What Is a Digital Feedback Loop in Automotive?

A digital feedback loop is a process of data collection and analysis used in the automotive industry to inform continuous improvements to vehicles via over-the-air software updates.

For most of the automotive industry’s history, vehicle features typically have not evolved beyond the point of sale. However, with technology evolving so rapidly in other areas of their lives, consumers now expect OEMs to incorporate new features continuously.

Two-way communication between the vehicle itself and software developers is essential for such ongoing innovation. That is where the digital feedback loop comes in. Feedback is collected from vehicles, analyzed, and used by developers to create software updates. The updates, in turn, are downloaded to vehicles to complete the loop, and the cycle begins again. 

Understanding vehicle behavior

To accelerate innovation, developers need to know how well their features are meeting consumer expectations — how the features are performing, how frequently they are being used and if anything unexpected is occurring. Many customer-facing businesses solicit feedback on their products and services through social media platforms, emails, surveys, online reviews and customer-support interactions. Of course, user-generated feedback is subjective, and it is not always easy to extract a clear or consistent message from it — especially when it concerns something as complex as a vehicle.

As vehicles increasingly become more software-defined and connected to the cloud, OEMs have an opportunity to collect objective data on how a vehicle’s software and hardware systems are functioning in the field, through quantifiable measures like CPU performance, memory performance, event-driven data, and diagnostic and usability metrics.

Collecting that data and transmitting it back to OEMs is no easy feat. The amount of data produced when using a smartphone is a fraction of the data produced by a vehicle. Fortunately, 5G connectivity is making it easier than ever for a vehicle to connect with the cloud. And intelligent preprocessing on the vehicle can ensure that only essential data is transmitted, conserving bandwidth usage.

Once data has been collected, developers can use analytics tools to identify any problems that need fixing, as well as opportunities for enhancement or even entirely new features. Eventually, AI and machine learning could be used to analyze patterns in the data to generate insights that are harder for developers to derive.

In the future, cabin monitoring systems could also provide OEMs with near-real-time feedback on driver satisfaction by monitoring reactions and emotions — such as facial expressions that convey joy or frustration during a certain use case.

Instead of representing isolated events, data will be anonymized and aggregated at the fleet level to reveal meaningful trends. So instead of seeing just one driver’s experience with a certain feature, developers could observe tens of thousands of drivers consistently experiencing the same feature, prompting them to make improvements.

Leveraging cloud connectivity for processing power

The processing of large amounts of vehicle-generated data occurs in the cloud, where algorithms analyze data on a specific vehicle over time and aggregated data on fleets of vehicles.

For example, as battery management software evolves, a data-driven DevOps toolchain enables the creation of a cloud-based digital twin of the vehicle, using data collected from the digital feedback loop. Algorithms leverage the experiences of other vehicles all over the world in a variety of driving scenarios to learn about and help extend the life of EV batteries and the vehicles they power.

Cloud connectivity enables vehicles to utilize offboard processing power to run powerful algorithms, including those training AI and machine learning modules. Developers can see the interaction between different sensors and datasets so that potential issues, improvements and innovations become apparent.

Digital feedback loops help OEMs shorten development cycles by providing developers with immediate feedback from the field. This enables updates to be delivered in a timely and ongoing manner that aligns with the CI/CD methodology.

Wind River and Aptiv

Aptiv’s expertise in high-performance sensing and compute, application software development, edge analytics and vehicle architecture, combined with Wind River’s edge-to-cloud software portfolio, helps OEMs reduce costs and complexity, increase flexibility and unlock new business models through the digital feedback loop.