Because our autonomous vehicles are extremely heavy and share the road with other people and vehicles, the safety of our semi-trucks is paramount. Our vehicles must operate safely, regardless of any local conditions, including poor road conditions, construction, heavy traffic, cross traffic, pedestrians, cyclists, and non-compliant drivers.
As part of our ongoing safety commitment, we sat down with Adrian Thompson—who recently joined TuSimple as Vice President, Systems Engineering and Test—for a brief interview to understand his stories and experiences that will help to ensure that TuSimple autonomous trucks are as safe as possible.
In this role, Adrian is responsible for developing the Systems Engineering technical capabilities, tools, and processes needed to ensure efficient design and deployment of our autonomous vehicles.
As you can imagine, I’m really excited to have the opportunity to lead one of the most complex systems engineering and safety projects ever undertaken. Every day I’m finding new ways to apply my experience from the aerospace, defense, and autonomous vehicle spaces to the more entrepreneurial task of building the TuSimple Autonomous Freight Network. As a Systems and Safety Engineer, it just doesn’t get more challenging or rewarding.
I think the most important thing that attracted me to TuSimple was its focus. Focus on delivering a meaningful and world-changing new product, and on creating an outstanding business built from the ground up on rigorous systems engineering and safety.
I also really liked the welcoming and inclusive approach the leadership team is taking to attracting and retaining top-tier talent, allowing them to help accelerate the company’s development. It’s surprising: for a company this successful to remain open to—and almost require—great ideas, even from its freshly-minted employees, is excellent. There’s just this pervasive communal spirit to make TuSimple better that I think is rare in companies moving this fast in this type of high-tech business.
Well, a driver-out pilot is a major milestone on the road to developing a truly safe autonomous truck, one that can deliver the promise of autonomous driving across the entire freight network.
In part, it’s a demonstration that we’ve understood the technical details of autonomy and have successfully built safety deeply into our system design. But more importantly, it’s a demonstration that we’ve created the disciplined and agile organization we need to address the next level of technical and organizational challenges that we’ll face in scaling that success to a more continuous operation.
There’s no business or societal value to creating a self-driving truck that doesn’t drive safely. We either build an autonomous vehicle designed from the ground up to contain state-of-the-art safety features, diagnostics, and controls, or we lose sleep every time one of these vehicles enters a real-world environment.
To accomplish the upcoming pilot, my Systems Engineering and Testing team has two primary responsibilities:
We take this responsibility very seriously! Many people on my team live and work in the Tucson area along the route where we’ll be performing this driver-out pilot. If we’re not confident the pilot is safe to perform alongside our families, friends, and colleagues who share these same roads, then our job is not done.
While safety is both a practical and regulatory matter, at its core our gut check is that the safety of our vehicles is essential. This means that we:
This type of rigor takes time and commitment, but the entire TuSimple organization is committed to safety in autonomy as our primary goal.
The key to a strong safety culture is that leadership at every level of the organization is visibly engaged in driving safe designs and operations and leverages as many channels as possible to demonstrate that our highest priority is to ensure that our system is safe.
I think that in this area, TuSimple has been extraordinarily successful, with safety featured as a major topic in every meeting regarding design and operations, and with accelerated hiring in the System Safety and Systems Engineering areas to ensure increased rigor in decision-making across the board.
We’ve implemented consistent and persistent messaging on the criticality of safety to our business at our all-hands meetings and other broad fora and established overall safety ownership at the Executive Team level, with a clear system safety focus by my teammate Al Houry, our Director of System and Functional Safety.
We’ve developed a rigorous approach to system safety combined with a layered approach to safety of overall autonomous operations.
We’ve adopted best practices from aerospace and other safety-critical industries and are open to grass-roots ideas and recommendations about how to improve our system and operational safety. There’s also a strong public commitment to blameless safety concern reporting, and corrective and preventive action implementation.
We’re currently hiring more members of our safety team that will be critical as we scale and enhance our communication, media and training events that can help to embed safety even deeper into the culture of our company and our industry.
Let it not be said that we think the job is done after we complete the driver-out pilot. The way we see it, this is just the start of our long-term safety journey.
There is so much to learn. Obviously, we don’t expect to learn much more about the safety or reliability of our system directly from the pilot runs themselves, because that essential knowledge is a prerequisite to our being ready to perform the pilot in the first place.
However, as mentioned before, this pilot is more than just demonstrating the safety and capability of our system to drive in the real world without supervision. It’s also about us understanding more about what it takes as an organization to define, design, develop, and deploy an incredibly complex device in an agile-but-still-predictable way.
The process of developing the pilot provides a first look at where our processes, documentation, tooling, communications, operations, and contingency management vulnerabilities are, so we can focus our continuous improvement efforts efficiently in those areas to prepare for the next phase of development.
Well, I obviously wouldn’t be exaggerating if I said this technology is going to change the world as we know it.
It’s not just about the AI and the robotics technology itself, which everyone is already impressed with. It’s more about the simpler things like the ultra-high-reliability freight system that it makes possible, allowing people to transport almost anything across the country within 36-40 hours because we are not constrained by driver hours.
That type of reliability and speed means that everything from perishable foods to medicines or other time-sensitive goods can be made more readily available everywhere, anytime.
I wouldn’t be surprised if this results in a complete re-imagining of the entire logistics infrastructure worldwide by 2050, even if only Level 4 autonomy is reached and Level 5 autonomy is much further into the future.
We’ll begin to expect fresher fruit, overnight delivery of everything from groceries to industrial goods, and be able to run the clock of the world economy that much faster.
I’ve built my career working on amazing and truly revolutionary products for government organizations like NASA, aerospace leaders like GE Aviation and United Launch Alliance, and industrial giants like United Technologies and ALCAN Alumina Ltd, but nothing I can think of had this level of potential to change the way we humans live.
Every day I come to the office knowing that the work I do to lead our Systems Engineering and Test team means that:
Without this type of model-based Systems Engineering approach, and a relentless focus on safety, L4 autonomous trucking would be impossible.
Every day, I feel profoundly grateful to play this small but extremely crucial role in bringing L4+ autonomy to the world.
Adrian has an outstanding academic background coupled with professional experience leading the development of safety-critical systems in multiple industries. He was most recently the Director of Systems Engineering Driving Behaviors at Waymo and was Uber ATG’s Head of Systems Engineering and Testing before that. He came to the self-driving industry from aerospace and defense, where he was Director of Systems Engineering and Testing at L3 Harris Technologies. He also served in various technology leadership positions for GE Aviation and United Technologies Research Center. Among his major technical achievements, these are his favorites:
Adrian has a BSc. in Chemical Engineering with specialization in modeling and controls from the University of the West Indies, a Ph.D. in Control and Embedded Systems from the University of Toronto, and an MBA in Finance, Analytic Finance, and Entrepreneurship from the University of Chicago Booth School of Business. He received a Project Management Professional (PMP) certification in 2017 and Lean Six Sigma Black Belt training in 2018.
Adrian lives in Pittsburgh, Pennsylvania with his wife and two children, aged 12 and 15. He loves Caribbean music, soccer, and the outdoors.