very quickly. But, that value diminishes as new software takes its place,” he explained.
McCammon described a cyclical pattern that continues “on and on” as new technology is introduced. In terms of autonomous vehicles, the pattern is destructive: “By the time an OEM integrates a new technology and launches it, putting it in the hands of the consumer, it tends to be a lot less relevant. It’s no fault of the industry – they do a tremendous job. But what they are looking for is the ability to extend the value of the software and allow the tech to have a longer lifespan to support the use of the car.”
Essentially, McCammon says that the value of the software “must continue to grow,” instead of diminishing when it is replaced by updated versions.
Eyes on the prize
Referring to the three areas, McCammon thinks that each is as important as the next when considering autonomous driving. However, the latter tends to be the area of dominant focus.
“Quite frankly, when we talk about autonomy, in many cases the first two cases tend to get overlooked, and we spend a lot of time on technology. The truth is that the three things are symbiotic. They are closely tied to each other, and without the people and process, the technology has no relevance,” he remarked.
As for when the vision of autonomous driving will be realised, Clive Davies, Automotive Systems Architect, Altera, said, “If you look at what the OEMs and Tier 1s are thinking, generally there is a much more cautious approach considering the hype around autonomous vehicles. If you examine highway driving, we’re pretty much there already with automatic cruise control and lane departure – the key building blocks are already in position. I think we are going to see things like autonomous highway driving come in, and valet parking, and it will slowly build piece by piece. So there is no big bang date for autonomous vehicles, but I think it’s going to be a gradual evolution over time.”
Agreeing, Marques added, “I think we will see autonomy in vehicles by 2020. The OEMs are exploring limited autonomous applications like advanced driver assistance systems (ADAS), but quite frankly the semi-autonomous operation is just a pathway to autonomous. We still have a lot of issues from a regulatory standpoint and an interoperability controls standpoint that need to be sorted out that go beyond the car. The 2025 to 2030 timeframe is a more practical prediction for fully autonomous driving. But I think we will see some level of autonomy coming in very soon.”
Michael Nash
For information about this and other Automotive Megatrends webinars, follow this link
