The Connected Car’s Virtual Chassis:

Tuesday, June 2, 2020

#Functional Safety    #Automotive SPICE    #Team Integration

The Software Battle Of The Future

Press releases, stockholder meetings, and industry reports talk about minimally three future, macro-trends for vehicles: connected, electric and autonomous. When considering what will allow any product to thrive in these categories, a major part of the success shall be the infrastructure behind it, above it or around it; a virtual chassis both internal and external to the physical vehicle. Bob Lutz, the former head of Product Design for General Motors famously stated, “Styling is the great differentiator”, but that may not be as true for the Connected Car of the near-future. In a recent survey by Brandwatch, 8000 respondents listed both “Innovative products or services” and “Personalization” as two of the top 8 wants in their next vehicle.

Part of these wants is driven by the past thirteen years of global learning about innovative, personalized, connected products via the Apple and Google revolutions, and how cloud-based enablers play a significant role. Apple previously had handheld device failures like the Newton, but it wasn’t until the iTunes Store supported the original iPhone in 2007 that a market shift happened. Personalization was all about what could be added; not the hardware. New features could be downloaded, failed features could be updated (or deleted), customization within a standard platform was made easy, and complex interactions could progress from archaic voice menus to natural language. 

So what effort does that great, sustaining connected experience require? Google’s former Engineering Manager estimated in 2015 their services required two billion lines of code, and that has assuredly grown since then. To put that in perspective, the Boeing 787 Dreamliner had approximately one percent of that (14M lines of code) for both on-plane and off-plane software combined, an F-35 Fighter Jet nearly doubled that (23M lines of code), and even a complex autonomous vehicle is theorized to have 500 million lines of code embedded within its steel walls. “The amount of code on the vehicle will certainly increase,” says Roger Berg, Vice President of Research and Development at DENSO, “but putting a number on how much code will be required outside the vehicle is unfathomable.” Intel projects upwards of 4 terabytes of data being generated by every car every day, which would require systems to store, analyze, adjust, display, and react to the data as desired.

To truly understand the enormity of the software battle ahead, there are at least three User Experience touchpoints to consider: the Use, the Maintain, and the Buy touchpoints.

The “Use” Touchpoint

Traditionally, automotive manufacturers have planned mostly for this one touchpoint: “Use”. How will the customer use the gear shifter? How will she use the radio, parking brake or turn signals? The product in full used to be the bumper-to-bumper automobile, and in some regards this will still be true. “A supplier’s traditional business won’t go away,” says Berg. “There will always be the need to provide the major systems we’ve always provided.” What will change, though, is the volume of off-vehicle code that will be developed to enhance the driving experience.

Examples of this are already in production and many are invisible to the average customer. For instance, drivetrains and chassis have started to send information to the cloud for analysis, long-term recalibration and engineering updates to improve the experience. These electronic drivetrains monitor, collect and send such data that can be used to make adjustments towards smoother, quieter, and more pleasurable rides based upon a feedback loop. In fact, Tesla recently released an over-the-air update that improved performance and range by 10%. A more visible example is dynamic navigation systems which parse probe information, road closures and real-time data to update directions based upon revised conditions. 

The future, though, will have even more extreme cases of how the off-vehicle systems will support the daily usage of the vehicle. Multiple manufacturers have realized that autonomous systems will likely need real-time operations to support edge cases. Nissan foretold of this at The Consumer Electronics Show (CES) in 2017, and described off-vehicle operators creating real-time obstacle-avoidance scenarios based upon learnings from Mars explorations by the National Aeronautics and Space Administration (NASA). “Part of the system will always be edge computing, but autonomous vehicles cannot decide everything internally,” states Guillaume Devauchelle, Vice President of Research and Development at Valeo. “There are so many occasional decisions that must come from outside the vehicle. A live, remote controller located in a remote country might have to assist 2-3 times per minute when the vehicle recognizes obstacles or difficult situations like a gesturing policeman. This must be done for now to support the Safety Goal.”

The “Maintain” Touchpoint

Vehicle maintenance seems to be the most obvious battleground for the Connected Vehicle in the next 3-4 years since the average household spends 1.5% of its annual income on vehicle maintenance and repairs. And in that same Brandwatch study, “Quality” was the most desirable feature by far with 30% of the respondents stating its importance.

And not shockingly, the Connected Car can absolutely assist. “Instead of looking at starter motors after they have been replaced,” says Berg, “we can use connectivity to monitor the behavior of systems in operational mode. When we have enough data and the associated Artificial Intelligence algorithms, we can model, detect and predict what needs maintenance or replacement before catastrophic failure within a certain amount of accuracy. That provides a new level of value above being a traditional parts supplier.” That means easier maintenance for the end customer, reduced warranty for the manufacturers – to the tune of millions of dollars – and better quality on the next vehicle.

All of this requires remote tools, dashboards, Artificial Intelligence algorithms, interfaces for communication, and systems which can compare the build configuration of a given vehicle to symptoms already encountered. Lots of information. Some of the effort can be streamlined with existing tools (e.g. Anark for information management, collaboration and coordination; numerous Artificial Intelligence engines for analyzing data), but handling all of the permutations of vehicle designs and customers’ use cases requires first-and-foremost a new level of User Experience strategy along with a long-term vision shared among Marketing, I.T. and Engineering.

The “Buy” Touchpoint

The current pandemic has made the Buy touchpoint brutally relevant as analysts predict a greater number of customers will purchase vehicles online with a vastly different user experience. This has already started to include “virtual walk-arounds”, online-scheduled test drives at home (e.g. car temporarily provided to the prospective buyer’s home), and home-delivery by a masked [wo]man that helps the buyer avoid an anxious dealership visit. It is not hard to imagine, though, connected vehicles will soon navigate themselves autonomously overnight to the garaged address of the buyer, thereby decreasing the pain of such a forced reality. And once the customer has taken possession of the vehicle, connected features might be bought from afar ranging from Dodge/Uconnect Market’s Fuel Mobile Pay to Tesla’s “Cheetah Stance” (i.e. lowering the front axle). 

Additionally, the connected car’s enablers have now opened novel Buy touchpoints. “New, non-traditional customers have initiated discussions of long-time services where uptime for their fleet equates to utilization and revenue,” says Berg. “Those customers will need to know when every part typically trends towards failure and when to schedule repairs around that reliability. Such offerings need to have complex algorithms, easy interfaces and a clear value propositions, which is an exponential add on top of the design and coding that already occurs.” 

“Managing offerings will change from country to country,” asserts Devauchelle. “For instance, the EDeliver4U autonomous robots that deliver meals in congested urban centers could include facial recognition for customers including user feedback based upon expressions or gestures. Something like that might be desirable in markets like China, but would not be offered or permitted in all regions. Privacy restrictions or customer acceptance must be considered.” Therein, not only must the base scale of the project increase significantly for such offerings, but the variance for purchase decisions within a region must have flexible levers based upon regulations or customer experience feedback. “On both the hardware and software side there’s a significant investment, so any ‘platform’ – both on and off the vehicle — is built on stable, long-term bricks. The business can swap out those building blocks depending upon what that vehicle or business needs.”

In the end, to provide an iPhone-type offering rather than a Newton, all of the touchpoints must be managed with a cohesive strategy, a long-term vision and a virtual chassis.


Do you need to improve your automotive product development, to increase efficiency, or to comply with ASPICE and Functional Safety? You are at the right place.