Driving Success with the Software-Defined Product

Moving a mechanical industry to SDVs is a solvable challenge. It’ll just take planning and attention to detail.

As more and more advanced ADAS features are developed, tested and implemented, new safety requirements need to be defined to help mitigate risk. (Siemens)

Across industries, more and more products are becoming smart and connected devices. The cellphone is the quintessential example of this development. It has evolved beyond a single-purpose device – to make phone calls – into a Swiss Army Knife for the modern and digital world. But the ubiquity of electronics like cellphones and modern computers has accelerated the potential of many other products and industries. Electrical, electronic, and software systems bring capability to products traditionally developed around mechanical systems: airplanes, heavy equipment, and transportation.

The shift toward smart and connected devices is especially notable in the automotive sector, as cars are being referred to as computers on wheels. (Siemens)

This transition is especially noticeable in the automotive sector, where cars are inching toward being computers on wheels. The addition of software and electronics increases the scope and complexity of the vehicle definition and requirements. It also makes software and electronics engineering core disciplines in the vehicle program. This is a significant change for an industry traditionally dominated by mechanical disciplines.

Adding electronics and software to mechanical and electromechanical development workflows requires a new base of understanding through testing, training, and hiring to gain that experience. Businesses need to be able to integrate these additional workflows without drastically slowing lead times and reducing profitability. As electronics and software become a required dimension of more products we rely upon, the same rigor needs to be applied as it would for mechanical systems.

The benefits and challenges of this transition are evident in automated driving features in vehicles. Lane departure warnings, adaptive cruise control, blind-spot detection, and many other driver assistance features enhance the driver experience but rely on the integrated development of mechanical, electrical, electronic, and software systems to function safely and securely. As automobiles adopt a cellphone-like software-defined focus, it is important to understand the differences in risk. Something going wrong in a cellphone does not have the same impact as it does on a quarter-ton vehicle with human passengers and pedestrians.

To ensure that development remains safe and profitable, it is important to outline a product's needs, expectations, and results as early as possible and provide continuous updates. Designers, manufacturers, suppliers, and OEMs need to adopt software-defined processes to integrate the hardware and software stack into the overall development effort. Software-defined development can help automotive companies deliver products that meet growing customer expectations and comply with increasing safety and security regulations.

Designing a safe product

Vehicles already carry a high degree of risk during operation, and safety has been a part of vehicle development for a long time. There are governmental agencies dedicated to the safety standards of vehicles. But the inclusion of electronics and software in the operation of these products has changed how risk is assessed and where responsibilities lie. Safety needs to be understood in context with other product requirements. The safest car likely wouldn’t move, rendering it unable to fulfill the purpose of transport.

Traditionally, cars have had requirements for pedestrian, passenger, and driver safety in the event of a crash. Adaptive driver assistance systems (ADAS), however, are implemented to avoid collisions altogether. That is a laudable goal, but it also moves regulators, OEMs, and the public beyond their understanding of risk and fault. This disconnect will only widen with time and as autonomy develops if there is no mechanism to characterize, analyze, and iterate on these metrics.

With actionable data and a means to continuously improve their understanding of the situation, manufacturers can bind the software and electronics systems to adhere better to safety requirements. A company, for instance, could require speed limits in adaptive cruise control systems or high-level autonomy systems to pull data from multiple sources to determine a safe speed of travel. Computer vision can acquire data from physical signage, GPS data could provide listed speeds, and other vehicle traffic could help with speed differentials.

Defining safety requirements early on helps mitigate risk while operating the product, and it can streamline the development of software and electronics through a systems approach. If multiple systems require similar, if not exact, data points for operation, a manufacturer can simplify the vehicle architecture and distribute that information to the relevant systems.

Regardless of what a company does to make their vehicles safer in a new world of product development, it starts with the system definition of what is executed in hardware, software, or mechanical realms. From there, designers, manufacturers, and users can help guide development of the products to best reflect the needs of the complete system.

The most powerful capability is the holistic approach of the digital enterprise to create digital twins in the automotive industry. The comprehensive digital twin represents all aspects of today’s vehicles: software, electronics, mechanical and electrical wiring systems. It becomes possible to predict the performance of the production unit and the products themselves, both during development and then across a vehicle's lifetime. Comprehensive digital twins ensure OEMs produce what customers expect in terms of individualization and drive concepts.

The digital twin is the precise virtual model of a vehicle. It displays development throughout the entire lifecycle and allows operators to predict behavior, optimize performance, and implement insights from previous design and production experiences. Car manufacturers obtain direct feedback on their actions. Whenever they change settings or create new scenarios, they can simulate the impact to make better and more confident decisions.

Designing a secure product

Electronics and software, especially internet-connected varieties, provide many ease-of-use and convenience features to new car designs. But these also bring in additional risks.

Traditionally, a car has needed to be safe in a collision and secure from mechanical intrusion, but software brings the need for cybersecurity. For automated vehicles, this risk involves potential remote control of the car, but even for non-automated vehicles, there can be a risk of bypassing electronic or software locks on the vehicle or intercepting/interfering with over-the-air (OTA) update systems.

Resolving the hardware and software stack in a vehicle is one of the biggest challenges. On top of the hardware systems in the vehicle and its digital twin, there is embedded software running critical systems as well as application software to control the infotainment features in the cabin. Some of these systems might be created in-house, but many will rely on suppliers to deliver key software and subsystems. Ensuring security across the entire vehicle means validating the software across the digital enterprise to mitigate vulnerabilities within individual systems and their interactions with the rest of the vehicle. This is where extensive traceability and continuous validation in the digital twin help secure a vehicle. By defining the intent and validating the system in the digital twin, OEMs can ensure security.

In the event of a missed variable or an unforeseen requirement in the system, it is valuable to update and repair vehicle software after the vehicle has left the lot. OTA updates can correct security holes, add features, and even correct operation characteristics, but without proper oversight, OTA updates could create another vector for intrusion. Vehicles need to be able to validate that the updates they are receiving are the right ones, and OEMs need to know what vehicle an update is going to, as models and revisions can impact the necessary fix. This makes data retention and accessibility extremely important across the entire digital enterprise. An OEM might need code revisions from one supplier, but another making the part for another region or model is fine. Being able to track and act on data produced throughout the value chain is the heart of becoming a digital enterprise.

Security is not just limited to vehicles and products. As companies share data with their supplier network to hit targets and regulators to confirm compliance, there is a risk of exposing intellectual property (IP). Whether it is the specifications for the vehicle itself, the production processes involved, or any other form of protected IP, it is essential that making details available to necessary parties does not become a breach. This is where comprehensive digital solutions can be so helpful. Rights access can be assigned to different data pools or types and different users can be provided a subset of data that is strictly relevant to their role.

Designing a successful product

The realm of product design, production, and operation has changed significantly in the past few decades, and the expectation is that it will keep evolving. Electronics and software are the newest domains of importance in these complex products. They bring both benefits and challenges to products and product development. A software system can be more easily updated to reflect a new understanding of the product and its environment, but it is a drastically different engineering discipline for OEMs that started with primarily mechanical products. Electronics can provide precise, accurate, and repeatable control of a product, but they must be engineered to survive in harsh conditions and for many years.

There are always tradeoffs to be made in a product, its design, and its production, but by shifting to a software-defined product methodology, a business can readily outline, validate, and iterate on the best solutions in the digital twin. This does not just apply within the factory environment. With secure data from the field, a digital enterprise can optimize the functionality of existing fleets or fix errors in software systems from edge-case scenarios as they are uncovered to improve safety. As business models change for these vehicles, OEMs can provide ongoing support through equipment leasing programs similar to those in the realm of heavy equipment. The flexibility and agility afforded by digitalization apply to the product and the process. Becoming a digital enterprise is critical in developing successful transportation and other similarly positioned products at the speed and scale demanded.

Nand Kochhar is vice president of automotive and transportation for Siemens Digital Industries Software. Dale Tutt is Siemens’ vice president of industry strategy.