How does game technology drive the automotive industry?

On November 28, Musk replied to a post on X, mentioning that xAI would open an AI game studio, hoping to make games great again!

Source: X@Elon Musk

At first glance, some might complain that Musk is neglecting his main duties again? But based on past experience, Musk's love for games may hide deeper considerations. Nowadays, game technology is no longer limited to games themselves but has been applied to various industries, with the automotive industry being intricately connected to game technology.

Of course, it's possible that Musk didn't think that much and it's just overthinking on the part of electric vehicle enthusiasts. Regardless, let's take this opportunity to discuss how game technology shines in the automotive industry.

Games Are the 'Behind-the-Scenes Pushers' of Autonomous Driving

Those who play games should know the famous GTA, also known as 'Grand Theft Auto' which we played as children. While we create various 'car accidents' in GTA5, someone has already used this game to develop cutting-edge technology.

A group of gamers from Intel Labs and Darmstadt University in Germany discovered that GTA5 is a highly realistic Los Angeles city simulator with real roads, vehicles, pedestrians, traffic signals, and other transportation facilities. It can also simulate special scenarios that are difficult to occur in real life, such as being hit by a train, falling off a cliff, or jumping over peaks.

Source: GTA5 game screenshot

Therefore, they invented a clever method to extract data from the GTA game. Researchers created a new 'software layer' between the game and computer hardware, which automatically classifies different targets on the game's roads. This data can be used in machine learning algorithms to identify cars, pedestrians, or other targets.

This directly made some automakers and tech companies uneasy. The limited time and space in the real world have always been a headache for autonomous driving companies. According to RDI data: For autonomous driving algorithms to reach the level of human drivers, at least 17.7 billion kilometers of driving data need to be accumulated. This means that even if a team of 100 test vehicles drives at a speed of 40 km/h for 24 hours non-stop, it would take about 500 years.

The realistic simulation of the real world in GTA5 has inspired automakers: Since the boundary between games and reality is becoming increasingly blurred, can the autonomous driving environment be simulated in the game world to collect data and optimize algorithms?

Theoretically, it is indeed possible. Collecting data through game simulations can not only complete billions of kilometers of test mileage at a relatively low cost and high efficiency but also improve R&D test efficiency through scene diversity and density, further shortening the test mileage goal.

In a BYD job posting, the requirements for the autonomous driving image engine development position clearly mention that applicants need to be familiar with the Unreal Engine and related functional modules.

Source: Boss Zhipin

The Unreal Engine is a tool developed by Epic, containing various tools and resources needed to create games or movies. It can handle high-quality 3D graphics and physical simulations, so it is often used to produce games and movies with high visual requirements. For example, 'Black Myth: Wukong' was developed using the Unreal Engine.

On the Unreal Engine's official website, a special 'Automotive' section elaborates on how Unreal Engine technology is applied in the automotive industry. For example, how to utilize data formats from the animation film industry to drive the development of autonomous vehicles, robots, and drones, and how technology can help Ferrari navigate the metaverse.

Source: Unreal Engine official website

Using the Unreal Engine, many companies have embarked on the journey of building virtual simulation platforms based on game engines, setting the stage for an intense competition around simulation platforms.

Giants Compete for Autonomous Driving Simulation Platforms

Today, the autonomous driving simulation platform has initially formed a complete industrial chain system.

In 2016, Intel Labs tested autonomous driving in GTA, and companies like Google, NVIDIA, and Microsoft also developed autonomous driving simulation systems based on game engines.

A simulation system is, in simple terms, a virtual world replicated from the real world, also known as a 'digital twin'. It replicates everything from geographical features, buildings, and road layouts to physical rules like gravity, day-night cycles, and weather changes.

Source: Unreal Engine official website

In this world, cars are deployed like NPCs to complete safety tests for autonomous driving functions. Testers can operate the system to change various parameters at any time, simulating different types of road environments, such as increasing the number of pedestrians, setting obstacles, and simulating sudden pedestrian crossings or forced lane changes to test the adaptability of autonomous driving functions.

China is also actively deploying simulation platforms. In early 2020, the 'Smart Car Innovation and Development Strategy' issued by eleven ministries and commissions clearly mentioned the need to focus on supporting the development of virtual simulation, hardware-software integrated simulation, real-world road testing, and other technologies and verification tools.

In 2018, Tencent developed the virtual simulation platform TAD Sim, combining cloud computing capabilities to build a cloud-based virtual city simulation for testing and verification of complex environments. Sun Chitian, technical director of TAD Sim's simulation technology, said: The autonomous driving simulation test platform born from the game engine can quickly achieve a basic setup from 0 to 70 points using game technology.

Meanwhile, about one-third of the TAD Sim team has a background in game development, and when recruiting new members, they specifically mention 'game development experience is preferred' in the job descriptions.

Source: Tencent autonomous driving virtual simulation test scenario

In 2020, Huawei and Alibaba also officially entered the simulation platform market. At the 2019 Shanghai Auto Show in April, Huawei's autonomous driving cloud service Octopus was first exhibited, with simulation testing being one of its service capabilities. On January 9, 2020, Huawei's autonomous driving cloud service was first launched in Changsha Xiangjiang New Area.

Alibaba, on the other hand, released the world's first 'hybrid simulation test platform' for autonomous driving on April 22, 2020. This platform uses simulation technology that combines virtuality and reality, introducing real road test scenarios and cloud trainers. It can simulate an extreme scenario in just 30 seconds, and the system can virtually test over 8 million kilometers per day.

In addition to these tech giants, autonomous driving companies like WeRide and vertical enterprises like 51WORLD have also launched their own autonomous driving simulation test systems.

51WORLD has even proposed the 'Earth Cloning Plan', a grand plan to replicate Earth in the virtual world and create an ultra-large simulation simulator.

Source: 51World Earth Cloning Plan

It is foreseeable that through continuous iterations by these tech companies, future simulation platforms are likely to become universal next-generation engines, empowering various industries including automobiles and aerospace. At the same time, more possibilities will emerge in the automotive industry.

The Fusion of Virtuality and Reality: Game Technology Has Become the Cornerstone of the Automotive Industry

In automotive R&D centers, the Unreal Engine has long been involved in vehicle design and development. Hao Yongjun, Senior Manager of Visual Realization at Li Auto, said they used the Unreal Engine to design an entire car in just one afternoon, allowing for a comprehensive view of every aspect of the car.

It's not just automakers that are interested in game technology. As game technology becomes more realistic in recent years, many professional race car drivers have also turned their attention to racing games, hoping to enhance their daily training through them.

These games not only replicate the layout of real-world racetracks and the performance of race cars but also simulate tire wear and fuel consumption that are quite realistic. At the same time, they mimic the racing formats and various rules of real-world racing. This means that aside from not feeling the cheers of the live audience, there is little difference from real-world racing.

Source: Forza Horizon 5 game screenshot

In fact, on relatively popular racing games like Forza Horizon 5, using a steering wheel for control can already provide a certain level of immersion, let alone professional racing games customized for race car drivers. GT Academy has successfully trained several endurance race and GT series drivers by selecting the best gamers, some of whom are still active in racing today.

The prototype of what we used to call 'intelligent transportation' was actually first realized conceptually through games and then applied to real life.

For example, in the game 'Cities: Skylines', players can freely plan urban roads, including designing main roads, secondary roads, branch roads, and other types of roads, setting one-way streets, traffic lights at intersections, and traffic signs.

Source: Cities: Skylines game screenshot

By rationally planning road layouts, such as increasing the number of branch roads to reduce traffic pressure on main roads and setting reasonable traffic signal timings to guide vehicle flow, players can effectively alleviate traffic congestion in the game's cities and improve traffic efficiency. This also provides a useful reference for urban traffic planning and management in reality.

At the same time, using game simulations can also alleviate traffic congestion to a certain extent. For example, when dealing with large-scale traffic congestion at specific locations after a concert, traffic management departments need to make predictions and responses in advance based on actual conditions. At this time, using game technology for simulation is a necessary and efficient step.

Once, games were considered 'electronic poisons' and were controversial. However, times have changed, and today's game technology has transformed into a key force driving the development of the automotive industry and even the entire technology sector.

Final Thoughts

From the accidental discovery in GTA5 to major companies building autonomous driving simulation platforms based on game engines, to games like Cities: Skylines providing conceptual prototypes and practical references for intelligent transportation, the presence of game technology is everywhere. It has broken the boundary between virtuality and reality in ways beyond imagination, playing an irreplaceable role in multiple key links within the automotive industry.

However, this is just the beginning. Will game technology completely reshape our travel modes and transportation systems? Will future cars become a new carrier that integrates gaming entertainment with efficient travel? We are standing at the crossroads of technological convergence, and with the continuous development of game technology, more innovations and breakthroughs await us in the future.

Cover Image Source: Unreal Engine official website

Source: Lei Technology