The Next Level Autonomous World

By Hong Kong Economic Times

The world stands at a tipping point for the rapid adoption of autonomous technologies. Driven by advancements in 5G networks, the Internet of Things (loT) and artificial intelligence (AI), autonomy has evolved from the realm of science fiction into practical reality that can be seen in daily life, with automobiles being a typical representative of this.

In a report by Goldman Sachs, partially autonomous cars, or Level 3 vehicles, will account for up to 10 percent of global new car sales by 2030, while fully autonomous cars, or Level 4 cars, could represent around 2.5 percent of total sales in the same period. The report further suggests that China will see the highest adoption rate of autonomous vehicles (Level 3 or above), reaching as much as 90 percent by 2040.¹

As autonomous technologies reshape transport, industry and defense, engineers are at the forefront of this revolution. Yes, the critical question remains: when machines begin to “think”, how should engineers redefine their roles?

Autonomous driving, a transformative smart technology for transportation, has shown vast potential across logistics, military operations, and public transit.

Current state of autonomous technology

Autonomous technology refers to systems capable of executing tasks and making decisions independently, without direct human intervention.

Unlike traditional automated systems that operate within preprogrammed instructions in pre-set scenarios, autonomous systems dynamically adapt to their surroundings using sensors, artificial intelligence, and machine learning. Advances in autonomous systems have been transforming transport, manufacturing and logistics through smart factory robots, selfdriving cars, drone fleets, and driverless trains and unmanned cargo ships.

In 2014, the Society of Automotive Engineers (SAE) introduced a framework to classify level of driving autonomous vehicles, from Level 0 (no automation) to Level 5 (full automation), in the context of motor vehicles and their operation on the roadways. This framework remains the industry’s benchmark for driving automation.

Commercialisation of autonomous driving has accelerated in recent years, with substantial progress being made across multiple application scenarios. Most automakers now offer Level 2 systems. Mercedes-Benz, with its Drive Pilot ADAS³, is currently the only automaker delivering Level 3 vehicles. While fully autonomous L5 systems remain distant due to technological, ethical and regulatory challenges, in the foreseeable future, L4 solution is viewed as the most practical near-term reality.

Current autonomous technologies demonstrate varying capabilities across different levels. Notable progress can be seen in low-speed autonomous solutions in controlled environments such as industrial parks, mining operations and port logistics, where their operational value has been clearly demonstrated.

Though autonomous systems have demonstrated impressive capabilities, such as navigating complex environments, avoiding obstacles, and making real-time decisions, it also has many limitations. For example, sensor accuracy can vary under different weather conditions. Processing units must balance performance and energy efficiency. Decisionmaking algorithms are not always reliable and can struggle with extreme cases and unpredictable human behaviour.

Globally, competition in autonomous vehicles is fierce. While the United States and Germany continue to lead in technology, Asian nations are making significant strides. Japan and South Korea are rapidly emerging as key players in autonomous mobility innovation. China is strategically positioning itself to achieve groundbreaking progress in selfdriving technology.

Core enablers of next-level autonomy

Internet of Things (IoT)

The Internet of Things (IoT) refers to a network of sensorequipped physical devices that continuously transmit realtime data to centralised processing systems for interpretation, analytics, and automated decision-making. It plays a transformative role in advancing autonomous vehicle technology by enabling real-time data exchange, enhancing situational awareness, and creating smarter transportation ecosystems.

By enabling Vehicle-to-Everything (V2X) communication, the IoT helps autonomous vehicles interact with traffic infrastructure, other vehicles, and pedestrians. Vehicle-to- Vehicle (V2V) systems share speed and position data to prevent collisions, while Vehicle-to-Infrastructure (V2I) connects autonomous vehicles (Avs) to smart traffic lights and road sensors. This integration reduces reliance on onboard sensors, providing 360° situational awareness even in blind spots.

IoT devices collect information on traffic patterns, weather conditions, and road obstacles, which is shared with other vehicles and infrastructure to optimise navigation and decisionmaking. It provides real-time monitoring through sensors in vehicles and infrastructure, enabling proactive adjustment to driving behaviour. The IoT also enhances perception as it can integrate data from various sources to provide a more comprehensive understanding of the driving environment, while supporting advanced driver-assistance systems through functions such as automated parking and platooning.⁶

5G Networks

5G networks are the backbone of autonomous driving. providing low latency, high-speed data transmission, and improved reliability—essential for the effective operation of autonomous vehicles. The capacity of 5G to deliver multi-gigabit-per-second data speeds enables vehicles to process vast amounts of information in real time, allowing immediate responses to changing road conditions and surrounding environments. This capability is especially crucial for applications such as real-time traffic updates, hazard detection, and vehicle-to-vehicle communications.

Moreover, the low latency feature of 5G networks—often as low as one millisecond—enables the rapid exchange of information between vehicles and infrastructure. This fast communication is vital for critical applications such as collision avoidance systems, where decisions must be made in fractions of a second. As autonomous vehicles rely on sensor data from various sources, including cameras, LiDAR, and radar, the enhanced connectivity afforded by 5G facilitates seamless integration of this sensor data with cloud computing resources. This allows for advanced processing capabilities, such as complex algorithm execution for path planning and environmental mapping, which are necessary for higher levels of autonomy.

5G also supports the Internet of Vehicles, which enables vehicles to connect with other vehicles, infrastructure, and cloud services. This interconnectedness allows for collective data sharing and analysis, improving overall traffic management and enhancing the efficiency of transportation systems. In addition, the integration of C-V2X (Cellular Vehicleto- Everything) technology within 5G networks creates a robust framework for communication that can adapt to various scenarios, ensuring that vehicles remain responsive and informed.⁷

With 5G technology, cars can be remotely controlled by external operators acting as a traffic controller. Such remote operation is only feasible with a 5G network, which provides essential features such as extremely low latency and guaranteed network resources.

Generative AI (GenAI)

There are three key roles of Gen AI in autonomous technology, namely developing end-to-end AI models that improve vehicle decision-making, generating synthetic datasets to train AI systems for diverse driving situations, and improving humanmachine collaboration through advanced monitoring systems.

Raquel Urtasun, Founder and CEO of self-driving technology company Waabi, states: “There is a massive leap in AI happening right now, resulting in smarter end-to-end AI systems that can learn much more efficiently, are interpretable, and can generalise to every possible scenario on the road.”

These advances result in autonomous vehicles with superhuman capabilities that will enhance road safety and transform transportation, she said.⁸

By adopting AI to offer a human-like driving experience, transformer neural networks, a key breakthrough in autonomous driving, can perform “birds-eye view” perception to achieve accurate environmental perception around the vehicles. They utilise an attention mechanism to better comprehend input sequences, allowing for a global contextual interpretation of the driving environment. This results in improved modelling of the vehicle's surroundings in three-dimensional space, aiding path planning for autonomous vehicles. Unlike traditional convolutional neural networks, which may produce conflicting data, transformers provide coherent and precise occupancy modelling, thus facilitating safer and more reliable autonomous driving experiences.⁹

Level 3 autonomous vehicles

Realising L3 level autonomous driving (conditional driving) requires meeting a complex set of technical requirements. While the market is still dominated by L2 cars, many attempts have been made to achieve LEVEL 3 autonomy.

In 2021, Mercedes-Benz became the world’s first vehicle manufacturer to receive a globally valid system approval for conditionally automated driving in accordance with SAE Level 3.

The system permits temporary disengagement from driving tasks, allowing users to remove both visual attention and manual input from the steering apparatus until an intervention request is issued. It is already on the market in Germany and the United States. The company is also among the first to receive approval to test LEVEL 3 systems in Beijing.¹⁰

According to Mercedes-Benz, a DRIVE PILOT-enabled vehicle can take over the dynamic driving task up at speeds of up to 64 km/h (and up to 95 km/h on German motorways) on suitable freeway sections and during high traffic density. Sensors of DRIVE PILOT include LiDAR laser beams for 3D environment, Radar electromagnetic waves for distance and speed measurement, a road moisture sensor to detect road surface wetness and ultrasonic sensor to use sonic impulses to detect the near vehicle surroundings. It also comes with redundant steering and braking actuators and a redundant power system to ensure manoeuvrability in the event of a system failure.

Besides Mercedes-Benz, Japanese carmaker Honda Motor was the world’s first carmaker to sell a vehicle with LEVEL 3 selfdriving technology approved by the Japanese Ministry of Land, Infrastructure, Transport and Tourism. However, production was limited to 100 units due to high cost.

In China, a number of automakers including BYD have been allowed to carry out tests on LEVEL 3 vehicles on public roads. Zeekr Group, Xpeng and Guangzhou Automobile Group said in March that they will start selling electric vehicles with LEVEL 3-ready autonomous driving capabilities.¹¹

It is also expected that leading advanced driver-assistance system players will introduce 71 vehicles equipped with Level 3 autonomous driving technology in Europe in 2025. This surge will result in Level 3 penetration reaching 21.2 percent of new car sales in Europe by the end of the year.¹²

Since LEVEL 3 autonomous driving requires the vehicle to fully take driving control under specific conditions, with the driver intervening only upon system request, this poses core technical challenges in the collaborative operation of multi-sensor systems, decision-making algorithms that mimic human driving strategies for complex traffic interactions, and redundancy and safety to provide support in the event of system failures.

The development of LEVEL 3 autonomous technology also faces challenges in complex scenarios such as extreme weather conditions. The perception accuracy may decrease when it is raining. While water droplets and fog particles reduce image clarity.

Asian countries are ramping up efforts to test autonomous driving in real-life scenarios

Market potential across sectors

Autonomous technology is poised to impact various sectors with its disruptive potential and wide range of application scenarios. Whether in mobility, aerospace and defense, or mining, construction, and industry, this technology shows great market potential.

With the continuous advancement and expanding applications of drone technology, unmanned aircraft and drones are poised to become integral components of future aerial transportation systems. Autonomous driving technology will enable these aerial vehicles to independently execute take-off, cruising, and landing procedures, significantly enhancing flight efficiency and safety.

The global AI and robotics market in aerospace and defense was valued at US$32.5 billion in 2024 and is estimated to grow at a compound annual growth rate (CAGR) of 7.7% between 2025 and 2034.¹³ Autonomous systems, such as drones and unmanned combat vehicles, are increasingly utilised for surveillance, reconnaissance, and logistics, improving operational efficiency and reducing risks to human personnel in dangerous missions. AI helps predict when planes need repairs, improves flight efficiency, and makes air travel safer. New technologies such as coordinated groups of drones, could revolutionise military operations. As both military and civilian aviation seek more cheaper, and advanced solutions, autonomous technology is expected to grow rapidly with a transformative impact.

In the mining sector, the autonomous mining equipment market size is projected to reach US$3.14 billion in 2025, up from US$2.94 billion in 2024. It is expected to grow further to US$3.88 billion by 2029 at a CAGR of 5.5%. A key driver of this growth is the increasing concern for miners’ safety, as autonomous equipment such as robotic loaders, laser sensors, and driverless truck can help ensure workers’ safety. Autonomous systems can also improve mining efficiency; for example, the autonomous water truck launched by Caterpillar allows mine operators to digitally monitor water consumption and reduce waste.

In agricultural fields, autonomous farming machinery can perform precision operations such as seeding, fertilising, and harvesting. Looking ahead, as technology continues to progress and applications diversify, autonomous vehicles will gradually become embedded in daily life, delivering safer, more efficient, and convenient mobility solutions.

Case study: Hong Kong's autonomous vehicle implementation

To facilitate wider trials and industry adoption of autonomous vehicles in Hong Kong, new legislation and a regulatory regime for autonomous vehicles came into force on 1 March 2024, providing a flexible regulatory framework for AVs. In November last year, the Transport Department issued the first pilot licence for autonomous vehicle under section 4(1) of the Road Traffic (Autonomous Vehicles) Regulations (Cap.374AA). The application of the pilot licence was approved to conduct trial for 10 autonomous vehicles in North Lantau.

After setting out the vision of materialising autonomous vehicles trial on public roads in the Smart Mobility Roadmap for Hong Kong in 2019, the government established a US$1 billion Smart Traffic Fund to support local organisations and enterprises in conducting research and applying vehicle-related innovation and technology. As of November 2024, 12 approved projects have focused on autonomous vehicles. These trials cover various road sections, including public roads and diverse application scenarios such as the West Kowloon Cultural District (Navya Arma), the Hong Kong Science Park (MobiGem and CM Pro), Zero Carbon Building (MobiToTo), Tai Po Industrial Estate (autonomous tractor) and individual private residential estates.¹⁴

In the 2024 Policy Address, Chief Executive John Lee Kachiu announced the work direction for promoting the development of low-altitude economy (economic activities in airspace below 1,000 metres), including the establishment of the Working Group on Developing LAE led by the Deputy Financial Secretary to formulate development strategies and action plans. The government is planning to take forward the regulatory sandbox pilot projects progressively starting from this year to explore more application scenarios for low-altitude flying activities.¹⁵

In June 2024, the Airport Authority Hong Kong announced it was awarded the tender for the development of autonomous vehicles and the associated transportation system for carrying visitors between the Hong Kong Port of the Hong Kong-Zhuhai- Macao Bridge and SKYCITY at Hong Kong International Airport, the first autonomous mass transportation system in Hong Kong.

Passengers traveling on autonomous vehicles from the Hong Kong Port will arrive at SKYCITY in just three minutes, and vice versa. Each vehicle can accommodate up to 16 passengers. Initially, the system will transport 500 passengers per hour in each direction, with plans to expand capacity to 2,000 passengers per hour per direction, according to the Airport Authority Hong Kong.¹⁶

The autonomous technology has also become available in daily life in Hong Kong. In August 2024, two autonomous driving shuttle buses has been put into use for residents to rise in the 2.5-kilometer-long inner loop of Fairview Park in Yeun Long. The vehicle has 15 sensors, including video camera and optical radar. The sensors, along with high-definition electronic maps and image processing technology, allows the AI system to analyse road conditions and achieve precise positioning. 5G technology is also used to monitor the vehicle’s condition in real time through smart lampposts. The speed of the shuttle bus is limited to 30 km/h while a human driver is assigned on the vehicle to control manually in case of emergencies or accidents.¹⁷

Beyond the road

The development of autonomous systems is rapidly expanding beyond traditional road vehicles, with breakthroughs in flying cars, urban air mobility, and military applications.

In March, China’s Ehang, an autonomous aerial vehicle manufacturer, announced that it became the world’s first to obtain a certificate for operating a pilotless aerial vehicle capable of carrying human passengers for its EH216-S twoseater aerial vehicle. The vehicle has a maximum design speed of 130 km/h, and a maximum range of 30 kilometres. It uses 16 rotors for redundancy to ensure stable flight.

“This milestone officially marks the launch of China’s humancarrying flight era in the low-altitude economy, allowing citizens and consumers to purchase flight tickets for low-altitude tourism, urban sightseeing, and diverse commercial humancarrying flight services at related operation sites in Guangzhou and Hefei,” the company said.¹⁸

However, some autonomous initiatives face hurdles. France initially planned to launch a flying taxi service for attendees of the 2024 Paris Olympics, but the plan was abandoned due to significant public backlash and scepticism regarding its feasibility and benefits. Additionally, the technology struggled to gain regulatory approval and failed to convince officials about its safety.

Meanwhile, militaries worldwide are testing and deploying new AI-enabled autonomous vehicles which could change the battlefield. For example, in Ukraine, autonomous technology is being utilised primarily through AI-equipped drones capable of carrying out targeted strikes. These drones have been deployed to conduct missions against Russian oil refineries and other strategic targets. The technology allows for real-time data gathering and analysis, enhancing operational effectiveness.

Autonomous systems are also being integrated into military strategies for tasks such as surveillance and reconnaissance. The use of AI enables these systems to make more decisions independently, which can improve responsiveness in combat situations. This increasing reliance on autonomous technology reflects a broader trend in modern warfare, with both sides exploring the potential of AI to enhance their military capabilities.

Under Level 3 autonomy, drivers can safely look away from the road and release the wheel until alerted by the vehicle to retake control

Engineering challenges and opportunities

Autonomous driving technology has brought convenience and innovation to modern transportation, yet its widespread adoption hinges on balancing innovation with safety, public trust, and regulatory frameworks.

This gives engineers the responsibility to ensure these technologies serve humanity’s interests and to promote the advancement of policies that establish standardised regulations and V2X communication protocols for global deployment.

Autonomous technology is the trend of future development, playing a key role in easing traffic congestion, reducing environmental pollution, and facilitating long-distance delivery. The use of advanced technologies such as LiDAR, radar and high-precision mapping systems will enable a high level of autonomy for smart transport. The advancement of autonomous technology presents both challenges and opportunities for engineers, including those working in autonomous systems safety and edge AI deployment. Engineers who embrace the evolving demands in autonomous technology will shape the future of mobility and automation.

References

1. Partially Autonomous cars forecast to comprise 10% of new vehicle sales by 2030 (2024) Goldman Sachs. Available at: https://www.goldmansachs.com/insights/articles/partially-autonomous-carsforecast-to-comprise-10-percent-of-new-vehicle-sales-by-2030 (Accessed 2 Apr. 2025).
2. Partially Autonomous cars forecast to comprise 10% of new vehicle sales by 2030 (2024) Goldman Sachs. Available at: https://www.goldmansachs.com/insights/articles/partially-autonomous-carsforecast-to-comprise-10-percent-of-new-vehicle-sales-by-2030 (Accessed 2 Apr. 2025).
3. Doll, S. (2024). CEO says Ford is ‘getting close’ to Level 3 autonomous driving that enables ‘hands and eyes off’. [online] Electrek. Available at: https://electrek.co/2024/06/20/ceo-ford-getting-close-level-3-autonomous-driving-hands-eyes-off/ (Accessed 2 Apr. 2025).
4. McCorry, P. (2022). South Korea Accelerates Commercialization of Autonomous Vehicles. [online] LeddarTech. Available at: https://leddartech.com/white-paper-south-korea-acceleratescommercialization-of-autonomous-vehicles/ (Accessed 2 Apr. 2025).
5. Ruan Zheng (2024). Empowering smart vehicles to navigate intelligent roads. [online] Cctv.com. Available at: https://5gai.cctv.com/2024/01/02/ARTIjN7JcLY1lQe8ekhkXu8H240102.shtml (Accessed 2 Apr. 2025).
6. Vermesan, O., Bahr, R., Falcitelli, M., Brevi, D., Bosi, I., Dekusar, A., Velizhev, A., Alaya, B., Firmani, C., Simeon, J.-F., Tcheumadjeu, L., Solmaz, G., Bisconti, F., Mauro, L., Noto, S., Pagano, P., Ferrera, E., Alejandro, G., Castillo, G. and Bonetto, E. (2020). IoT Technologies for Connected and Automated Driving Applications. [online] River Publishers. Available at: https://www.riverpublishers.com/pdf/ebook/chapter/RP_9788770221955C6.pdf (Accessed 16 Apr. 2025).
7. Titus, Nzeyimana & Extension, Kiu Publication. (2024). The Impact of 5G Technology on Autonomous Vehicles. 3. 36-39.
8. Alonso, M., Koster, A., Jordan, P. and World Economic Forum (2025). How GenAI is driving the development of vehicle autonomy. [online] World Economic Forum. Available at: https://www.weforum.org/stories/2025/04/how-genai-is-helping-drive-vehicle-autonomy/ (Accessed 16 Apr. 2025).
9. Louis, L. (2024). Accelerating Transformer Neural Networks for Autonomous Driving | Ambarella. [online] Ambarella. Available at: https://www.ambarella.com/blog/accelerating-transformer-neuralnetworks-for-autonomous-driving/ (Accessed 16 Apr. 2025).
10. Group, M.-B. (2021). First internationally valid system approval for conditionally automated driving. [online] Mercedes-Benz Group. Available at: https://group.mercedes-benz.com/innovation/productinnovation/autonomous-driving/system-approval-for-conditionallyautomated- driving.html (Accessed 16 Apr. 2025).
11. Reuters Staff (2025). China’s Zeekr, Xpeng to offer cars with L3 autonomy as auto battle heats up. Reuters. [online] 19 Mar. Available at: https://www.reuters.com/business/autos-transportation/chinaszeekr-unveils-first-ev-with-level-3-ready-autonomous-drivingcapabilities-2025-03-18/ (Accessed 16 Apr. 2025).
12. Stathousis, G. (2021). Level 3 autonomous driving in Europe - AUTO2X. [online] AUTO2X. Available at: https://auto2xtech.com/level-3-automation-in-1-in-5-cars-in-europe-by-2025/?srsltid=AfmBOoqBs783fAkVUGcX7sRnoMeGfguGnhlP302FvhpBQAp1aNCoTRQL (Accessed 16 Apr. 2025).
13. Paul, P. and Suraj Gujar (2024). AI and Robotics in Aerospace and Defense Market – By Component, By Deployment, By End Use & Forecast, 2025 – 2034. [online] Global Market Insights Inc. Available at: https://www.gminsights.com/industry-analysis/ai-and-robotics-inaerospace-and-defense-market (Accessed 16 Apr. 2025).
14. Info.gov.hk. (2024). Transport Department issues pilot licence for autonomous vehicle. [online] Available at: https://www.info.gov.hk/gia/general/202411/29/P2024112900386.htm (Accessed 17 Apr. 2025).
15. Info.gov.hk. (2024a). LCQ18: Promoting the development of lowaltitude economy. [online] Available at: https://www.info.gov.hk/gia/general/202411/06/P2024110600282.htm (Accessed 16 Apr. 2025).
16. Hong Kong International Airport. (2024). Airport Authority Awards Tender for Hong Kong’s First Autonomous Mass Transportation System. [online] Available at: https://www.hongkongairport.com/en/mediacentre/press-release/2024/pr_1728 (Accessed 16 Apr. 2025).
17. 星島環球網（2024）。 自動駕駛小巴元朗錦繡花園試載客，限速30公里有 安全帶感應器 | 星島環球網. [online] Available at: https://www.stnn.cc/c/2024-08-22/3915356.shtml (Accessed 16 Apr. 2025).
18. Ehang.com. (2025). EHang | EHang’s EH216-S eVTOL Operators Obtain Air Operator Certificates. [online] Available at: https://www.ehang.com/news/1196.html (Accessed 16 Apr. 2025).