GNSS+IMU+MM Vehicle Integrated Navigation System

With the rapid development of positioning services, vehicle positioning requirements have risen from navigation-level to lane-level accuracy. In urban canyon environments with tall buildings and viaducts, GNSS signals may be blocked or interfered, leading to missing positioning results or poor precision. This inherent weakness of active positioning cannot be eliminated merely by algorithms. Multi-sensor fusion schemes such as GNSS+IMU have attracted wide attention, since passive IMU can compensate for the shortcomings of satellite navigation.

GNSS is a space-based radio navigation system. It calculates real-time positioning using pseudorange, carrier wave, ephemeris, time and clock offset from satellites. It provides all-weather 3D coordinates, velocity and timing information on the Earth surface and near-Earth space. Its advantages are high precision and non-divergent error. However, it is vulnerable to occlusion by trees and buildings, as well as multipath effects from reflective surfaces.

INS is an autonomous navigation system relying on IMU. It does not depend on external information or radiate signals outward, and works in air, on ground and underwater. Based on Newton’s laws, it measures carrier acceleration within the inertial reference frame, integrates acceleration over time and converts results into the navigation frame to obtain velocity, yaw and position. It works without external signals and can be installed freely for full-scene navigation. Its drawbacks are limited absolute precision and time-divergent error, which makes it complementary to GNSS.

Map matching combines position data with map information to determine the precise road position of the vehicle and support accurate vehicle navigation control.

Dead Reckoning is a tracking navigation algorithm. Given the current coordinates, it calculates the next position using movement distance and heading measured by IMU. This article focuses on IMU-aided integrated navigation built upon GNSS solutions.

Vehicle navigation has developed for decades, while new challenges emerge as precision requirements increase:

• Yaw recalculation: Position drift caused by signal occlusion under viaducts or urban canyons.
• Position failure: Low estimation accuracy in signal-free zones such as tunnels and underground parking lots, resulting in large exit positioning errors.
• Incorrect road matching: Misidentification between main lanes, side lanes and upper/lower viaduct roads.

Yaw recalculation and position failure stem from the nature of GNSS and are hard to resolve fundamentally. Wrong road matching occurs when adjacent roads are too close to distinguish with limited positioning precision; the root cause is that only position data is used without leveraging other sensor information.

The above positioning technologies have respective strengths and complement each other. The GNSS+IMU+MM fusion scheme adopts DR algorithm and POS/heading data to improve positioning reliability. The major vehicle positioning issues are resolved step by step:

1. Data Fusion: Process position, velocity, time and heading output from GNSS in real time. Evaluate GNSS data quality and switch positioning strategies accordingly.
2. Sensor Compensation: When GNSS signal is weak or lost, DR based on IMU provides navigation support. GNSS measurements calibrate velocity sensor scale factor and IMU error parameters including gyro scale factor on vertical axis and three-axis gyro bias. With compensation, DR can maintain centimeter-level positioning for a short period without satellite signals.
3. Scene Recognition: Determine vehicle map location by combining high-precision map data and real-time environmental perception from LiDAR or millimeter-wave radar to assist environmental awareness and decision-making.

Take MS-200 high-precision vehicle integrated navigation unit as an example. It adopts high-performance MEMS inertial sensors, high-precision GNSS board and vehicle odometer for integrated navigation. It outputs centimeter-level positioning, velocity and attitude in real time and retains acceptable navigation precision after GNSS loss. It applies to passenger cars and unmanned vehicles equipped with intelligent driving systems.

Intelligent driving systems require real-time and autonomous acquisition of precise vehicle position, velocity and attitude. MS-200 features high measurement precision, small size and low power consumption to meet such demands.

Core sensors can be customized for different projects. Functional-safety IMU serves as the core sensor. The Z-axis (vehicle heading) gyro can be upgraded to higher-precision versions to improve pure inertial navigation performance during GNSS outages. Its RTK-enabled GNSS module supports multiple constellations including BDS B1/B2, GPS L1/L2, GLONASS L1/L2 and GALILEO E1/E5b, with single-antenna positioning and dual-antenna heading options.