NGP

NGP (Navigation Guided Pilot) is one of the core features marking the evolution of Advanced Driver Assistance Systems (ADAS) towards high-level intelligent driving. As a pivotal technology in the automotive intelligence race, different carmakers adopt distinct naming conventions (such as XPeng's NGP, NIO's NOP, and Huawei's NCA). In essence, it leverages high-definition (HD) maps (or mapless vision-based solutions), multi-sensor fusion perception, and high-computing-power autonomous driving chips to deliver point-to-point navigation-assisted driving within designated areas (such as highways, urban expressways, and complex urban open roads). Once the driver sets a navigation destination and activates the system, the vehicle can autonomously handle a full suite of driving tasks, including speed regulation, overtaking, lane changing, navigating ramps, negotiating complex junctions, and obstacle avoidance.

Core Technical Architecture and Operating Principles

The technical architecture of NGP primarily comprises multi-sensor fusion perception hardware, high-computing heterogeneous domain controllers, HD map/navigation engines, and underlying drive-by-wire chassis actuators (such as steer-by-wire and brake-by-wire):

Multi-Modal Sensor Fusion and 3D Spatial Modelling:

The system's perception layer is typically composed of automotive-grade LiDAR, long-range millimetre-wave radar, high-resolution cameras, and ultrasonic sensors. Using sophisticated deep learning algorithms, the vehicle fuses heterogeneous data in real time within on-board processing chips (such as the NVIDIA Orin or higher-tier platforms) to construct a high-precision 360-degree dynamic 3D environmental model, accurately identifying lane markings, traffic lights, moving vehicles, pedestrians, and smaller road elements such as traffic cones.

Behavioural Planning Decision-Making and Autonomous Game-Theoretic Control:

Relying on global path planning via HD maps and real-time navigation feeds, the system's decision-making and planning algorithms emulate the logic of experienced drivers to compute the optimal driving trajectory in real time. When encountering slower traffic ahead, the system proactively evaluates safe clearances in adjacent lanes and the speed of oncoming vehicles from behind, autonomously activating the indicators to execute smooth overtaking lane changes. As it nears the exit ramp, the system performs multi-lane transitions well in advance to seamlessly enter the slip road, achieving end-to-end automated control.

Safety Guidelines and Physical Limitations 

Strictly Prohibiting Hands-Off Driving, Sleeping, Phone Distractions, or Diverting Attention from Driving While NGP is Active: 

NGP remains an Advanced Driver Assistance System and is fundamentally assistive in nature, which does not alter the driver's legal responsibility as the primary person accountable for vehicle safety. It must strictly not be treated as a fully autonomous driving system (FSD or Level 3 and above). Drivers must stay alert to road conditions at all times, keeping their hands lightly on the steering wheel, ready to intervene instantly should the system encounter unrecognisable roadworks, severe weather, or aggressive cut-ins.

Strictly Prohibiting Blind Reliance on NGP in Severe Weather, Roads Without Clear Lane Markings, Heavily Waterlogged Stretches, or Sudden Major Road Diversions:

High-level intelligent driving systems rely heavily on physical environmental perception and navigation data. Engaging NGP under high-risk conditions—such as when sensors are caked in mud, or when visual cameras are blinded by strong backlighting or torrential rain—is strictly prohibited. In scenarios involving manual traffic hand signals, complex traffic cone configurations at construction sites, or unmarked rural roads, the system is prone to perception dropouts or path-planning conflicts, requiring the driver to promptly disengage the system and take manual control.