Vehicles of The Future: Weiyang Zhang Of TuSimple On The Leading Edge Technologies That Are Making Cars & Trucks Smarter, Safer, and More Sustainable

An Interview With David Leichner

Solid technical judgment: Sound technical judgment is indispensable when designing or troubleshooting algorithms. When tasked with incorporating a new function in an algorithm, it’s vital to weigh short-term against long-term solutions, and sometimes, sift through multiple implementation methods. A well-honed technical acumen helps in discerning the most effective approach, balancing timelines with the pros and cons of each method.

The automotive industry has been disrupted recently with new exciting technologies that have made cars and trucks much smarter, much safer, and much more sustainable and more environmentally friendly.

What other exciting disruptive technologies will we see in the next few years? How much longer will fossil fuel powered cars be produced? When will we see fully autonomous vehicles? Can we overcome the challenge of getting stuck in traffic? As cars become “moving computers”, do we have to worry about people hacking our cars? How else will our driving experience be different over the next five years? To address these questions, Authority Magazine started a new interview series about “Exciting Leading Edge Technologies That Are Making Cars & Trucks Smarter, Safer, and More Sustainable.” In this series we are talking to leaders of automotive companies, automotive tech companies, EV companies, and other tech leaders who can talk about the vehicles of the future. As a part of this series, I had the pleasure of interviewing Weiyang Zhang.

Throughout his nearly four years at TuSimple, Weiyang Zhang has focused on the development of advanced decision-making, motion planning, and control algorithms for long-haul self-driving trucks. These algorithms are designed to safely interact and negotiate with human drivers on highways. Notably, he helped develop a patented algorithm for “proactive lane changing,” enabling TuSimple trucks to safely and efficiently perform lane changes. Before joining TuSimple, Zhang worked at the Isuzu Technical Center of America (ITCA), where he crafted a specialized coverage motion planning algorithm that efficiently navigates entire grounds while dynamically avoiding obstacles. This algorithm has been successfully deployed in Isuzu trucks and resulted in two scholarly publications in the Society of Automotive Engineers (SAE). Zhang holds dual master’s degrees in electrical and computer engineering and mechanical engineering from the University of Michigan, Ann Arbor.

Thank you so much for joining us in this interview series! Before we dive in, our readers would love to “get to know you” a bit better. Can you tell us a bit about your ‘backstory’ and how you got started in the automotive industry?

I’ve always been enamored with engineering due to its capacity for tackling complex challenges and generating substantial societal impacts. As I started my graduate studies at the University of Michigan, Ann Arbor in 2017, autonomous driving technology emerged as a hotspot, with the university boasting a plethora of related resources and research projects. This domain particularly aligned with my interests as it presented technical challenges, industry potential, and positive societal impact. It’s an exciting fusion of classic vehicle dynamics and control with emerging technologies like AI and robotics. This inspired me to pursue dual master’s degrees in mechanical engineering and electrical and computer engineering, establishing a solid foundation for my ventures into the autonomous driving domain.

I joined the TuSimple planning and control team in 2020, concentrating on the development of sophisticated decision-making, motion planning, and control algorithms for long-haul autonomous trucks. My focus has been on addressing interactive driving behaviors on the highway, including lane-changing and merging, to enable the self-driving truck to navigate safely and smoothly.

Can you share the most interesting story that happened to you since you began your career?

The most interesting event unfolded the day when I sat in a TuSimple truck, testing the algorithm I had developed for autonomously handling merging scenarios. With heavy traffic ahead and numerous vehicles on the ramp, the truck was tasked with making proper decisions: either yielding to or overtaking for each on-ramp vehicle. As I nervously monitored the truck’s performance, the safety driver and I discussed potential actions. He told me about the action he would take under this scenario when driving manually which was different from the action that my algorithm decided to execute. He was prepared to take over anytime he felt necessary. However, it was very interesting and impressive to witness my algorithm managing to navigate through the scenario, accurately predicting the intentions of on-ramp vehicles and executing appropriate actions correspondingly. This test revealed that my algorithm could potentially outperform experienced drivers in some aspects, underscoring the promise of autonomous driving technology when paired with precise perception systems and well designed, decision-making algorithms. This real-world success not only validated our hard work but fueled my passion for further advancing autonomous driving technology.

Ok wonderful. Let’s now shift to the main focus of our interview. Can you tell our readers about the most interesting projects you are working on now?

The most interesting project I’m currently working on is developing a decision-making and motion planning algorithm for highway lane-changing scenarios. This can be considered as the second stage of the “proactive lane change” algorithm. Before delving into that, let me first introduce the original proactive lane change algorithm to the readers.

Traditional lane-change algorithms for self-driving trucks employ a “lane-following and wait” strategy, requiring the truck to wait passively for a naturally occurring gap in the target lane. This often leads to missed lane-changing opportunities and operation inefficiencies. To tackle this challenge, the “proactive lane change” algorithm was devised. This approach proactively searches and selects feasible windows in the target lane for the truck to move into, based on a comprehensive set of factors including the truck’s current state, the dynamic behavior of vehicles in the target lane with future prediction based on deep learning, as well as mandatory safety and traffic law requirements. Once a feasible window for lane change is locked, the algorithm employs advanced motion planning and control techniques that command the truck to make active adjustments in its longitudinal position and speed, aiming to align itself perfectly with the selected window while ensuring these adjustments are executed smoothly and safely. Furthermore, I want to highlight that this algorithm played a vital role in TuSimple’s first driver out, fully autonomous driving test on public roads — a global first! It’s very exciting to see my algorithm enabling TuSimple’s self-driving trucks to seamlessly negotiate with other human drivers, without any human interventions, and thereby successfully making the lane change.

Now, onto the next phase which is both challenging and intriguing. It aims to extend the algorithm to handle non-compliant drivers — those who disregard right-of-way rules, speed excessively, or exhibit erratic driving behaviors. These behaviors are unforeseen and hard to predict and could even be challenging for experienced human drivers to handle, making algorithm development a complex task. I’m very excited to participate in the development of this algorithm for its challenging nature and its essence for safe deployment of self-driving trucks on the road. I am currently collaborating closely with my teammates, exploring new algorithms by combining machine learning and optimization theory, and striving for a synthesis of intelligent and safety-guaranteed behavioral planning and execution.

To ensure our development is on track and safe, we have been developing our algorithms by following TuSimple’s rigorous development pipeline. It encompasses design review, algorithm implementation, unit tests, simulations, road test scenario regenerations, and on-road tests. Each phase in this pipeline not only boosts the development of the algorithm but also ensures that we are inching closer to the real-world deployment of safe and efficient self-driving technology. With the collective expertise and relentless efforts of our team, I am confident that the algorithm we are developing will significantly enhance the decision-making capabilities of our self-driving trucks, making them safer and more intelligent for all road users.

How do you think this might change the world?

The development and refinement of proactive lane-changing algorithms, especially in handling non-compliant drivers, driver’s gestures and construction information, holds significant promise for the safe deployment of self-driving trucks.

By efficiently managing complex and unpredictable traffic scenarios, these advancements can contribute to reducing road accidents, enhancing traffic flow, and potentially saving lives. Additionally, by demonstrating successful real world applications of autonomous driving technology, it propels the automotive industry closer to a future where autonomous vehicles are a reliable and commonplace mode of transportation.

Keeping “Black Mirror” in mind, can you see any potential drawbacks of this technology that people should think more deeply about?

Certainly. The exploration of autonomous driving technology, especially developing decision-making algorithms for autonomous vehicles, can sound like you’re opening a Pandora’s box of ethical and security concerns, much like scenarios depicted in “Black Mirror” and some fiction movies. In a world where technology governs critical aspects of daily life, the potential for malfunctions or cyberattacks poses serious risks, especially on the road where split-second decisions matter.

One of the primary ethical concerns revolves around machine versus human judgment in critical situations. For instance, how an autonomous vehicle responds to unforeseen obstacles or makes decisions during a potential crash scenario. The ethical dilemma extends to who is responsible in accident scenarios — is it the algorithm, the developers, or the vehicle owner?

Moreover, if not designed correctly, autonomous vehicles could become targets for cyberattacks. The reality is that as vehicles become more connected, the attack surface for hackers expands. This must all be considered when designing a safe autonomous vehicle, which TuSimple has done. Specifically, we’ve instituted cybersecurity measures to protect our autonomous driving system from malicious attacks, following ISO 21434 guidelines. Specifically, we employ continuous monitoring services to detect any abnormal software events, ensuring a proactive identification of potential threats. Concurrently, all communications are encrypted to safeguard our onboard systems from unauthorized control or data breaches by hackers, providing a robust shield against cyber intrusions. In case of a cyber incident, our system auto-logs key data, aiding swift analysis, recovery, and fortification of our cybersecurity measures.

Furthermore, the technology could be leveraged for surveillance purposes. With an array of sensors and cameras, autonomous vehicles could be used to collect an unprecedented amount of data on individuals and communities. This raises privacy concerns, and without robust data protection measures that companies like TuSimple implement on every system, individuals’ privacy could be infringed upon.

As this technology develops, it’s imperative that companies meticulously dissect these ethical and security challenges while not letting them become roadblocks to development. We should embrace these challenges head-on. After all, new challenges often unveil new horizons, catalyzing innovations like the birth of new industries such as autonomous vehicle maintenance, the development of secure infrastructure for autonomous navigation, and new job roles in monitoring and managing autonomous fleets. These changes could foster a new ecosystem of innovation and opportunity. I’m very optimistic for the future and I would recommend everyone to fasten their seatbelts and enjoy the journey.

What are a few things that most excite you about the automotive industry as it is today? Why?

The automotive industry is undergoing a riveting phase of evolution that is poised to redefine mobility and transportation. Here are a few key aspects that excite me the most:

1. Advancements in autonomous and assisted driving: The growing focus among OEMs to develop Advanced Driver-Assistance Systems (ADAS) and fully autonomous driving systems is exhilarating. This shift is propelling a surge in the development of autonomous driving related algorithms, which in turn, advances the growth of autonomous driving technology. The integration of cutting-edge technologies like AI and machine learning fuels these advancements further, promising a future where autonomous vehicles could significantly enhance road safety and traffic efficiency.
2. Shift towards clean energy vehicles: The increasing emphasis by OEMs to develop clean energy vehicles, such as electric and hydrogen-powered vehicles, is a pivotal step towards sustainable transportation. This shift not only addresses environmental concerns but also showcases the industry’s commitment to innovation and a greener future.
3. Collaborations and partnerships: The increasing number of collaborations between automotive companies, tech giants, and startups is fostering a culture of open innovation. These partnerships are accelerating the pace of advancements in autonomous driving, electric vehicles, and smart mobility solutions, promising a future of integrated and sustainable transportation ecosystems.

What are a few things that most concern you about the automotive industry as it is today? What must be done to address these challenges?

Well, there are a few aspects that currently concern me about the automotive industry, which pose challenges that need addressing to ensure the smooth progression of autonomous vehicle technologies.

1. Standardization and legal framework: As we previously discussed, there’s a pressing need to establish a unified standard for autonomous vehicle algorithms, alongside a clear legal framework addressing liability and outlining the criteria for different levels of autonomy. The lack of standardized terminology and varying regulations across different regions or companies hinders collaborative efforts and confounds the public’s understanding of autonomous technology. This requires a concerted effort from the entire industry to develop and adhere to common standards and terminologies, and to work with regulatory bodies to create clear, supportive legal frameworks.
2. Financial investment challenges: Developing autonomous driving technology demands a substantial financial investment. The challenge for startups, especially in the current macroeconomic climate, is to secure adequate funding to continue their innovative work. On the other hand, large OEMs might also be deterred from investing heavily in autonomous driving technologies due to a focus on short-term profits, which could, to some extent, stifle the advancement of autonomous driving.
3. Hardware costs: The cost of autonomous driving vehicles is also a major concern as it necessitates the use of high-resolution LiDAR, multiple cameras, GPS, and high-performance SoCs. These hardware requirements significantly drive up the cost of autonomous vehicles, making them less accessible to the average consumer and potentially slowing the adoption and expansion of autonomous driving technology.

Based on your vantage point as an insider in the automotive industry, what other exciting disruptive technologies will we see in the next few years? Can you share some of the new developments that will make vehicles smarter, safer, and more sustainable?

Of course! There are numerous efforts underway in the automotive industry aimed at enhancing vehicle performance. Here are some highlights:

1. Clean energy vehicle development: The automotive industry is indeed gravitating towards the development of new energy vehicles like electric or hydrogen fuel cells. Particularly for electric vehicles, advancements in battery technology are crucial. Recent innovations such as solid-state batteries, lithium-sulfur batteries, and sodium-ion batteries are game-changing. They promise higher energy density, improved safety, lower costs, and better environmental performance compared to traditional lithium-ion batteries​.
2. V2X infrastructure development: The integration of Vehicle to Everything (V2X) technology is another exciting frontier. By enabling vehicles to communicate with each other and with infrastructure like traffic signals, V2X technology can significantly enhance the intelligence of autonomous vehicles. It facilitates early danger detection, smoother traffic flow, and a more efficient transportation experience by providing real-time traffic data.
3. Manufacturing innovations: Innovations in manufacturing design, like modular units and one-piece body construction, are also on the horizon. Modular design allows for more efficient manufacturing processes, cost reductions, and quicker adaptation to market demands​​. On the other hand, the one-piece body manufacturing, also referred to as mega- or giga-casting, has been recently spotlighted by Tesla’s efforts. Tesla’s giant casting machine can produce the rear underbody frame of a car in one piece, reducing the number of parts and manufacturing costs by 40%​​. This methodology could revolutionize body construction, making vehicles cheaper and quicker to produce.
4. Enhanced human-self-driving vehicle interaction: With the spread of robotaxis, the development of high-performance autonomous driving vehicles and services becomes more important. These advancements are geared towards ensuring a user-friendly and convenient experience for self-driving passengers, with interactive interfaces and real-time adaptability to passenger needs and preferences, thus elevating the overall passenger experience in autonomous vehicles​. Just imagine the cool experience of sitting in a self-driving car that navigates based on your preferences, while the lighting, music, and atmosphere are all tailored to your liking.

These technological advancements signify a shift towards a more connected, safe, and sustainable automotive landscape. They not only contribute to the evolution of autonomous driving but also address broader societal and environmental challenges posed by traditional automotive systems. I’m personally looking forward to seeing all of these innovations become a reality.

In your opinion, how much longer will fossil fuel powered cars be produced? When do you think EVs will be the majority of vehicles in use? Can you explain?

Honestly, the timeline for the phasing out of fossil fuel-powered cars and the shift to EVs is subject to a variety of factors. I can only give a very rough estimation that this transition could become apparent over the next decade or two. The rate of adoption for EVs is expected to increase significantly in the coming years, with some projections indicating that EV sales could surpass those of fossil fuel vehicles by around 2033 in major auto markets, and continue to grow thereafter.

However, the exact timeline is contingent on several factors including regional policies, automaker commitments, market dynamics, infrastructure development, technological advancements, and cost factors. For instance, aggressive policies in regions like California are pushing for a substantial portion of new vehicles to be electrified by 2035, and major car manufacturers like General Motors have pledged to phase out fossil fuel vehicles by 2040. Moreover, the continued development and deployment of charging infrastructure, advancements in battery technology, and the reduction in costs associated with EVs are crucial elements that will influence the pace of this transition.

On a personal note, being a person with a primary background in mechanical engineering, I have a certain appreciation for the engine. Looking ahead into the future, it’s conceivable that while the majority of vehicles on the road might be electric, high-performance engines could still find their place in specialized grounds, much like how equestrianism is enjoyed today. The thrill of driving a high-performance engine vehicle could become a niche, yet cherished experience. This unique blend of tradition and technology might foster a new culture of automotive enthusiasm, preserving the legacy of traditional automotive engineering while embracing the inevitable electric future.

When do you think we will see fully autonomous vehicles deployed in a mainstream way? What do you think are the main barriers to reaching that stage?

Actually, the transition toward mainstream deployment of fully autonomous vehicles (AVs) has indeed started. The process is expected to be gradual initially, but there may come a point where suddenly, we’ll find fully autonomous vehicles becoming commonplace on our roads. For example, leading players in autonomous passenger cars have already rolled out robotaxi services in several major U.S. cities and are steadily scaling up. In the realm of self-driving trucks, our company has successfully conducted the first driver out (no safety driver in the vehicle) demo on public roads in the U.S. Given these advancements, over the next 10 years I do believe we will see more widespread adoption of fully autonomous vehicles.

Regarding barriers, several factors could potentially delay the mainstream deployment of fully autonomous vehicles. As previously discussed, the establishment of standard and legal frameworks, the high costs associated with autonomous vehicles, and the confidence of investors in this technology are notable challenges. Furthermore, public perception also plays a crucial role. For example, some individuals are conservative or skeptical about self-driving technologies. Despite these hurdles, I remain confident that they will be surmounted in due time. Being a participant in this transformative journey, I am thrilled to have a front row seat to witness and contribute to overcoming these challenges, propelling us toward a future where autonomous vehicles are an integral part of our transportation ecosystem.

How else will our driving experience be different over the next five years?

Well, over the next five years, we will gradually see more self-driving vehicles on the road. I can anticipate that some readers might encounter our trucks on the highway in the next few years. Driving along with autonomous vehicles will gradually become a common scene. Interactions between human drivers and self-driving vehicles could lead to safer roads for all.

Speaking of safety, as autonomous vehicles are designed with safety as a paramount priority, we could expect to see an uptick in safety measures on the road. This could stem from the technology of the vehicles themselves and possibly from stricter road safety regulations to accommodate autonomous driving. Furthermore, with the introduction of V2X communication technologies, traffic management could be significantly improved. This technology promises to reduce congestion and improve the overall flow of traffic by allowing vehicles to communicate with each other and with traffic infrastructure.

Moreover, as cars become more autonomous, the in-car experience will also evolve. I expect people could use their commute time more productively or for relaxation, as they won’t need to focus on driving. The interior design of vehicles may also change to accommodate these new ways of spending time on the road.

Eventually, I would expect that self-driving vehicles can blend well into day-to-day traffic, and we can negotiate and interact with them no differently than we do with human drivers. The driving experience is bound to change in positive directions, and I’m looking forward to seeing how these advancements will shape our daily commutes.

What are your “5 Things You Need To Create A Highly Successful Career In The Automotive Industry?

Drawing upon my own journey in the automotive sector, particularly in the realm of autonomous driving, I believe the following five elements can significantly contribute to carving out a successful career in this dynamic industry:

Perseverance: In the domain of autonomous driving, algorithm development often requires a relentless pursuit of precision and efficiency. For instance, the “proactive lane change” algorithm I developed underwent several iterations, scrutinized by different sets of eyes and had to meet a myriad of safety standards. Despite the demanding process that spanned months, the eventual deployment of the algorithm in trucks was immensely gratifying. The journey underscored the importance of patience and tenacity in seeing a project through to its fruition.

Effective communication: Effective communication is a linchpin in algorithm development. During my stint working on decision-making algorithms, it was interesting to observe how individuals subconsciously infused their driving styles into algorithm design. In one instance, a debate ensued on whether our truck should yield or proceed when facing a high-speed vehicle in the opposite lane at an intersection. Diverging opinions emerged, underscoring the importance of articulate communication in understanding varying perspectives and arriving at a consensus or an even more innovative solution.

Continuous learning: The self-driving landscape is continually evolving, and staying abreast of the latest algorithms is imperative. The advent of Artificial Intelligence, especially Large Language Models (LLMs), has marked a significant milestone. Leveraging the prowess of these advanced models is becoming increasingly crucial in honing decision-making algorithms for autonomous vehicles, which underscores the importance of continuous learning and skill enhancement in this field.

Solid technical judgment: Sound technical judgment is indispensable when designing or troubleshooting algorithms. When tasked with incorporating a new function in an algorithm, it’s vital to weigh short-term against long-term solutions, and sometimes, sift through multiple implementation methods. A well-honed technical acumen helps in discerning the most effective approach, balancing timelines with the pros and cons of each method.

Optimism and resilience: The autonomous driving sector, laden with higher risks and uncertainties compared to other automotive domains, demands a resilient spirit. The road to commercialization may be long and winding, and the impact of one’s efforts might not immediately reflect in stock prices or investment influx. Hence, maintaining a steadfast belief in the transformative potential of this technology and nurturing a positive outlook toward the eventual realization of autonomous driving on a grand scale is essential.

You are a person of great influence. If you could inspire a movement that would bring the most amount of good to the most amount of people, what would that be? You never know what your idea can trigger. :-)

Haha, I’m flattered. There are a few things I want to inspire including promoting sustainable transportation, open-sourced development for the autonomous driving platform, and defining safety standards. However, the movement I want to inspire most is to build an understanding of autonomous driving technologies, given we are currently at the key time frame when self-driving technology is scaling up.

As I mentioned previously, some people don’t understand self-driving technology and are skeptical about its future. I’d love to kickstart a movement where we could have open conversations with the community. I hope I can get the chance to hold casual meetups where people can chat about self-driving cars. No jargon, no pressure. Just real talk about how this tech can shape our streets and neighborhoods. And of course, giving people a chance to voice their thoughts and concerns. It’s all about making tech feel less like a stranger and more like a friendly neighbor. It could be a fun way to make self-driving tech less intimidating and more exciting. By doing so, I would say it will be most beneficial for the development of self-driving and let most people accept and love self-driving technology.

How can our readers further follow your work online?

If readers are interested in my work, I invite them to stay updated through TuSimple’s official website and by subscribing to the TuSimple YouTube channel. We periodically share our progress on these platforms. Additionally, I welcome connections on LinkedIn where I’m always open to discussing any questions or concerns regarding self-driving technology. You can also delve into my published work on my Google Scholar page.

Thank you so much for the time you spent doing this interview. This was very inspirational, and we wish you continued success.

About The Interviewer: David Leichner is a veteran of the Israeli high-tech industry with significant experience in the areas of cyber and security, enterprise software and communications. At Cybellum, a leading provider of Product Security Lifecycle Management, David is responsible for creating and executing the marketing strategy and managing the global marketing team that forms the foundation for Cybellum’s product and market penetration. Prior to Cybellum, David was CMO at SQream and VP Sales and Marketing at endpoint protection vendor, Cynet. David is a member of the Board of Trustees of the Jerusalem Technology College. He holds a BA in Information Systems Management and an MBA in International Business from the City University of New York.