Today at TU-Automotive Detroit, Aurora Labs, the creators of
Self-Healing Software to future-proof automotive software, announced
Auto Validate, a Line-of-Code Maintenance™ machine learning product that
provides dynamic relationship analysis showing clear evidence of the
software running on the electronic control unit (ECU) and any changes
made during the lifetime of the system.
Auto Validate creates software functionality relationship maps to
compare software code from version-to-version. Software functionality
relationship maps will help automotive manufacturers streamline the type
approval process by making it easy to pin-point exact tests required,
provide evidence for regulatory authorities, save time and decrease
costs.
The type approval process consists of country-specific procedures and
regulations for vehicle approval. In most countries, including all of
Europe, China, India and Japan, automotive manufacturers must
demonstrate compliance to a national regulation authority. In North
America, automotive manufacturers self-regulate, however, evidence is
required for when questions arise asking for detail on software bug
fixes and software feature updates, proving sufficient compliance.
Auto Validate will help answer a growing requirement based on new type
approval compliance regulations. The United Nations Economic Commission
of Europe (UNECE), which includes members from Europe, North America and
Asia, is expected to ratify new regulations, detailed in the WP.29’s
report, by the end of 2019. According to WP.29, an amended type approval
for a new software update is needed unless the software update is only
fixing bugs or applying a security patch. WP.29 also states that if the
new software functionality only affects a limited amount of installed
vehicle software, the tests required to receive amended type approval
will also be limited.
Auto Validate assigns a unique digital thumbprint to changes in existing
functionality. This offers evidence that a specific bug-fix, or
cyber-fix did not add new functionality and new type approval
certification is not required. In addition, this process can be used to
document which specific systems have been affected by a software update
minimizing the amount of testing that is required for type approval
certification when new functionality has been added to the vehicle.
“It is an undisputed fact that the amount of vehicle software will be in
the hundreds of millions of lines of code,” said Zohar Fox, CEO, Aurora
Labs. “The type approval process is already lengthy and expensive. With
the increase in software feature updates, software-bug and
cyber-security fixes and new compliance regulations, the type approval
process is becoming more daunting. We look forward to partnering with
automotive manufacturers to help them adapt to the new dynamics of the
type approval process, liability and warranty,” he added.
About Aurora Labs
Aurora Labs is pioneering Self-Healing Software for connected cars to
enable automotive manufacturers to proactively respond to future vehicle
software architectures, processes, and services. Agile user-centric
software development processes create a plethora of continuous
opportunities and risks for the OEMs, even after the car has left the
production line. Aurora Labs’ Line-Of-Code Maintenance™ technology is
the foundation of its In-Vehicle Software Management solution. Using
machine learning algorithms to uniquely address all four stages –
detect, fix, update and validate — of a software health solution
providing a safety-net, Aurora Labs future-proofs the next generation of
software-driven automotive features. From detecting line-of-code faults
to predicting downtime events, fixing errors on-the-go, enabling
reliable and cost-effective rollouts of new automotive features to all
ECUs in the vehicle without any downtime for the user and validating
changes to the software, Aurora Labs is paving the way for the age of
the self-healing car.
Follow Aurora Labs on LinkedIn
View source version on businesswire.com: https://www.businesswire.com/news/home/20190603005052/en/