Seeing through the noise: multipath mitigation reaches a new level with ALCI (Adaptive Long Coherent Integration)

How adaptive long coherent integration delivers superior measurement and positioning performance where conventional receivers fall short.

Multipath interference is one of the most persistent challenges in GNSS. In dense city streets, under tree canopies, or anywhere signals bounce off surfaces before reaching the antenna, conventional receivers struggle to tell the true line-of-sight signal apart from its reflections. The result is position errors that erode confidence precisely where reliable navigation matters most.

At u‑blox, we address this challenge in a fundamentally new way, through well‑designed hardware featuring one of the most stable local clocks among comparable products, without sacrificing the real‑time performance demanded by safety‑critical and high‑dynamic applications. The result is Adaptive Long Coherent Integration (ALCI): a technique that adaptively incorporates Sensor‑Aided Long coherent Integration (SALI), sensor‑less long coherent integration, and u‑blox legacy tracking loops across different environments, hardware conditions, and use cases.

By coherently integrating the signal over an extended window, we project received signals onto a line-of-sight hypothesis in the delay-Doppler domain. Multipath components of a time-variant channel are separated and suppressed. What remains is a cleaner measurement of the true line-of-sight path – identified by its shortest delay. An additional multipath warning flag and NLOS indicator give the positioning engine the information it needs to further filter unreliable measurements, which is especially valuable in deep urban scenarios.

Validated through extensive road tests in Switzerland, Germany, the USA, and Japan, a SALI test receiver operating with only 14 tracking channels and tracking exclusively Galileo E1 and BDS B1C demonstrates remarkable performance gains across all challenging environments. The 95th‑percentile horizontal error is reduced by 7×, 3×, and 2× in foliage, deep‑urban, and suburban scenarios, respectively. The proportion of position fixes within 2 m error nearly doubles. In the Stanford diagram, the hazardously misleading information rate across all environments decreases by 3×. Importantly, the overall CPU load on an ARM Cortex‑M3 is reduced by approximately 25%, as high‑rate carrier and code tracking loops are no longer required.

Looking ahead, the approach extends naturally to L5 wideband signals and holds promise for anti-jamming, anti-spoofing, and high-rate 100 Hz PVT output. Combined with L1/L5 dual-band processing, sub-meter accuracy in challenging environments becomes a realistic near-term target.

Download the full ION GNSS+ paper and presentation below:

Real-Time Multipath Mitigation with Sensor-Aided Long Coherent Integration (SALI) – White paper

Real-Time Multipath Mitigation with Sensor-Aided Long Coherent Integration (SALI) – Presentation

Courtesy of u-blox