Lidar-based localization doesn’t have high accuracy in open scenarios with few features, and behaves poorly in robot kidnap recovery. To address this problem, an improved Particle Filter localization is proposed who could achieve robust robot kidnap detection and pose error compensation. UAPF adaptively updates the covariance by Jacobian from Ultra-wide Band information instead of predetermined parameters, and determines whether robot kidnap occurs by a novel criterion called KNP (Kidnap Probability). Besides, pose fusion of ranging-based localization and PF-based localization is conducted to decrease the uncertainty. To achieve more accurate ranging-based localization, linear regression of ranging data adopts values of maximum probability rather than average distances. Experiments show UAPF can achieve robot kidnap recovery in less than 2 s and position error is less than 0.1 m in a hall of 40 by 15 m, when the currently prevalent lidar-based localization costs more than 90 s and converges to wrong position.

Real-time performance and global consistency are extremely important in Simultaneous Localization and Mapping (SLAM) problems. Classic lidar-based SLAM systems often consist of front-end odometry and back-end pose optimization. However, due to expensive computation, it is often difficult to achieve loop-closure detection without compromising the real-time performance of the odometry. We propose a SLAM system where scan-to-submap-based local lidar odometry and global pose optimization based on submap construction as well as loop-closure detection are designed as separated from each other. In our work, extracted edge and surface feature points are inserted into two consecutive feature submaps and added to the pose graph prepared for loop-closure detection and global pose optimization. In addition, a submap is added to the pose graph for global data association when it is marked as in a finished state. In particular, a method to filter out false loops is proposed to accelerate the construction of constraints in the pose graph. The proposed method is evaluated on public datasets and achieves competitive performance with pose estimation frequency over 15 Hz in local lidar odometry and low drift in global consistency.

Object detection algorithms require compact structures, reasonable probability interpretability, and strong detection ability for small targets. However, mainstream second-order object detectors lack reasonable probability interpretability, have structural redundancy, and cannot fully utilize information from each branch of the first stage. Non-local attention can improve sensitivity to small targets, but most of them are limited to a single scale. To address these issues, we propose PNANet, a two-stage object detector with a probability interpretable framework. We propose a robust proposal generator as the first stage of the network and use cascade RCNN as the second stage. We also propose a pyramid non-local attention module that breaks the scale constraint and improves overall performance, especially in small target detection. Our algorithm can be used for instance segmentation after adding a simple segmentation head. Testing on COCO and Pascal VOC datasets as well as practical applications demonstrated good results in both object detection and instance segmentation tasks.