The software startup is leveraging a robotic operating system to advance AV capabilities.
Multiple segments of the automotive industry have been embracing open source software, especially Linux-based code over the last decade.
Here we focus on an open framework called robotic operating system (ROS). Apex.AI is leveraging ROS, adding new automotive capabilities discussed below.
ROS has been around for over decade and increasingly used by autonomous vehicle (AV) developers. Apex.AI has greatly improved ROS for automotive applications.
I summarize the importance of ROS and then profile Apex.AI. With its focus on the automotive industry and software-defined vehicles, the company is gaining support, making it a future player in AV development and a likely acquisition target.
ROS is a software framework focused on developing systems and applications for robotics products. Robotics includes most types of autonomous systems—terrestrial, airborne and seaborne. The table below provides an overview of ROS. Sources include the Open Robotics website.
ROS is not an operating system as it does not control and manage hardware resources—the key function of an OS. Instead, ROS works with an OS, preferably a real-time operating system.
ROS encompasses many functions that are included in middleware. ROS middleware is focused on building a software ecosystem for autonomous and robotics devices. This development-centric approach leverages the traditional strengths of open source development.
Software in the ROS ecosystem fall into three categories:
Both language-independent tools and the main client libraries (C++, Python, Lisp) are released under license, and are free for both commercial use and research.
The final version of ROS 1 was released in May 2020. ROS 2 is greatly improved, leveraging a large ecosystem of software libraries and development tools to build robotics applications.
The second-generation version incorporates a new API that supports real-time programming and a wider variety of computing environments.
ROS Ignition includes robotics tools such as simulation libraries and cloud services. Five Ignition versions have been released. Another simulation development tool, Gazebo, is focused on 3D simulations with robotics devices, obstacles and other scenarios. Gazebo is on its eleventh version.
ROS also provides services such as low-level device control, hardware abstraction, message passing between processes, package management and others. Software development processes also can be represented as a graph architecture.
The Open Source Robotics Foundation (OSRF) was established in 2012 as a California corporation. OSRF is a nonprofit, public benefit entity and tax-exempt charity under IRS section 501c. A subsidiary in Singapore focuses on the regional development community. In 2017, the foundation changed its name to Open Robotics. Since then, it has grown to more than 50 employees.
Open Robotics is guided by a 19-member steering committee that includes representatives from Amazon, Bosch, Intel, LG Electronics, Microsoft, Samsung, Sony, Toyota Research Institute, Ubuntu and Wind River. Apex.AI is also a steering committee member.
An early version of ROS was developed by students at Stanford University, some of whom later launched Willow Garage to continue ROS development.
Willow Garage released its first ROS version in March 2010, attracting a large number of university developers in the U.S. and elsewhere.
In February 2013, OSRF took over ROS development. Since then, Open Robotics has expanded the ROS ecosystem and updated the software technology.
Open Robotics is currently focused on five industry sectors: aerospace, automotive, healthcare, logistics and maritime.
Apex.AI was founded in 2017 in Palo Alto, with offices in Berlin, Munich and Stuttgart, Germany. It has received more than $70 million in funding, including investments from Airbus, Continental, Hella, Jaguar Land Rover, Toyota, Volvo Group and ZF, according to the company’s website.
CEO Jan Becker has more than 20 years of AV experience. Becker was a member of the Stanford team competing in the 2007 DARPA Urban Challenge. Later, he headed autonomous driving development at Bosch North America. He was also a co-author of the SAE autonomy level definitions.
Becker has lectured at Stanford University for more than 15 years, focusing on AVs and ADAS technologies.
CTO Dejan Pangercic, who coordinates development activities at Apex.AI, has more than a decade of experience in robotics and software frameworks, including ROS-based software development.
The company’s strategy focuses on advancing the ROS 2 framework to the next level by rewriting the ROS 2 code with an emphasis on automotive features. The upgraded version called Apex.OS is a real-time automotive development kit and platform. Apex.AI has added hardware abstraction, meaning Apex.OS code can run on multiple microprocessor architectures. The strategy is also focused on software-defined vehicles and especially safety-critical systems.
Apex.AI has received ISO 26262 ASIL D certification from leading German testing company TÜV Nord. The product, Apex.OS Cert., is positioned as an OS framework and development kit for AVs and software-defined vehicles.
Apex.AI uses C++ to develop its software, offering middleware libraries in C, C++, Python and Rust languages. That strategy allows Apex.AI to ready its technology for production ADAS and AVs.
Apex.OS was introduced as a software platform that can be integrated across all in-vehicle domains while extending into cloud functionality. As a fork of ROS 2, Apex.OS is billed as real-time, reliable and deterministic for use in safety applications.
The robotic OS is aimed at two main customers: OEM developers implementing complex AI software integrated with multiple ECUs; and AV developers implementing safety-critical applications. Apex.OS also can be used with other development frameworks, including AUTOSAR Adaptive.
Apex.OS 1.3 and Apex middleware can be used together. The latter is designed as an integration of Eclipse Cyclone DDS and Eclipse iceoryx, enabling real-time data transfer and so-called zero copy communication. Zero copy is a crucial feature for developing software with strict real-time requirements.
The figure below shows how Apex.OS integrates with Apex.Middleware and other products. Note that a real-time OS is needed on the top of ECU hardware.
Apex.Autonomy serves as a set of certified building blocks for developing autonomy functions. It includes a lidar-based L4 autonomous driving stack for building safety-critical applications.
Apex.AI has disclosed some information on its partners and customers, with other relationships likely forged but unannounced.
Continental and Apex.AI have been collaborating since 2020, using Apex.OS in vehicle electronic systems. As a Tier-1 supplier, Continental needs better software capabilities, and Apex.AI provides safety-critical software tools.
The Green Hills Integrity RTOS kernel has achieved ISO 26262 ASIL D certification, and has adopted the ISO 21434 automotive cybersecurity standard. Integrity is a good companion for many Apex.AI customers.
Apex.AI has also partnered with Tier IV, a Japanese AV software startup. Tier IV developed an open source autonomous software platform called Autoware. The partners will combine their software and offer it to their customers.
Toyota’s Woven Planet is also using Apex.OS for safety applications such as AV software. Woven Planet is implementing Apex.OS in its Arene platform. Toyota’s software development platform includes advanced tools, APIs and safety building blocks to shorten software development cycles for new mobility and automation applications.
Another customer, Voyage, is using Apex.OS to develop its fixed-route vans. Voyage was recently acquired by Cruise.
ZF is both an investor, with a 5-percent ownership stake, and Apex.AI customer. It is developing safety-critical software based on the Apex.OS.
ROS is an established, popular open source software development framework and platform for creating robotics systems and applications. It has attracted strong technology industry participation under the stewardship of Open Robotics.
Apex.AI has leveraged the ROS ecosystem to develop products that meet auto industry needs. Apex.OS has received the highest rating for safety-critical software: ISO 26262 ASIL D certification. Its companion products, Apex.Middleware and Apex.Autonomy are likely to gain similar safety ratings.
This is an excellent strategy. Apex.AI’s technology should prove useful for auto OEMs, Tier-1 suppliers, AV software developers and those designing safety-critical systems.
It also likely Apex.AI will become an acquisition candidate.
This article was originally published on EE Times.
Egil Juliussen has over 35 years’ experience in the high-tech and automotive industries. Most recently he was director of research at the automotive technology group of IHS Markit. His latest research was focused on autonomous vehicles and mobility-as-a-service. He was co-founder of Telematics Research Group, which was acquired by iSuppli (IHS acquired iSuppli in 2010); before that he co-founded Future Computing and Computer Industry Almanac. Previously, Dr. Juliussen was with Texas Instruments where he was a strategic and product planner for microprocessors and PCs. He is the author of over 700 papers, reports and conference presentations. He received B.S., M.S., and Ph.D. degrees in electrical engineering from Purdue University, and is a member of SAE and IEEE.