GB 47955 New Standard Takes Effect: Voyager ISL Addresses New ADAS Requirements with an "Eyes-Brain-Body" Technical Framework
Safety is the top priority for every trip.
Data published by the World Health Organization (WHO) shows that approximately 1.16 million people died in road traffic accidents worldwide in 2025.
How can driving risks be reduced? Intelligent vehicles are offering a new solution.
As Level 2 (L2) advanced driver-assistance systems (ADAS) gain widespread adoption, the industry’s focus has shifted from the quantity of onboard functions to the safety and reliability of the entire system.
On June 27, GB 47955–2026 Intelligent and connected vehicle — Safety requirements of combined driver assistance system — China’s first mandatory safety standard specifically targeting L2 combined driving assistance systems — was officially issued.
This new standard ushers in an era governed by formal safety benchmarks for ADAS.
What core capabilities must a truly reliable ADAS system possess?
We interviewed engineers from Voyager ISL to break down the technical challenges brought by the new national standard.
Voyager ISL Engineers’ Insights on GB 47955
Q1 What core new safety constraints does the new standard introduce compared with previous industry specifications?
Perception Algorithm Engineer
The new standard establishes a unified safety access threshold. Instead of merely evaluating the number of available functions, it imposes mandatory full-link safety requirements for the entire system, with particularly prominent updates to perception-related specifications.
In the past, many products relied on a single sensor channel for environmental detection. The new national standard mandates extensive mandatory testing under real-world road conditions, including tunnels, rain and snow, temporary road construction, and mixed traffic with non-motor vehicles.
It also explicitly requires multi-channel redundancy design for perception systems. If one sensor channel malfunctions, other sensors can continue stable operation to identify surrounding road users and obstacles.
Furthermore, the standard raises higher bars for perception modules beyond basic image capture: systems must classify detected objects and conduct hierarchical risk assessment, outputting credible environmental data for downstream decision-making modules.
In the framework of this new standard, stable environmental comprehension via perception becomes a critical threshold for products to achieve compliance.
Hardware merely enables the system to “see”; genuine safety value lies in the ability to “understand” road conditions.
Q2 The new standard enforces strict risk exit mechanisms for out-of-bound operating conditions and sudden hazards. What operational logic underpins Voyager ISL’s decision-making framework to manage such risks?
Decision-Making Algorithm Engineer
Simply put, the decision-making layer continuously predicts the movement of surrounding traffic while cross-verifying real-time road data transmitted from perception modules. It evaluates whether the vehicle is operating within the valid operational scope of the ADAS functions.
If the decision layer detects rising road risks, degraded sensor recognition performance, or the vehicle drifting outside the functional usable zone, it will not abruptly disable assistance functions.
Built on Voyager ISL’s proprietary VT-Vulcan® hierarchical decision control architecture, the system executes a gradual risk mitigation workflow: multi-stage driver alerts first, followed by gentle speed adjustment, then phased handover of vehicle control authority back to the driver.
Throughout this control sequence, the decision unit restricts harsh vehicle maneuvers such as emergency hard braking, unintended acceleration, and sharp steering wheel turns, balancing occupant comfort with driving safety.
Every driving strategy is rigorously regulated to uphold the baseline of vehicle safety.
Q3 Why does the new standard mandate enhanced driver monitoring for all L2 vehicles?
Perception Algorithm Engineer
No matter how intelligent ADAS technology becomes, the driver remains the primary party liable for driving safety — a principle reaffirmed in the regulatory text. Therefore, intelligent vehicle designs must monitor both road conditions and the driver’s state.
Previously, most models assessed driver engagement solely via steering wheel torque detection, a method with notable limitations.
The new standard requires upgraded driver monitoring capabilities to accurately gauge driver attention levels and issue timely warnings.
Driver Monitoring System (DMS) functionality cannot be achieved by simply installing a camera; recognition performance is compromised by nighttime lighting, sunglasses, and complex variable illumination.
Voyager ISL’s VT-DMS driver monitoring system leverages visual perception technology to comprehensively identify the driver’s head posture, gaze direction, and hand positioning, delivering robust human-machine collaboration support for ADAS.
Smart vehicles must not only interpret the road ahead, but also understand the driver behind the wheel.
Q4 What will define competitiveness in the future ADAS market?
Platform Architecture Engineer
I believe future competition will no longer center on the sheer volume of onboard functions. Instead, the integrity, stability, and safety of the complete system will be the decisive differentiators.
Human driving relies on three core components: eyes to observe surroundings, the brain to make judgments, and the body to execute physical actions. Vehicles follow an identical logic.
The “Eyes” handle complex environmental perception; the “Brain” delivers decision-making and path planning; the “Body” executes vehicle control.
Centered on intelligent perception, intelligent decision-making, and precise control execution, Voyager ISL continues to develop mass-production-ready ADAS solutions, enabling intelligent vehicles to not only “see” road conditions, but also “understand” them and “act” with stability.
Eyes | Intelligent Perception
VT-Sensing Visual Perception + VT-Navi Multi-source Fusion Positioning
Integrated vision-computing architecture, proprietary ISP chips, superior low-light performance, resistance to positioning drift caused by tunnels and tall buildings.
Consistently capture accurate, authentic environmental data.
Brain | Intelligent Decision-Making
Mass-production-ready highway NOA and automatic parking assist (APA) solutions built on VT-Vulcan® cabin-chassis integrated architecture, compatible with mainstream automotive chips.
Deliver environmental comprehension, traffic behavior prediction, and path planning to generate stable, safe vehicle control strategies.
Body | Execution & Control
VT-Vulcan® translates decision-making commands into precise vehicle control over steering, braking, and acceleration, enabling seamless coordination with the full vehicle electronic system for safe maneuver execution.
New Standard, New Starting Line
The release of GB 47955 marks a pivotal shift in ADAS industry competition — from a race to pack individual features to a contest of comprehensive system performance.
Moving forward, the quantity of intelligent vehicle functions will matter far less than consistent safety and reliability across complex real-world driving scenarios.
Voyager ISL will continue advancing its technical ecosystem covering intelligent perception, intelligent decision-making, and precise vehicle control to elevate the overall performance of driving assistance systems.