Software-defined vehicles (SDVs) can improve safety, but do they also create new risks?
As the growth of software-defined vehicles (SDVs) transforms the automotive industry, automotive experts are finding that SDVs bring numerous safety benefits as well as risks. With the features and functions of SDVs primarily managed through computer software instead of hardware, the industry is seeking to harness the benefits and mitigate the risks.
Safety benefits of SDVs
Because SDVs feature a central computer and numerous electronic control units (ECUs) that communicate with each other, they can process massive amounts of data collected by sensors, onboard cameras and other devices, leading to a variety of safety benefits. These technologies and other advanced driver assistance systems (ADAS) can help alert drivers of potential collisions with other vehicles, detect pedestrian and roadside obstacles, and, in the case of autonomous vehicles (AVs), take over driving and help avoid incidents.
With the wealth of data collected and ever-growing computing power, SDVs can also use artificial intelligence (AI) to identify traffic and driver behavior patterns and help avoid incidents. AI-powered imaging software can also enhance images during poor weather conditions.
In regions with intelligent infrastructure, SDV systems can be linked with infrastructure systems to provide safety benefits. For example, intersection cameras connected to roadside computers can transmit data to vehicles for navigating intersections. Interconnected systems can account for multiple modes of traffic, including automobile, pedestrian and cyclist traffic and vehicle interactions in multi-lane traffic.
Predictive maintenance is another benefit of SDVs. Onboard diagnostic systems can detect wear and tear of vehicle components in real time to guide repairs or maintenance before problems occur.
Because SDVs offer real-time access to new features, safety improvements can be integrated continuously. As new features are developed, manufacturers can distribute them remotely, enabling the vehicles to be updated without ever entering a shop.
How SDVs create new safety risks
Ironically, some of the same SDV traits that benefit safety also create new safety risks related to functionality and security. Because software can generally be modified more easily than hardware, problems can also be introduced through those modifications — either unintentionally by manufacturers or intentionally by hackers.
Unintentional software glitches may range from minor annoyances to critical safety-related issues. Even with sophisticated quality control systems, the complexity of SDV software means bugs are inevitable. When an issue affects functions such as braking, steering and acceleration, vehicle occupants and others could be exposed to safety risks.
Other unintentional glitches could occur during update and installation processes. For example, if an over-the-air (OTA) update fails due to network or cloud-related problems, the vehicle owner may have limited abilities to remedy the situation.
Problems related to cybersecurity present another set of risks. Hackers can exploit security vulnerabilities to manipulate critical vehicle functions, creating potentially hazardous situations. In certain cases, attacks could render systems nonfunctional or significantly impaired. Network-related issues can also be exploited in systems that lack authentication and encryption.
Autonomous SDVs present additional risks. If sensor data or AI-related models are corrupted, interactions with cameras, sensors and other devices could be compromised, potentially causing the vehicle to misinterpret its surroundings.
Cybersecurity issues could also expose personal data and lead to privacy breaches. Attackers can access data via compromised cloud infrastructure, potentially misguiding critical systems, delaying critical updates, or triggering false alarms, leading to safety concerns and operational disruptions.
Managing SDV safety risks
Automotive experts have been ramping up efforts to mitigate SDV safety risks. Efforts have included new regulations and intensified efforts to improve quality management and security.
Regulatory efforts have been developed at both the national and international levels. The International Organization for Standardization (ISO) established ISO 26262 (road vehicles functional safety) as an international standard for functional safety of electrical or electronic (E/E) systems. The standard defines functional safety for automotive equipment applicable throughout the lifecycle of all automotive E/E safety-related systems, aiming to address possible hazards caused by the malfunctioning behavior of electronic and electrical systems in vehicles.
ISO also developed ISO 21448 (road vehicles — safety of the intended functionality, or SOTIF). According to ISO, this standard provides a “framework and guidance on measures to ensure the safety of the intended functionality (SOTIF), which is the absence of unreasonable risk due to a hazard caused by functional insufficiencies.” Potential insufficiencies could include functionality specifications at the vehicle level or performance specifications of E/E system elements.
Working in conjunction with SAE International, ISO developed ISO 21434 (road vehicles cybersecurity engineering), which provides cybersecurity guidelines for the development lifecycle of road vehicles. The standard is related to a European Union (EU) regulation and the UNECE Cyber Security Management System (CSMS) certification, mandatory for vehicle-type approval and defined in the overarching UN Regulation 155.
The United Nations’ World Forum for Harmonization of Vehicle Regulations (WP.29) introduced an automotive regulatory framework that mandates cybersecurity management system audits for automakers and suppliers. The framework is intended to foster innovative vehicle technologies while continuously improving global vehicle safety, decreasing environmental pollution and energy consumption, and improving anti-theft capabilities.
The European Commission’s Directorate-General for Communications Networks, Content, and Technology established the “Software Defined Vehicle of the Future (SDVoF) initiative” that emphasizes collaboration across European original equipment manufacturers (OEMs) and suppliers. In the U.S., the National Highway Traffic Safety Administration (NHTSA) has published numerous reports addressing new vehicle topics, including cybersecurity, crash avoidance technologies, and safety applications for communications technology.
Bridging the old with the new
The innovative SDV approach will require diligent application of established methods and the introduction of new methods to identify and address risks before they turn into safety issues. This may include more intense testing and quality management practices, as well as new technologies, such as AI-based safety verification. For example, AI methods could use real-world data and machine learning to test systems before implementation.
Other methods might include safety lifecycle analyses, including hazard analyses, risk assessments, and adherence to standards from design through decommissioning; more thorough validation, including simulation and hardware-in-the-loop (HIL) testing; more robust software development practices, including code reviews and automated testing; additional safety architectures employing redundancy and error detection; and secure coding practices to address cybersecurity issues.
Along with technical methods, collaboration among different industry groups will also be key. Automobile manufacturers, OEMs, suppliers, consultants, regulatory agencies, and other stakeholders all play key roles in SDV development and need to share knowledge gained to improve SDV safety.
