We are building an operating system for autonomous vehicles that is designed to never fail
Experts predict that by 2030 at least 15 percent of cars will be fully autonomous. There is no shortage of opinions and hot takes on the latest sensors, use cases, and business models. One topic will be critical for a successful introduction of self-driving cars into market, but it rarely gets as much attention: Reliability, robustness, safety and security of the software that drives the vehicle in lieu of a human driver. Self-driving car software could stall just like laptops stall. A stalled laptop is frustrating but crashed software on a car may have more severe consequences.
Today I am proud to announce Apex.AI. We are focused on creating robust, reliable, safe, and secure software for autonomous mobility systems. Apex.AI is transitioning autonomous vehicle software from R&D to commercialization so that it meets highest safety and reliability standards, a need for both automakers and consumers. We recently closed our $15.5MM Series A funding, led by Canaan with participation from our seed investor Lightspeed.
We believe in a modular and diverse approach to software. Hence, we have designed a modular software stack, that allows for easy integration into existing systems as well as easy integration of 3rd party software. We think highly focused specialists can develop a better overall system than one company developing a full-stack — which will likely not fit all applications. Our software is not designed for peak performance — it’s designed to never fail. We’ve built redundancies into the system design to ensures that single failures don’t lead to system-wide failures.
To understand the problem we’re solving, a brief overview of AV software is helpful.
Background on the current systems
Autonomous vehicles use sensors to perceive objects and other traffic participants. The sensor data feeds into a computer and the computer runs software to understand the environment and to make driving decisions based on this understanding. This software is very complex and contains computer vision, 3D perception, machine learning, localization, decision making, trajectory planning, and control algorithms. These algorithms need a software framework to run in.
ROS, the Robot Operating System developed since 2008, is used by the majority of companies developing applications for autonomous vehicles and by almost all academic labs. ROS is analogous to iOS’s or Android’s SDK in the mobile stack, providing a software framework for developers to build applications on top. ROS has found widespread adoption because it is open source, has a strong community behind it and has seen a decade of application development, and provides a rich ecosystem in addition to the software framework, such as sensor drivers, algorithms, visualization, simulation, build tools and much more. ROS is excellent for rapid prototyping and development but is not designed to run in safety- and security-critical applications. As the industry shifts from development to productization and from testing to real-world applications, the highest level of robustness and reliability is needed.
Apex.AI’s approach to a modular, safe and secure system design
Our first two products are Apex.OS and Apex.Autonomy.
Apex.OS is a robust and reliable software framework or meta-operating system for autonomous mobility. It is an automotive version of ROS 2 and abstracts hardware-specific, real-time, and embedded software development so developers can focus on building production-grade autonomous driving applications on top of Apex.OS. We modify the popular and widely used ROS software framework to ensure compatibility to the development standard and have added mechanisms for real-time, functional safety, and security. Apex.OS will be certified to the highest automotive functional safety level (ASIL D) according to the automotive safety standard ISO 26262.
Apex.Autonomy provides developers with building blocks for autonomy. We believe that our customers should be able to pick and choose software building blocks from an ecosystem to compose a software stack that is right for their application — and not take a full-stack that has been developed by a single 3rd party. Instead we have started to build those blocks, e.g. automotive grade LiDAR data processing, and are also partnering to expand this into an open ecosystem.
There are different approaches to building software for autonomous systems and some have chosen to create an entirely new system from scratch. We believe ROS provides significant benefit — including widespread adoption, a huge developer community, and a decade of application development. Apex.OS is meant to transition and bolster ROS, not replace it.
Ultimately, Apex.OS is application agnostic, so any autonomous mobility system — cars, robots, drones, airplanes — can use it.
Dejan Pangercic, my cofounder and the CTO of Apex.AI, was one of the early contributors to ROS and has worked on software frameworks for over 10 years. I have worked on autonomous vehicle software for over 20 years. We know the people, the technology and the challenges of the autonomous vehicle industry. We’re eager to help solve this problem because it requires deep and innovative technology, application of thorough engineering practices, and a deep understanding of the industry — a task that complements our backgrounds well.
Our people are our secret sauce. Over our 20+ year experience, we’ve seen the power of different perspectives. We think collaboration, openness, and diversity beats individualism. We have assembled an awesome team to tackle the challenge to provide the mobility industry with safe and reliable software .