TACO: Trajectory Aligning Cross-view Optimisation

2026-05-04

AI Summary

This work proposes a high-precision, low-power vehicle localization method for GNSS-denied or GNSS-degraded environments. Relying on only a single initial GNSS fix, the approach tightly couples inertial measurement unit (IMU) data with fine-grained cross-view geolocalization (CVGL). Key innovations include a loop-closure triggering mechanism, a forward-biased five-point multi-crop search, heading residual gating, and anisotropic noise modeling, integrated within a framework combining a closed-form cross-track error model, an Unscented Kalman Filter, and factor-graph optimization with loop-closure verification. Evaluated on the KITTI dataset, the method reduces the median absolute trajectory error from 97.0 meters to 16.3 meters—a 5.9× improvement—while achieving per-frame fusion latency under 0.1 ms and maintaining a camera duty cycle of only 5–10%.

Abstract

Cross-View Geo-localisation (CVGL) matches ground imagery against satellite tiles to give absolute position fixes, an alternative to GNSS where signals are occluded, jammed, or spoofed. Recent fine-grained CVGL methods regress sub-tile metric pose, but have only been evaluated as one-shot localisers, never as the primary fix in a live pipeline. Inertial sensing provides high-rate relative motion, but accumulates unbounded drift without an absolute anchor. We propose TACO, a tightly-coupled IMU + fine-grained CVGL pipeline that consumes a single GNSS reading at start-up and thereafter operates on onboard sensing alone. A closed-form cross-track error model triggers CVGL before IMU drift exceeds the matcher's capture radius, and a forward-biased five-point multi-crop search keeps inference cost fixed at five forward passes per fix. A yaw-residual gate rejects fixes that disagree with the onboard compass, and an anisotropic body-frame noise model scales each Unscented Kalman Filter update by per-fix confidence. A factor graph with vetted loop closures provides an offline smoothed trajectory. On the KITTI raw dataset, TACO reduces median Absolute Trajectory Error (ATE) from 97.0m (IMU-only) to 16.3m, a 5.9 times reduction, at <0.1 ms per-frame fusion cost and a 5-10% camera duty cycle. Code is available: github.com/tavisshore/TACO.

Problem

• Cross-View Geo-localisation
• Inertial Navigation
• Absolute Positioning
• Sensor Fusion
• Trajectory Estimation

Innovation

• Cross-View Geo-localisation
• Tightly-coupled IMU-CVGL
• Unscented Kalman Filter
• Factor Graph
• Absolute Trajectory Error