The V&V process for autonomous vehicles will be a daunting but necessary task.
Siemens has sponsored this post.
Written by Nand Kochhar, vice president of Automotive and Transportation for Siemens Digital Industries Software.
The U.S. autonomous car market size is estimated to reach $37.56 billion by 2029, growing at a CAGR of 20.5% between 2024 and 2029. However, the widespread implementation of automated vehicles (AVs) in the real world (Figure 1) depends on the verification and validation (V&V) for designing, manufacturing and of course, operation. Hence, AVs must safely and reliably drive on roads in all weather and traffic conditions on urban, suburban and back-country roads.
To continue the march towards autonomy, companies must transform the methods with which vehicles are developed and put into the market. While great opportunity exists in the future, companies face several impediments to supplying connected, automated and software-defined vehicles to the world.
Building AVs is a particularly complex process because the entire vehicle is a system of mechanical, electrical, electronic, network and software systems. State-of-the-art components from each of these domains are required to create the most sophisticated vehicles ever produced.
Thus, the V&V of AV safety, reliability and performance in all traffic scenarios is a daunting task. Projections indicate AV platforms will need to complete the equivalent of billions of miles of testing to ensure their safety and reliability. And, because the vehicle involves interconnected and interdependent systems, the complexity compounds.
Connecting real world and simulation data via digitalization
The development of advanced driver assistance systems and AVs is a data-driven engineering process. Numerous measurements are generated, analyzed and incorporated back into the design at each step in the lifecycle. Translating the raw data gathered into engineering insights that drive improvements and optimizations is where competitive advantages are won and lost.
Companies can successfully address the complexities of these challenges by using a mixed-reality, digitalized approach to the development, building, verification and validation of their vehicles and driving systems.
Real-world data collection is critical to providing accurate V&V. Typically, the information collected from a physical test is immense. Data-collection software can perform an initial analysis to distinguish what is pertinent to the testing objective at hand. This enables teams to make well-informed decisions on data storage and processing priority.
Essential hardware elements, sensors, actuators, controllers or complete autonomous driving systems need to be tested. Hardware components are verified and validated using simulators to mimic on-road driving scenarios with varying environments, traffic and road conditions.
Hardware-in-the-loop simulation software can help with the testing of sensor systems. These solutions can test camera-based perception systems used for advanced driver assistance systems by integrating the actual camera sensor into the testing environment. This helps increase simulation fidelity because the system is processing real sensor data. Camera projection boxes, as well as camera injection setups, allow engineers to test the camera-based perception systems under challenging conditions.
This information becomes more powerful when converted into the virtual domain. Real-world tests enrich and inform simulated vehicle environments and driving dynamics. Scenarios captured during real-world testing can be imported and recreated in a simulation solution (Figure 2). Within the simulation environment, engineers can change parameters of the scenario (such as weather conditions or vehicle speeds) so that they can interrogate all facets of system performance in different driving environments.
Using data from physical testing and simulation, AV engineers can quickly identify on-road edge cases and assess vehicle behavior in all driving scenarios. As the AV is an integrated system, its development platform also needs to be integrated to test and retest the operation of the vehicle in realistic virtual scenarios throughout the design process. By implementing both design and simulation on one platform, test results and simulation data can be readily reincorporated into the vehicle design. This produces a closed-loop feedback system that improves not only the design but also operation and physical characteristics of the vehicle.
When this approach is used, a comprehensive digital twin of the AV then can be created. This empowers an efficient closed-loop AV development lifecycle that spans from design to verification and validation and even through to in-field maintenance.
Encountering unsafe scenarios in a virtual environment, not on the road
High-fidelity simulations using a comprehensive vehicle digital twin also provide a virtual environment for identifying unknown unsafe scenarios. Engineers can combine the knowledge from known real-world situations with mathematical prediction and simulation methods to uncover possible alternative critical scenarios. They can discover and analyze these scenarios more efficiently in a virtual environment, reducing the number of unknown-unsafe scenarios and the risk incurred when deploying AVs.
As stringent regulations are adopted by governments that focus on road safety, they will likely guide the future of virtual vehicle testing and simulation technology standards to help streamline consistency and acceptance of AV certification.
Collaboration is key to building confidence
In the United States and worldwide, standards were developed for vehicle certification. As regulations matured, confidence in the products was built. The engineering environment for AVs also needs to build that confidence, to prove the connection between physical testing and virtual testing so that the authorities can confidently approve standards that will benefit everyone, manufacturers and customers alike.
Three areas are crucially important to fully autonomous vehicles being introduced successfully: public acceptance, technology and regulations. The automotive and transportation industries have a challenging task ahead of them to address the needs of all three areas.
Standards should help balance out this tension between technology and regulations. They would help guide companies, like Siemens, that develop tools for AV manufacturing companies. These tools could then help ensure that the standards are in alignment with the technology that they are developing and vice versa. Standards, and the verification and validation they call for, may be driving the speed with which we will see AVs become common.
The process of designing, building, testing and scaling autonomous cars is complex and time-consuming. As such, advancement must rely heavily on collaboration and partnerships, which will drive innovation, improve functionality and ensure safety for everyone on the road.
About the Author
Nand Kochhar is the vice president of Automotive and Transportation for Siemens Digital Industries Software. He joined Siemens in 2020 after nearly 30 years with Ford Motor Company, where he most recently served as Global Safety Systems Chief Engineer. In this capacity, Kochhar was responsible for vehicle safety performance of all Ford and Lincoln brand products globally. He also served as Executive Technical Leader, CAE and as a member of Ford’s Technology Advisory Board. Kochhar’s tenure at Ford also included executive engineering leadership across a range of disciplines including in product development, manufacturing, digitalization, simulation technology development and implementation.  
