Addressing the challenge of automotive OTA update -

Addressing the challenge of automotive OTA update


TOKYO — The automotive industry often describes a connected vehicle as a “smartphone on wheels.” Nothing, however, is further from the truth.

Senichi Yoshioka knows this as well as anyone.

Today, Yoshioka is senior vice president and chief technology officer of Renesas’ automotive business. But he’s a mobile business veteran. He used to work at Renesas Mobile, a smartphone apps processor company that existed several years ago.

During our interview at Renesas’ headquarters, Yoshika casually took his smartphone out of his pocket and cradled it on his palm. He said, “Architecturally speaking, this thing is so much simpler than a connected car.”

Basically, a smartphone has one big apps processor — all that’s needed to run apps. In contrast, vehicles come with a hundred or more ECUs. They are connected to a variety of vehicle communication networks. Each network uses different communication protocols to send data.

Despite this complexity, as they scramble to run a host of vehicle applications and services on their cars, automakers are under the gun to connect their vehicles to the cloud. The now obligatory applications include over-the-air (OTA) software updates, feature upgrades, vehicle maintenance/anomaly detection and protections from cyberattacks.

But here’s the rub.

Unlike Tesla, which designed its vehicle architecture from the ground up, traditional makers of gas-guzzling vehicles suffer from their legacy in automotive architecture. Vehicle models are based on disparate architectures that have proliferated over time, due to a patchwork of add-on features, ad-hoc improvements and proprietary extensions.

In short, OTA for the connected vehicle is a whole different animal compared to OTA for a smartphone, Yoshioka explained.

Vehicle computer
As a typical vehicle’s value shifts rapidly toward software, Yoshioka said, “We think it’s critical for us to a) reduce the complexity of in-vehicle network communication; and b) simplify in-vehicle API and encapsulate systems controlled by MCUs.”

As a leader of the automotive MCU, Yoshioka said, “It is incumbent on us to propose to the automotive industry a new ‘vehicle computer,’ which can simplify the vehicle networks, translate communication protocols and take care of security.”

A vehicle MCU installed in a hub should be smart enough so that it can compress data before handing it off from Ethernet to CAN bus for example, he added.

So, which MCU fits the bill for such a job in a hub of a vehicle?

Yoshioka said, “We are using R-CAR Gen 3 right now.” But as functions further integrate and vehicle domain centralization advances, Yoshioka said Renesas is plotting a road map to launch a newly dedicated vehicle computer in 2023, he explained.

Click here for larger imageVehicle computers will evolve as in-vehicle networks change. (Source: Renesas)

Vehicle computers will evolve as in-vehicle networks change. (Source: Renesas)

Vehicle data deeply buried inside a car
As Colin Bird-Martinez, a senior analyst with IHS Markit, once told us, “To pull off automotive OTA [from the cloud to ECUs], you need two things: deep knowledge in computing hardware and familiarities with a variety of in-vehicle network communication topologies.”

The same applies to getting vehicle data out of the car, sucking it up and uploading it to the cloud.

An oft-overlooked complication is that conventional vehicles on the move generate a lot of data. A car generates data that shows its wheel speed, wheel angles, wear and tear on components, anomalies in driving behavior and routing, and more. From a software developer’s point of view, this is a goldmine of worthy data that can turn into “information” that ADAS and autonomous vehicles can use.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.