May 30, 2021

Paper Reading Part1

Vision Localization

  1. Monocular Localization in HD Maps by Combining Semantic Segmentation and Distance Transform (IROS 2020, KIT)
  2. Road Mapping and Localization using Sparse Semantic Visual Features (ICRA 2021, Alibaba)
  3. Compact 3D Map-Based Monocular Localization Using Semantic Edge Alignment (IROS 2021, Alibaba)
  4. Monocular Localization with Vector HD Map (MLVHM): A Low-Cost Method for Commercial IVs (Sensors, 2020)
  5. Long-Term Urban Vehicle Localization Using Pole Landmarks Extracted from 3-D Lidar Scans (2019)
  6. DA4AD: End-to-End Deep Attention-based Visual Localization for Autonomous Driving (ECCV 2020)
  7. Exploiting Sparse Semantic HD Maps for Self-Driving Vehicle Localization (IROS 2019, Uber.ATG)
  8. Lane Endpoint Detection and Position Accuracy Evaluation for Sensor Fusion-Based Vehicle Localization on Highways (Korean, 2018, sensors)
  9. Visual Semantic Localization based on HD Map for Autonomous Vehicles in Urban Scenarios (ICRA 2021, Huawei)
  10. HDMI-Loc:Exploiting high definition map image for precise localization vis bitwise particle filter (IROS 2020)
  11. DT-Loc: Monocular Visual Localization on HD Vector Map Using Distance Transforms of 2D Semantic Detections (IROS 2021)

Crowdsourced Mapping

  1. Real-Time HD Map Change Detection for Crowdsourcing Update Based on Mid-to-High-End Sensors (sensors 2021)
  2. CarMap: Fast 3D Feature Map Updates for Automobiles

  1. nvidia drive mapping ecosystem (material in files folder)

  2. here edge live mapping (material in files folder)

Lidar Localization

  1. Tightly-Coupled Multi-Sensor Fusion for Localization with LiDAR Feature Maps (tusimple, ICRA 2021) (combines the geometric features of lidar with normal distribution features, and fuses information such as IMU and wheel encoder)

  2. Direct LiDAR Odometry: Fast Localization with Dense Point Clouds (a novel submapping method, keyframe association, FAST-GICP + NANO-FLANN)

  3. CoFi: Coarse-to-Fine ICP for LiDAR Localization in an Efficient Long-lasting Point Cloud Map

  4. Fast and Robust Registration of Partially Overlapping Point Clouds

  1. Patchwork++: Fast and Robust Ground Segmentation Solving Partial Under-Segmentation Using 3D Point Cloud

  2. OpenStreetMap-based LiDAR Global Localization in Urban Environment without a Prior LiDAR Map generates descriptors for the lidar scan and the OSM map separately, then performs matching-based localization

SLAM

  1. CamVox: A Low-cost and Accurate Lidar-assisted Visual SLAM System (2020, SUST)
  2. TIMA SLAM: Tracking Independently and Mapping Altogether for an Uncalibrated Multi-Camera System (Sensors 2021)
  3. Structure-from-Motion Revisited (CVPR 2016)
  4. OV2SLAM : A Fully Online and Versatile Visual SLAM for Real-Time Applications (ICRA 2021)
  5. Generalizing to the Open World: Deep Visual Odometry with Online Adaptation (CVPR 2021)
  6. LVI-SAM: Tightly-coupled Lidar-Visual-Inertial Odometry via Smoothing and Mapping (ICRA 2021)
  7. LIO-SAM: Tightly-coupled Lidar Inertial Odometry via Smoothing and Mapping (IROS 2020)
  8. RoadMap: A Light-Weight Semantic Map for Visual Localization towards Autonomous Driving(ICRA 2021)
  9. PicoVO: A Lightweight RGB-D Visual Odometry Targeting Resource-Constrained IoT Devices (ICRA 2021) (edge based vo use edges as illumination-robust features and track them through image alignment upon Distance Transform)
  10. Simple But Effective Redundant Odometry for Autonomous Vehicles (ICRA 2021)
  11. AVP-SLAM: Semantic Visual Mapping and Localization for Autonomous Vehicles in the Paarking Lot (IROS 2020)
  12. DynPL-SVO: A New Method Using Point and Line Features for Stereo Visual Odometry in Dynamic Scenes

  1. Lightweight Semantic Mesh Mapping for Autonomous Vehicles (ICRA 2021)
  2. Submap-Based SLAM for Road Markings (kit, 2017)
  3. SA-LOAM: Semantic-aided LiDAR SLAM with Loop Closure
  4. EDPLVO: Efficient Direct Point-Line Visual Odometry

  1. MetroLoc: Metro Vehicle Mapping and Localization with LiDAR-Camera-Inertial Integration

  2. A decentralized framework for simultaneous calibration, localization and mapping with multiple LiDARs

A unified state; multiple agents each run their own EKF and communicate state over the network

  1. A General Framework for Lifelong Localization and Mapping in Changing Environment

Uses a Chow-Liu spanning tree to sparsify the pose graph (the marginalization involved in removing a submap causes dense fill-in)

  1. DSOL: A Fast Direct Sparse Odometry Scheme

  2. Vision-based Large-scale 3D Semantic Mapping for Autonomous Driving Applications

  3. LOCUS 2.0: Robust and Computationally Efficient Lidar Odometry for Real-Time 3D Mapping

  4. Nonlinear Optimization of Multimodal 2D Map Alignment with Application to Prior Knowledge Transfer (non-rigid alignment, distance transform, gradient map)

  5. Large-Scale Scene 3D Reconstruction for Unmanned Systems (Prof. Shen) download file here

Calibration

  1. Online Camera-LiDAR Calibration with Sensor Semantic Information (IROS 2020, Pony.ai)
  2. A Novel AVM Calibration Method Using Unaligned Square Calibration Boards (Sensors, 2021)
  3. Automatic Calibration of an Around View Monitor System Exploiting Lane Markings (Sensors, 2018)
  4. LiDAR-Camera Calibration using 3D-3D Point correspondences (2017)
  5. Online camera pose optimization for the surround-view system (Tongji, 2019, zhanglin)

BEV Perception

  1. Lift, Splat, Shoot: Encoding Images from Arbitrary Camera Rigs by Implicitly Unprojecting to 3D (ECCV 2020)
  2. Orthographic feature transform for monocular 3D object detection (2019 BMVC)
  3. BEVSegFormer: Bird’s Eye View Semantic Segmentation From Arbitrary Camera Rigs
  4. VectorMapNet: End-to-end VectorizedHD Map Learning
  5. Associative Embedding: End-to-End Learning for Joint Detection and Grouping

LocalMap

  1. Augmenting Lane Perception and Topology Understanding with Standard Definition Navigation Maps (2022 NVIDIA)
  2. Traffic Flow-Based Crowdsourced Mapping in Complex Urban Scenario (qintong 2023)
  3. Fast map matching, an algorithm integrating hidden Markov model with precomputation
  4. VectorMapNet: End-to-end Vectorized HD Map Learning
  5. MapLite 2.0: Online HD Map Inference Using a Prior SD Map

Math in SLAM

  1. a micro Lie theory for state estimation in robotics (2020)
  2. MegBA: A High-Performance and Distributed Library for Large-Scale Bundle Adjustment https://github.com/MegviiRobot/MegBA

Parking Slot Detection

  1. Vacant parking slot detection in the around view image based on deep learning

Comments