Intempora releases software toolchain for validation of autonomous driving algorithms with Microsoft Azure


Intempora S.A, a software editor specialized in real-time multi-sensor software solutions, has unveiled IVS – Intempora Validation Suite - a new set of software tools for cloud-based testing and validation of perception and deep learning algorithms and relies on big-data and high-performing architecture provided by Microsoft Azure for managing high-volumes of recorded automotive sensors data.

Today, test systems and prototype vehicles are up and running around the world in order to record, simulate and process billions of kilometers. The road to fully autonomous future is tough and fraught with challenges.

A new standard called SOTIF (Safety of the Intended Functionality - ISO/PAS 21448) provides guidelines to address the way to guarantee the operational safety of functions (such as vision-based perception systems, neural network algorithms…) in the absence of a fault, hence complementing the functional safety standards such as ISO 26262.

Due to the high volume of sensor data generated by autonomous vehicles (up to tens of Petabytes), the diversity of scenarios and the unpredictability of driving situations, perception algorithms are hard to certify and automotive engineers need to find new methods and solutions to perform statistical validation since formal proof becomes too complex and unaffordable.

Among these challenges, focusing tests on relevant scenarios and corner cases considering a given function (e.g. obstacle detection, accurate positioning, etc.) is key in order to reduce the amount of kilometers to test while ensuring good situations coverage and hence optimizing the IT infrastructure resources.

Intempora unveils IVS - Intempora Validation Suite, a set of new software tools for the benchmarking, testing and validation of Advanced driver assistance systems (ADAS) and Highly Automated Driving (HAD) functions including deep learning and perception algorithms.

IVS enables automotive and IT engineers to centralize, store and share their recordings of driving datasets, annotate and index them with tags (obstacles, weather and traffic conditions, road network information, vehicle data, etc.), search for particular situations, previews results, then post-process these huge amounts of recorded driving data on parallel computing clusters in order to validate their perception algorithms with specific scenarios.

“Data management and test automation for ADAS/HAD functions validation have been a longtime request by our users all over the world, especially with their huge amount of recorded autonomous driving sensors datasets. Today we are proud to release the Intempora Validation suite (IVS) powered by Microsoft Azure to answer the needs of the automotive industry,” said Nicolas du Lac, CEO Intempora.

Sanjay Ravi, General Manager Automotive Industry, Microsoft Corp. said, “Intelligent data analytics and machine learning algorithms are key to the implementation of autonomous driving scenarios. By integrating IVS with Microsoft Azure‘s cloud and AI tools, Intempora helps the automotive industry tackle autonomous driving challenges that haven’t been solved yet.”

Most Popular

Digital Editions

PMV Middle East - May 2020

Subscribe Now