The third edition of UL 4600 was released in 2023 to add specific requirements for the use case of autonomous trucking and to address changing industry trends.
Here, we explain what’s happening in the autonomous trucking industry, why UL 4600 is important for autonomous trucking specifically, and how static analysis helps to overcome challenges for this evolving technology.
Read on or jump ahead to the section that interests you most:
➡️ start Your Free Static Analysis Trial
While some just dream of fleets of autonomous trucks efficiently delivering goods across the country, others are already at work to ensure they can do so safely and resiliently. With the update to ANSI/UL 4600, the Standards for Safety for the Evaluation of Autonomous Products, Edition 3, in play, embedded software teams now have better safety guidance just as self-driving technology is ramping up to make shipping faster, more cost-effective, and more efficient in the face of driver shortages and rising transportation costs.
Autonomous trucking software is on the cusp of widespread adoption. To prepare for the inevitable transition from closed-course testing to widespread deployment, OEMs need to understand the current state of trucking automation and how to effectively implement software safety practices.
What Is Autonomous Trucking?
Autonomous or driverless trucks operate with minimal to no human input. Instead, autonomous trucks or autonomous tractor-trailers rely on sensors — usually combinations of cameras, LiDAR, and radar — to feed environmental data into algorithms and actuators that control the vehicle.
The Society of Automotive Engineers (SAE) have defined levels of driving automation that are adopted across the industry, which range from Level 0 (completely manual) to Level 5 (completely autonomous). While the lower levels with an automated driving system (ADAS) are already in play for autonomous trucks, software development teams will need to ensure functional safety for the higher autonomy levels, ideally Levels 4 (High Automation) and 5 (Full Automation) for the freight industry to benefit from autonomous trucking.
Autonomous Trucking Development Today
Trucking is the dominant mode of inland freight transportation in many countries, with several trucking companies competing to be the first to operationalize autonomous driving. Driverless technologies offer significant potential value for fleet operators: They can reduce operating costs, overcome driver shortages, and improve efficiency.
For example, driverless semi trucks (or lorries) can employ truck platooning more effectively than human drivers, where vehicles follow each other at the same speed to improve fuel economy and reduce their impact on traffic.
While there have been a few setbacks in recent years — for example, TuSimple, Navistar and UPS shut down their “Driver-out” self-driving truck system in 2023 and Waymo, Embark and Locomation are no longer actively developing autonomous trucks — there are many more new entrants working toward wide-scale deployment:
Traditional OEMs are nearing full operationalization, with trucks already being tested in North America. In partnership with Aurora, Volvo revealed its first production-ready autonomous truck in May, and Daimler Truck reported that its Freightliner Cascadia semi-trucks are meeting closed-course acceptance tests in October of 2024.
These announcements indicate that the automotive industry is driving autonomous trucking plans forward — so software development teams in automotive manufacturing will need to get familiar with the challenges, solutions, best practices, and compliance with UL 4600 to ensure they are prepared for these exciting advancements.
Trucking Automation Software Challenges
There are five main challenges for teams developing autonomous trucking software:
It’s hard to reuse autonomous software built for cars.
The development and testing of self-driving cars focus on short, low-speed routes with stop-and-go traffic. In contrast, autonomous trucks will operate on long-haul highway routes at higher speeds and will encounter less traffic, more variable terrain, and transitions between urban and rural roads.
More critically, a Class 8 driverless semi-truck has a gross vehicle weight eight times that of the average passenger vehicle — before loading its cargo. This means autonomous driving software has to account for a larger turn radius, longer stopping distances, and the presence of a trailer that can weigh as much as 14,000 U.S. pounds unloaded.
Software has to handle many use cases.
Trucking automation software must accommodate different vehicle classes, cargo conditions, and route types. While some fleets deliver consumer packaged goods within cities, most autonomous trucks are particularly good for long distances, which will need to be accounted for in the software. Still others transport hazardous, refrigerated, or liquid materials across international borders.
There is also the scalability factor: To build and test multiple branches of software to handle various scenarios at scale, developers should design systems to accommodate a wide range of input and control conditions before deploying them into real-world trucking environments.
Verification and validation require long highway driving.
Once acceptance tests are performed on closed-loop circuits, autonomous trucks must conduct road tests on highways, ideally for hundreds of miles across the country. That way, developers can ensure their autonomous trucks can cover the distances, runtime, and road conditions necessary for the long haul. The U.S. Department of Energy reports that the average semi-truck travels over 62,000 miles annually.
Security must be a top priority.
Similar to their automotive counterparts, driverless trucks must address the following security concerns:
- Protect connected infrastructure and endpoints such that there is acommon baseline of trust between nodes.
- Track and adapt to vulnerabilities that will continue to grow as malicious actors realize new opportunities to attack and destabilize critical trucking networks.
- Secure the manufacturing supply chain with vendors who may not have had to deal with software security before.
- Comply with cybersecurity regulations and best practices, such as ISO/SAE 21434.
Safety compliance must be accounted for.
Safety will be the key differentiator between manufacturers that make it to market and those stuck in verification and validation activities. Fully unmanned trucks operating at highway speeds present real concerns, including difficulty in handling unexpected situations and decision-making capabilities, in addition to more typical concerns about undefined behaviors and software malfunctions.
The challenges of developing safe trucking automation systems lie in their components. Everything from sensors to decision-making algorithms to vehicle motion control must be scrutinized. Given this complexity, manufacturers will find themselves relying on automated tools, like Perforce Static Analysis, to help with UL 4600 compliance.
Understanding UL 4600: Safety Principles and Processes for Autonomous Trucking
UL 4600 is the first safety standard designed specifically for autonomous and connected vehicles. Unlike a traditional UL satefy standard, UL 4600 takes a “safety case” approach with real-world applications in a specific environment — and the inclusion of autonomous trucks in UL 4600 Edition 3 includes trucking-specific examples. The standard helps developers build a safety case for carious aspects of system development and maintenance:
“It offers framework that leads designers of autonomous systems through the required thought process to ensure all possible complications have been considered. What are the safety questions that need to be considered in design? How do you think beyond design and for the lifecycle of the vehicle? Can quality and consistency be assured across manufacturers?
— Dr. David Steel, Executive Director of UL Standards & Engagement, ULSE, Inc.
The UL 4600 Edition 3 standard requires developers to follow a three-step approach for assessing and validating driverless truck safety:
- Make a measurable safety claim, where developers state how the autonomous truck should operate.
- Make an argument that proves the claim is true by describing the perception technologies and the systems that are triggered by them.
- Provide evidence that the system will perform as expected by providing simulation results, road test outcomes, and other proof that the autonomous truck will perform as stated.
The end result is a safety case arguing that an exceptionally robust combination of analysis, simulation, closed course testing, and public road testing have been performed — with evidence given — to ensure an appropriate level of system safety.
How Static Analysis Helps Achieve Autonomous Truck Safety
There is a specific requirement in UL 4600 for coding standard compliance, as the development process should be similar to that of IEC 61508 or ISO 26262, so developers should use static analysis to some degree and produce the results of source code analysis. Static code analyzers — like Perforce Helix QAC and Klocwork — support these goals by ensuring comprehensive code coverage and sufficient supporting evidence in these areas:
Amid the pressure of getting driverless trucks to market, these tools enable developers to focus on feature development rather than compliance activities.
We are delighted to announce our acquisition of Delphix, a best-in-class leader in Enterprise Data Management solutions. I want to share with you why I am personally excited about this major milestone in our company’s continued DevOps evolution and the benefits this acquisition provides to our customers.
Data is at the heart of how enterprises operate today and essential for successful software development, but accessing and managing that data is extremely challenging. Many teams do not have rapid access to solid, high-quality test data. Imagine something the size of a relational database, with all the data to collect and piece together to make it testable — this is both labor-intensive and very difficult to achieve.
All that changes with Delphix. This truly outstanding platform provides on-demand, easy access to data very quickly and in a safe way. Delphix protects and masks customer data giving teams the right data, securely and quickly, so they can focus on creating great software.
