LG Vehicle Solution (LG VS) experts explain that as Advanced Driver Assistance Systems (ADAS) takes on more driving situations, the boundaries of vehicle safety are becoming more complex. Advances in safety technology have increased confidence in vehicle performance, while also expanding the responsibility placed on safety systems.
As drivers rely more on ADAS, safety can no longer be considered solely in terms of system performance. It increasingly depends on how the roles of the vehicle and the driver are defined, and how they work together across different driving situations.
This shift is also reflected in the evolution of the New Car Assessment Programme (NCAP). Its scope has expanded from occupant protection to accident avoidance, with greater emphasis now being placed on driver attention and state. As Euro NCAP 2029 extends this scope further, the boundaries of safety evaluation are continuing to evolve.
KEY TAKEAWAY: NCAP is evolving beyond vehicle performance to include a deeper understanding of driver state. Integrating ADAS with in-cabin sensing will be increasingly important as this evaluation scope expands.
Myoung Hoon Kim | Senior Professional, Advanced Product Planning, LG Vehicle Solution
The NCAP has evolved into a globally recognized benchmark for vehicle safety, with frameworks adopted across Europe, the U.S., Korea, Japan, ASEAN, and Latin America. When I look at how NCAP has evolved, I see a clear expansion in what we consider vehicle safety. It began with crash protection, then expanded to accident avoidance as ADAS, including Automatic Emergency Braking (AEB), became part of the evaluation.
Today, we are moving beyond whether an ADAS function is simply available. What matters is how reliably it performs in realistic driving scenarios, including intersections and encounters with vehicles, pedestrians, and cyclists.
At the same time, we are looking more closely at the driver. Drowsiness and distraction are already part of safety evaluation, while Euro NCAP 2029 will extend the scope to cognitive distraction and impairment, including alcohol-related conditions and medical emergencies. From my perspective, this evolution makes the integration of ADAS and driver state monitoring through in-cabin sensing increasingly important. Safety evaluation is moving toward understanding both the driving environment and the driver as part of one system.
KEY TAKEAWAY: ADAS is evolving beyond perception accuracy to consider whether the driver understands the situation and is ready to respond. A Human-in-the-loop approach enables safety decisions to reflect both the driving environment and the driver.
Min Kyu Kim | Leader, AIDV Development Strategy, LG Vehicle Solution
The way we approach ADAS design is changing. In the past, development focused primarily on perception accuracy and how effectively the system could intervene when necessary.
Today, ADAS decisions need to consider both the driving situation and the driver's state. Even when a hazard is accurately detected, the appropriate response can differ depending on whether the driver recognizes the situation and is ready to respond.
This means the goal is no longer perception accuracy alone. ADAS also needs to understand the driver's awareness and readiness to accept assistance. One example is the shift from Eyes-on-Road to Eyes-on-Target, which looks beyond where the driver is looking to whether they are actually aware of and engaged with the driving target.
Ultimately, I see ADAS evolving from a system that makes decisions independently toward a shared decision-making framework built around a Human-in-the-loop approach, with the driver remaining an integral part of the safety response.
KEY TAKEAWAY: Understanding driver state requires multiple signals to be interpreted together, rather than relying on a single sensor. ADAS can then integrate that understanding with the driving environment to support real-time safety decisions. At LG, we are advancing integrated sensing and decision-making frameworks that combine driver state and environmental understanding for adaptive vehicle control.
Ho San Han | Leader, ADAS & Vehicle Intelligence Project, LG Vehicle Solution
As driver state becomes part of safety decision-making, relying on a single sensor is no longer enough. Recent NCAP developments and real-world cases are expanding the focus beyond inattention and drowsiness to impairment caused by alcohol or reduced cognitive ability.
From a development perspective, I believe we need to integrate multiple inputs, including driver gaze, reaction time, behavior patterns, and biometric signals. The goal is to understand the driver's overall condition rather than interpreting each sensor signal independently.
The next challenge is making that understanding actionable. Driver state needs to be reflected in the system in real time, allowing ADAS to consider both the external driving environment and what is happening inside the vehicle.
Ultimately, I see ADAS evolving toward an integrated system that recognizes the driving environment and driver state together, and incorporates both into vehicle control.
KEY TAKEAWAY: Over-reliance reflects the gap between increasingly capable Level 2 systems and drivers' expectations of Level 3 automation. Managing that gap requires ADAS to balance safety and convenience while keeping driver responsibility clearly defined.
Myoung Hoon Kim | Senior Professional, Advanced Product Planning, LG Vehicle Solution
As ADAS becomes more capable, drivers can begin to expect the vehicle to handle more of the driving task. The challenge is that a Level 2 system can create expectations closer to Level 3, even though the driver remains fully responsible for driving.
From my perspective, this expectation gap is at the heart of over-reliance. Simply limiting ADAS is not a realistic solution, especially as drivers have already experienced the convenience these features provide and increasingly expect them as part of the driving experience.
The real challenge is managing Level 3 expectations within the boundaries of Level 2. This means keeping the regulatory and responsibility framework clearly aligned with Level 2 while continuing to improve both safety and convenience.
At LG, we are approaching this by considering the driver's state, intent, and driving situation together. Rather than restricting technology, the goal is to develop ADAS that adapts to context while remaining within clearly defined Level 2 boundaries
KEY TAKEAWAY: Safety does not depend on a single sensor or algorithm being perfect. Integrating ADAS and driver state monitoring enables more reliable, risk-adaptive intervention.
Ho San Han | Leader, ADAS & Vehicle Intelligence Project, LG Vehicle Solution
ADAS and driver state monitoring both operate in highly variable real-world conditions. No single sensor or algorithm can interpret every driving situation or driver behavior perfectly.
From a development perspective, I believe the goal should not be perfect perception or decision-making. Instead, we need to understand the confidence of each algorithm and design systems that complement one another.
By integrating ADAS decisions with driver state monitoring, we can improve overall system reliability and adapt the timing and level of intervention to the actual driving risk. This is why we are developing a state-based architecture that accounts for the inherent limitations of individual algorithms.
For me, this approach represents a practical direction for safety design within the capabilities of today's technology.
KEY TAKEAWAY: AI can turn individual safety signals into a contextual understanding of driver risk over time. This enables more adaptive intervention while enhancing both safety and the driving experience.
Min Kyu Kim | Leader, AIDV Development Strategy, LG Vehicle Solution
Sensors and algorithms have inherent limitations. From my perspective, the next step is not simply improving their individual performance, but understanding the driving situation and driver state together in context. This is where AI can play an important role.
Drowsiness detection is a good example. Detecting drowsiness alone does not prevent every accident. The system also needs to determine when, how, and at what level to intervene within a human-in-the-loop framework.
AI can help by interpreting driver state as a pattern that changes over time, rather than as a single event. This allows the vehicle to anticipate changes in driver state and adapt the timing, method, and intensity of intervention to the situation. When necessary, the driving strategy itself can also be adjusted to help reduce risk.
For drivers, I see this creating a system that goes beyond warning to understand and protect them. For OEMs, AI can enhance existing safety systems with minimal architectural changes while creating new opportunities to differentiate the safety experience
Throughout this discussion, one direction has become clear: vehicle safety can no longer rely on individual technologies working in isolation.
ADAS, driver state recognition, and AI are increasingly coming together as an integrated decision-making framework that considers both the driving environment and the driver.
The next step is not simply higher performance. It is about understanding context and coordinating the right intervention at the right moment to deliver safer, more adaptive driving experiences.
At LG, we believe the future of vehicle safety lies in orchestrating ADAS, driver state recognition, and AI within a unified vehicle intelligence framework. Through this integrated approach, we aim to deliver safer, more adaptive, and context-aware driving experiences.
Byeong Min Kwon | Moderator, ADAS & Vehicle Intelligence, Advanced Development, LG Vehicle Solution
The URL has been copied to the clipboard.
Your email has been successfully registered.
The restocking notification has already been updated.
Add items to your MY LG wishlist
OTP authentication failed. Please verify your mobile number to use for mobile OTP login.
