Have Autonomous Vehicles become a reality yet?

As of 2025, while the skies aren’t filled with flying cars as once imagined decades ago, intelligent mobility has made significant strides. What was once a futuristic concept is now a tangible reality. Autonomous vehicles (AVs) are already operating among us, guided by artificial intelligence that interprets the environment, makes decisions, and responds instantly without human intervention. 

AVs, or robotaxis, are increasingly prevalent in certain countries. In regions like the United States and China, their deployment is becoming more common, especially in complex urban settings. For instance, in February 2025, Alphabet announced that its autonomous vehicle subsidiary, Waymo, achieved 200,000 paid rides per week in cities such as Los Angeles and San Francisco, a twentyfold increase compared to two years prior. 

While global adoption isn’t yet widespread, other nations like Japan and South Korea have initiated pilot programs in major cities, testing the technology in controlled environments. Simultaneously, the European Union is progressively developing regulatory frameworks to facilitate the gradual and safe integration of these systems. 

Thus, the question is no longer if we will see autonomous vehicles at scale in our cities, but when and under what regulations. 

What Is the Current Status? 

The Society of Automotive Engineers (SAE) defines six levels of automation, ranging from Level 0 (fully manual vehicles) to Level 5 (full autonomy vehicles capable of operating in any situation without human intervention). 

Currently, most commercial vehicles fall between Levels 2 and 3, offering driver assistance features but still requiring human oversight. Only a few companies have achieved Level 4 operations, which entail autonomous driving in specific environments and under certain conditions. 

Achieving Level 5 remains a medium- to long-term goal, primarily constrained by technical, ethical, and regulatory challenges. 

How Do Autonomous Vehicles Operate? 

The functionality of an autonomous vehicle relies on the integration of multiple technological layers working in unison to ensure precise and safe driving in dynamic environments. 

Firstly, environmental perception is achieved through a combination of advanced sensors, including high-resolution cameras, LiDAR, radar, and ultrasonic sensors. These sensors provide a three-dimensional, accurate representation of the surroundings, enabling the vehicle to detect obstacles, identify traffic signals, recognize pedestrians, and anticipate the behavior of other road users. Each sensor serves a specific purpose: radar performs well in adverse weather conditions, LiDAR is crucial for precise long-distance 3D visualization, and cameras facilitate the recognition of complex elements. 

All this information is processed by AI algorithms based on deep learning models, such as convolutional neural networks (CNNs) for computer vision and attention networks to understand environmental dynamics. Trained with millions of kilometers of real and simulated data, these systems enhance their predictive and adaptive capabilities through techniques like deep reinforcement learning, optimizing decision-making in complex situations. 

A key challenge is sensor fusion, which integrates data with varying frequencies and noise levels to generate a coherent three-dimensional model of the environment. This representation feeds into planning and control modules, allowing the vehicle to anticipate risks and respond accurately. Low latency and robustness against uncertainty are essential for reliable autonomous driving in real-world, changing scenarios. 

Additionally, Vehicle-to-Everything (V2X) connectivity plays a crucial role by extending the vehicle’s perception beyond its own sensors. It enables real-time communication with urban infrastructure, other vehicles, and even pedestrians’ mobile devices, creating an interconnected network that facilitates route optimization, movement coordination, and risk anticipation. For example, an autonomous vehicle can receive information from traffic lights to adjust its speed before reaching an intersection or receive alerts about other vehicles approaching at high speed. 

To manage this vast amount of real-time data, autonomous vehicles depend on high-performance hardware. Platforms like NVIDIA DRIVE and Qualcomm Snapdragon Ride, along with specialized chips developed by various tech companies, provide the necessary processing power to analyze and act upon millions of data points per second. These systems are essential to ensure that autonomous vehicles can react instantly to any situation, from a change in another vehicle’s speed to the need to brake for an unexpected obstacle. 

Collectively, these technologies enable AVs to operate safely and efficiently in highly complex and ever-changing environments. 

Remaining Challenges 

Despite significant technological advancements, the global adoption of AVs faces substantial obstacles: 

Regulation: A primary hurdle is the lack of an adapted legal framework. In most countries, legislation doesn’t permit the circulation of driverless vehicles, and the absence of international harmonization complicates their approval and the procurement of appropriate insurance. 

Ethics: Algorithms must be prepared to react to risk situations where any decision entails potential harm. This raises debates about how to prioritize lives and who is responsible for these decisions—the programmer, the manufacturer, or the user. Moreover, these ethical decisions can vary depending on cultural context. For instance, in Europe, an autonomous vehicle might prioritize avoiding harm to a child over an elderly person, whereas in Japan, due to cultural reverence for the elderly, the decision might be the opposite. 

Cybersecurity: Being connected to networks and digital systems makes AVs vulnerable to cyberattacks. A security breach can affect the vehicle’s functionality and, in extreme cases, be exploited to create hazardous situations on public roads. This vulnerability represents one of the most significant challenges for the safe adoption of this technology, necessitating the development of robust defense mechanisms and constant updates to counter new threats. 

Infrastructure: Most cities lack infrastructure compatible with the widespread circulation of autonomous vehicles. The transition will require significant investments in urban technology, such as intelligent signage, advanced communication networks, and integrated traffic management systems. 

The European Landscape 

As mentioned earlier, some countries have already launched operational autonomous vehicle services. In Europe, progress is more gradual, characterized by a stringent regulatory approach and a strong commitment to safety and technological interoperability. 

Countries like Germany and the United Kingdom are leading the transition on the continent. Germany was a pioneer, approving a law in 2021 that allows Level 4 autonomous vehicles to operate on public roads. In the UK, the government plans to permit driverless vehicle operations on roads by 2026, supported by the Automated Vehicles Bill. 

In Spain, the development of autonomous driving has also progressed in recent years, particularly in the experimental domain. The country has hosted several pilot projects in urban and highway settings, such as those developed under the European AUTOPILOT consortium. A recent example is a pioneering project conducted in the city of Leganés, where a driverless urban bus operated for five days without incident, demonstrating the viability of this technology in real-world environments. This initiative garnered significant attention and positive feedback from both the public and local authorities. 

At the community level, the European Commission has outlined a clear roadmap through its Sustainable and Smart Mobility Strategy, setting goals for automated, connected, and climate-neutral mobility by 2050. 

AI Development at ARQUIMEA 

At ARQUIMEA Research Center, we believe in the power of technology to transform the future of mobility, and we are committed to contributing to this development. 

Through our Artificial Intelligence division, we drive innovative solutions that enable safe and efficient technological advancements, researching and working in key areas such as Edge Machine Learning and Safe Autonomy. 

All projects at ARQUIMEA Research Center are part of QCIRCLE, an initiative funded by the European Union aimed at establishing a center of scientific excellence, thereby reinforcing our commitment to innovation and research in addressing current global challenges.