GPS Time Synchronization: Principles, Methods, and Applications in Mapping Vehicles

1. Overview

Mapping vehicles are equipped with cameras, LiDAR, inertial navigation units, and other sensors that generate image, point‑cloud, and trajectory data. To fuse these data into a coherent map, all sensor streams must be temporally aligned. This alignment is achieved by a time synchronization system that uses GPS‑based time dissemination.

2. GPS Timing Principle

GPS satellites carry high‑precision atomic clocks (cesium). The satellites maintain tight time synchronization, and a receiver can determine its clock offset relative to the satellites. At least four satellites are required to resolve the clock bias and compute a navigation solution. Once the receiver knows its clock offset, it can correct its local clock – a process called time transfer or 授时 .

The atomic clock works because electron transitions between energy levels emit radiation at a highly stable frequency, which serves as an ultra‑stable pendulum.

GPS time transfer provides two types of time signals:

1) A 1‑second Pulse‑Per‑Second (PPS) signal whose rising edge is synchronized to UTC with an error < 1 µs.

2) Serial data (NMEA‑0183) that carries full date‑time information, e.g., $GPGGA, $GPRMC.

Example NMEA RMC sentence:

$GPRMC,161229.487,A,3723.2475,N,12158.3416,W,0.13,309.62,120598,*10

3. GPS Timing Methods

3.1 PPS and NMEA Relationship

The PPS pulse arrives before the NMEA sentence. The interval between them varies by manufacturer (a few milliseconds to several hundred milliseconds).

Yellow line: PPS (rising edge marks the exact start of a UTC second). Blue line: NMEA time information (date, hour, minute, second).

3.2 Detailed GPS Timing Process

The GPS receiver, when in positioning mode, outputs a PPS pulse and serial NMEA data (TTL/RS‑232, ASCII, configurable baud 9600‑460800). The RMC sentence is sufficient for extracting UTC time.

The MCU uses an external crystal oscillator (OCXO preferred, 3 ppb stability) as its clock source. The synchronization steps are:

Extract UTC time (hhmmss, date) from the RMC sentence and set the MCU’s system clock.

Capture the PPS rising edge via an interrupt; reset sub‑second counters to align the system clock to the exact second.

After a few seconds, compare the GPS‑derived time with the MCU’s time to detect possible ±1 s jumps caused by PPS alignment issues.

Test results from various scenarios (overpasses, malls, ring roads, streets) over 5 h 38 min showed a second‑level error of 0 s and a sub‑second error ≤ 4 µs.

3.3 Camera Time Synchronization

Multiple cameras mounted on the vehicle capture images that must be matched with trajectory data. The time sync system provides a unique timestamp for each image:

Cameras operate in external‑trigger mode; the MCU supplies the trigger pulse and records its sequence number.

During exposure, the camera emits a pulse that the MCU captures, logging the exact timestamp and sequence.

The logged timestamps are stored alongside image files, enabling precise association with GPS‑derived positions.

4. GPS Timing Anomalies and Handling

Real‑world deployments encounter irregularities due to signal loss, lock acquisition delays, and environmental influences. Notable anomalies include:

4.1 PPS vs. Crystal Oscillator

Three alignment cases:

PPS edge aligns perfectly with the crystal clock – ideal but rarely observed.

MCU clock runs slower than 1 s; PPS arrives before the MCU reaches the next second. The system must truncate the current second early and jump to the next second (+1 s).

MCU clock runs faster; the MCU has already entered the next second before PPS arrives. The system must roll back to the start of the current second (no ±1 s adjustment).

4.2 PPS vs. GPS Time Information

Occasionally the NMEA time (e.g., GNRMC) drifts, causing a “time jump” where the PPS marks the start of second N, but the NMEA sentence still carries second N‑1. This results in a one‑second backward jump that must be detected and corrected.

Various mitigation strategies can be applied, such as buffering NMEA data, cross‑checking with PPS timestamps, or discarding out‑of‑order messages.

5. Conclusion

Understanding GPS timing principles and methods enables the design of robust, high‑precision time‑synchronization systems for sensor fusion in mapping vehicles. Future work may involve testing different GPS modules in challenging environments and refining the synchronization algorithms.

6. Appendix

Caption: PPS signal processing – anti‑interference and noise filtering.

Caption: GPS time signal level conversion.

Caption: Crystal oscillator signal conditioning – clock voltage control and DC filtering.