引導式測驗 · 硬 等級
LiDAR SLAM:同時定位與地圖製作測驗
檢查 LiDAR SLAM 座標系、掃描匹配、里程計漂移、閉環、運動失真、感測器融合和地圖可靠性。
問題 1 的 8
A moving LiDAR reports a point in its own coordinate frame. What is needed to place it in the map?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- A moving LiDAR reports a point in its own coordinate frame. What is needed to place it in the map?
- Only the measured range without a platform pose
- Only the transform between two adjacent scan points
- The sensor-to-body transform and estimated platform pose
- An IMU angular-rate sample without frame calibration
- How do scan matching and odometry contribute to SLAM?
- They provide an absolute geographic position for each beam
- They assume a complete perfect map is known initially
- They identify object classes without alignment
- They estimate relative movement from nearby observations
- Why can walls appear doubled after a sensor mount changes?
- The old sensor-to-body transform remains correct after remounting
- The LiDAR-to-body calibration may no longer match the physical transform
- A denser scan automatically duplicates every static wall
- Loop closure must always create two copies of every wall
- What does a valid loop closure add to the trajectory estimate?
- A constraint linking poses from a place revisited later
- Only a constraint between two consecutive scans
- A guarantee that all dynamic objects disappear
- An absolute external position measurement from the range sensor
- Why must a loop-closure candidate be verified in similar-looking corridors?
- Similar corridors always produce unique range patterns
- A false match affects only one unoptimized scan
- A false match can bend the optimized trajectory and map incorrectly
- The optimizer can correct wrong place matches without checking them
- A spinning LiDAR moves during one sweep. What error can deskewing reduce?
- All random range noise in the laser measurements
- Warping caused by points being collected at different sensor poses
- The need to calibrate the LiDAR-to-body transform
- Long-term drift caused by an incorrect loop closure
- Why might passing pedestrians be filtered from a static map?
- Their temporary returns can be mistaken for persistent structure
- Moving returns are always permanent map landmarks
- A static map is more accurate when it stores every passerby
- Pedestrian motion supplies an absolute world-frame origin
- Which observation alone gives the weakest evidence that a LiDAR SLAM trajectory is globally correct?
- Trajectory error against a suitable reference
- A validated loop-closure consistency check
- Measured alignment across a representative route
- A visually smooth local point cloud