车载惯性导航技术

A brief history of automotive dead reckoning (ADR)

The term dead reckoning comes from early maritime navigation methods where position was estimated based on speed, direction, and time, without external references. The word "dead" in this context is believed to derive from "deduced," as in deduced reckoning, a method of estimating position without external references. In the automotive world, this concept evolved into automotive dead reckoning (ADR), a technique that combines GNSS with onboard sensors to maintain accurate positioning when satellite signals are weak or lost.

Today, ADR is a cornerstone of advanced vehicle positioning systems, ensuring continuous and reliable navigation even in the most challenging environments.

Problems addressed by ADR

Modern vehicles often operate in environments where GNSS signals are degraded or completely unavailable, such as:

• Urban canyons with tall buildings
• Tunnels and underground parking
• Forested or mountainous terrain
• Multi-level road systems

In these scenarios, relying solely on GNSS can lead to position jumps, signal loss, or inaccurate heading estimation. ADR addresses these issues by fusing GNSS data with inertial sensors and vehicle dynamics, ensuring:

• Continuous positioning (availability) even without satellite visibility
• Improved heading estimation, especially at low speeds or standstill
• High-frequency updates (e.g., 50 Hz), enabling smoother and more responsive navigation

Technology: how automotive dead reckoning (ADR) works

Dead reckoning (DR), also called sensor fusion, estimates position based on motion data from onboard sensors, including accelerometers and gyroscopes, collectively known as an inertial measurement unit (IMU). 
Automotive dead reckoning (ADR) uses additional inputs such as wheel ticks or vehicle speed to improve system performance and reliability. ADR therefore requires access to vehicle systems (e.g., CAN bus), which is feasible in some first-mount and aftermarket applications. 

ADR combines different data inputs:

• GNSS signals, providing absolute velocity and heading
• Inertial measurement units (IMUs): accelerometers and gyroscopes
• Vehicle sensors: wheel ticks (WT) or vehicle speed, and drive direction
• Dynamic model: vehicle dynamics limitations, so called non-holonomic behavior (for instance, a car cannot move laterally or vertically)

A Kalman filter continuously fuses the inputs and estimates the values that are not known. In an initial phase, GNSS signals are compared with inertial measurements to estimate sensor biases and drifts. Alignment is also estimated, if not manually configured. The calibration is automatic, and it typically completes within a few minutes under strong GNSS signal conditions. Frequent changes in speed and direction, such as doing figure eights, can accelerate the calibration.

After completing the calibration, the filter corrects drift, smooths noise, and enables precise position, velocity, and heading estimation through continuous prediction and correction cycles. This enables robust estimation of position, velocity, and heading, even in difficult urban environment and during full GNSS outages.

ADR vs. UDR

ADR existed first and untethered dead reckoning (UDR), emerged from ADR, with "untethered" referring to the absence of direct vehicle speed or wheel tick inputs, relying solely on inertial sensors and GNSS data.  The main differences between the two technologies are shown below. 
 

Feature	Automotive dead reckoning  (ADR)	Untethered dead reckoning  (UDR)
Sensor inputs	GNSS IMU Vehicle sensors	GNSS IMU
Accuracy	Higher,  especially in GNSS-denied areas	High,  for limited outage periods less accurate standstill detection during GNSS outage
Use case	Automotive OEMs Fleet telematics for public transportation	Various kinds of fleet management (trucking, last mile delivery, car-sharing, long-term rental) Aftermarket navigation
Alignment	Typically using manual alignment configuration	Typically, fully self-calibrating
Conclusion	Best position accuracy	Simplest integration  Much better position accuracy compared to traditional GNSS receivers without DR

Applications for ADR

ADR is used across a wide range of mobility solutions:

Automotive first mount

• Integrated into car OEM navigation systems
• Essential input to vehicle-to-everything (V2X) systems for safer roads
• Supports motorcycles with dedicated dynamic models

Aftermarket telematics

• Used in fleet management and passenger information systems for buses and trains for efficient operational management and smooth customer experience

Shared micromobility

• Applied in e-scooters and e-bikes with dedicated dynamic models for more safety and compliance in regulated zones

Variants and product grades

ADR solutions are available in different performance and environmental grades:

Standard precision (SPG) versus high precision GNSS (HPG)
 

Variant	Standard precision GNSS  (SPG)	High precision GNSS  (HPG)
Accuracy	Meter-level	Centimeter to decimeter-level  (with RTK/PPP)
Use Case	General navigation	Lane-level positioning Autonomous driving
Cost	Lower	Higher

Product grades

Dead reckoning products come with different product grades. Automotive grades are typically selected by first-mount automotive customers for OEM integration and are available with an extended operating temperature range from -40°C up to 105°C. Professional grade products are typically used in all other applications. 
 

Grade	Operating temperature	Use case
Automotive	-40°C to +105°C -40°C to +85°C	Automotive first mount (OEM integration)
Professional	40°C to +85°C	Aftermarket telematics Shared micromobility

Products recommended by market segment

The following table shows the products recommended by market segment. Chip products are also available for high-volume OEM applications (but not shown in this overview). We recommend that customers ordering hundreds of thousands of pieces per year fill in the project information form.

		m-level	cm- to dm-level
Automotive	First-mount (OEM integration)	NEO-M9L ZED-F9L	ZED-F9K
Industrial	Aftermarket telematics: cars, trucks and buses	NEO-M9V	ZED-F9R
	Aftermarket telematics: trains	ZED-F9R	ZED-F9R*
	Shared micromobility	NEO-M9V	ZED-F9R

                                           ^{* RTK performance not specified on trains}

Conclusion

Urban environments will continue to challenge GNSS-based positioning. Untethered dead reckoning provides a resilient, scalable solution that enhances positioning availability and accuracy in complex environments. By intelligently combining GNSS and inertial sensing, ADR enables robust performance across a wide range of mobility applications.

Once a suitable product is identified from the table above, it is most common practice to get started with the respective evaluation kit and u-center.