Advancing Automotive Software Through Cloud-Based Simulation
The ever-evolving landscape of software development within the automotive industry is no secret, with a growing expectation from consumers for advanced levels of safety, autonomy, and convenience in their modern software-defined vehicles (What is a software-defined vehicle?). Delivering on all three of these pillars simultaneously without compromise is an intricate challenge, but it is one that is made more manageable through one core technique developers have been relying on to optimize their software lifecycle — the use of simulation. According to a June 2023 report by The Brainy Insights, the automotive simulation market is anticipated to reach $5.4 billion USD by 2030, and in this blog, we’ll gain some insight into how and why by first looking at how traditional simulation has enabled companies to:
Then, we’ll explore how cloud-based simulation is poised to amplify these fundamental business values even further — by eliminating the barriers of physical hardware and providing users with virtually unlimited scale and flexibility.
The Need for Simulation
Many cars on the road today contain upwards of 100 electronic control units (ECUs) and an ever-increasing array of sensors, cameras, radar, and LiDAR devices integrated into various systems to enhance the overall driving experience. In an effort to truly enable the potential of the ‘software-defined’ vehicle, OEMs are now working on their next-generation electrical architectures, which consolidate the functionalities of the numerous independent small systems onto a smaller number of high-performance hardware platforms. This consolidation reduces the number of ECUs, but also greatly increases the overall software complexity involved. No software can be shipped without a thorough testing process for safety and performance — and therefore as this surge in code deployment continues, naturally so too does the demand for more efficient testing as companies wish to remain competitive in their time-to-market.
While real-world testing is certainly essential in evaluating how software behaves in actual driving conditions, it is often time-consuming and cost-intensive — requiring physical prototypes, specialized equipment, and additional resources to navigate any associated safety concerns (public risk) and geographical constraints (weather, limited space). This is where simulation comes into play.
Through the creation of virtualized, simulated environments and digital twin devices that replicate highly complex real-world systems, conditions, and interactions, we can run a wide range of tests and scenarios without the need for costly physical hardware setup.
By increasing the amount of testing done earlier in the development lifecycle in simulated environments, we minimize our reliance on extensive real-world testing — reducing costs and accelerating time-to-market. This proactive and comprehensive testing approach also enhances risk mitigation as edge-case scenarios can be tested more thoroughly and bugs can be addressed well before any expensive vehicle recalls or redesigns are required.
While traditional simulation did indeed lift some of the physical barriers hindering more efficient automotive software testing, it too is ultimately bottlenecked by a reliance on the physical machines that are running the simulation technology. In contrast, a cloud approach involves hosting the simulation software and its computing resources on cloud servers through well-established cloud-service providers such as AWS (Amazon Web Services) or Microsoft Azure. This results in 3 key differentiating factors in favor of cloud-based simulation compared to traditional methods:
Whether it’s ADAS enablement, crash detection, or digital instrument cluster verification, embracing the cloud approach to simulation empowers teams to optimize their simulated testing processes through an enhanced level of flexibility and adaptability — ultimately accelerating their development lifecycle.
Importance of an Effective Virtual Lab
Now these benefits are all well and good, but they are only achievable through an often-overlooked component of simulation testing — an effective and easily accessible target-management- system (In the context of embedded software, a target refers to the virtual or physical device under test)
One of the biggest obstacles software developers are commonly faced with is that target hardware (ECUs) and physical labs are not always available in an operational state and may require long waiting times to access. Wind River Studio’s Virtual Lab overcomes this hurdle by providing global, unrestricted access to simulated and physical test environments to all team members through a secure, remote, web-based cloud portal. To see more on how reserving cloud-access to targets works in practice, take a look at this Virtual Lab demo.
Future Trends and Conclusion
One March 2023 study by Peerless Research Group reported that “24% of survey respondents are using the cloud for engineering simulation today, with another 24% planning to use it over the next 12 months,” and given the added flexibility and adaptability leveraging the cloud provides to a team’s development lifecycle, it’s clear to see why this trend is increasing.
With the introduction of even newer technologies like AI and machine learning being used to enable self-driving capabilities, the prevalence of simulation in the auto industry is only expected to increase— making a cloud-based simulation approach more appealing day-by-day.
To learn more about Wind River’s system simulation technology and how it can help amplify your business, explore the following additional resources, or contact us today.
About the Author
Saleh Bhatti is a Product Marketing Intern at Wind River — pursuing his bachelor’s degree in Management Engineering at the University of Waterloo. He is passionate about the intersection of tech and business.