Abstract: To address target loss caused by continuous geometric occlusions in outdoor environments, a following-target localization method incorporating corner constraints is proposed for mobile robots. First, an adaptive extended Kalman filter (EKF) robustly fuses LiDAR and draw-wire pan-tilt sensor data, mitigating mechanical jitter under high-dynamic conditions. Secondly, within blind zones, the target's azimuth is modeled as a uniform distribution using environmental geometry and sensor-measured relative positions. An optimal position estimate is derived via the minimum mean square error (MMSE) criterion for occlusion state estimation. Simulations show that under continuous multi-corner occlusions and non-stationary interference, the method limits the localization RMSE to 0.24 m, reduces recovery time to 0.8 s, and achieves a 100% corner-passing success rate. This effectively suppresses tangential divergence in blind zones, ensuring stable robot following in complex occluded environments.