Fixing software errors before they happen

Fixing software errors before they happen

Technology News |
By Christoph Hammerschmidt

With evermore autonomous vehicle systems, detecting and fixing software errors proactively is increasingly important for the automotive industry. To be future-proof, the industry needs solutions that create trust throughout the ecosystem by analyzing the software dependencies and mapping the relationships between functionalities to ensure that vehicles are safe and secure throughout their entire lifecycle.

Like any other industry, the automotive industry was completely under the influence of the Corona pandemic this year. But how does the industry look forward to the coming years? What developments do stakeholders and experts from the automotive sector expect in autonomous driving? What challenges do they see as the most urgent in view of the increasing number of software-based vehicle systems? In order to find answers to these and many other questions, we joined forces with leading market research and consulting firm Strategy Analytics to survey 200 automotive industry experts, analysts and journalists worldwide.


(Fig 1)

In view of the ever-increasing number of software-based vehicle systems, it is not surprising that the study participants consider the proactive detection of software errors to be of great urgency. 88 percent of those surveyed believe that it is important or very important not only to react to anomalies in software behavior, but to proactively predict them (Figure 1). Two-thirds of the participants in the study also assume that it will be more difficult or much more difficult in future to correct software errors after the vehicle has been delivered (Figure 2).


(Fig 2)

Identify software dependencies and proactively prevent errors

In modern vehicles, there are a wide variety of software systems from different manufacturers, some of which are interdependent. These dependencies between different control units are extremely complex. How can car manufacturers proactively predict whether changes to the software of one control unit will affect the behavior of other control units?

There are solutions that enable car manufacturers to analyze these dependencies between different software systems and thus predict possible sources of error. This makes it possible to proactively prevent potentially dangerous malfunctions. Our Line-Of-Code BehaviorTM technology, for example, relies on Artificial Intelligence and Machine Learning to manage the entire vehicle. This technology works in four steps:

  1. Validate: The structure, relationships, and dependencies of the software are checked in real-time using a dynamic algorithm that identifies and maps on a line of code execution level, what is really running on the ECU. The algorithms validate which triggers, signals, and code lines are affected by changes to the software due to an update, and therefore which functions require an amended type approval. This process provides the evidence required to greatly simplify the regulatory approval process.
  2. Detect: Our proprietary learning algorithms analyze the behavior and relationships between the millions of lines of software code in the vehicle. This allows car manufacturers to detect when drifting and deviation in the software behavior occurs, which could indicate a change in software configuration, a software error, or a hack. Line-Of-Code Behavior technology can predict which new trend in the software behavior can potentially become a problem that will lead to a system failure.
  3. Fix: When a problem is identified, the solution seamlessly isolates the affected software and maintaining uptime of the system by rolling back to the previous secure version without requiring dual banks of flash memory and all with zero downtime to the system. Basically, the car heals itself, preventing downtime and malfunctions.
  4. Update: The final step is the Over-the-Air (OTA) Update. Unlike current solutions that compare software binary files for changes, line-of-code updates create much smaller additive update files by identifying changes in the lines-of-code, thereby saving bandwidth transmission costs. Furthermore, using line-of-code updates enables the ECU to be updated without reprogramming the entire memory or taking it offline – the daily use of the vehicle is not affected.

Fixing errors via OTA update with zero downtime

A modern car has more than 100 separate engine control units that use different chipsets with different memory sizes, clock rates, and operating systems. These are connected via several in-vehicle networks with different protocols such as CAN, FlexRay, MOST, LIN and even Ethernet and are made available to OEMs by various vendors, integrators and suppliers.

OTA updates play an important role in this process. They make it possible to install the “repaired” software version on the vehicle and keeping the security patches up to date. The software of modern cars is already frequently managed remotely and updated via OTA updates. But the OTA update solutions for automotive software used today were originally designed for binary-image differential algorithms (BSDiff), which pushed the industry to costly solutions requiring redundant storage resources.

Aurora Labs’ line-of-code differential solution fixes the approach and allows even the smallest ECUs stay up-to-date without the additional cost of dual memory and for the larger ECUs that already have dual memory configured in, our approach updates them without reprogramming the memory leading to a faster update and guaranteed reduced system friction. Additionally, it will enable to swap between multiple versions and reduce the overall data sent over the air.

Validation of updates accelerates type approval and reduces costs

The United Nations Economic Commission for Europe’s (UNECE) World Forum for Harmonization of Vehicle Regulations has adopted new regulations to manage cyber risks and provide safe and secure software updates. The adopted regulation (WP.29) outlines the requirements and processes needed to assure that the software update management system (SUMS) guarantees safe, secure and reliable OTA updates. In Germany, adoption of these regulations is expected for all new vehicle introductions starting in 2022.

Line-Of-Code BehaviorTM technology enables the validation of the newly updated software, enabling the vehicle manufacturers to save costs when type-approving the car in adherence to WP.29. By analyzing the relationships between the software functions in the individual control units of a vehicle, it is possible to determine precisely whether an update of one software component will affect other components. In this way, vehicle manufacturers can easily and quickly provide all relevant evidence to the authorities during type approval.

This accelerates the type approval process and thus reduces costs – which is very much to the benefit of the car manufacturers. Our study showed that 67 percent of those surveyed would like a software solution that ensures vehicle safety without incurring additional costs. Only a third of the study participants stated that safety automatically entails higher costs for them (Figure 3).


(Fig 3)

 Line-of-code BehaviorTM technology, therefore, helps car manufacturers ensure the safety of connected and autonomous vehicle systems while reducing costs. At the same time, the Line-of-code-BehaviorTM technology offers drivers maximum convenience, as OTA updates provide the vehicle with innovative technologies on a permanent basis – with great cost savings for the vehicle manufacturers.

About the author:

Zohar Fox is Co-founder and CEO of Aurora Labs

Zohar Fox has 20 years of leadership experience in the IoT, automotive, and financial industries. As co-founder and CEO of Aurora Labs, based in Tel Aviv, Zohar Fox is pioneering self-healing software. This is a vehicle software management solution that uses machine learning and artificial intelligence to make software behavior more predictable.




If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles