Bibliographic record
Abstract
LiDAR Simultaneous Localization and Mapping (SLAM) technologies, which are the foundational technology of autonomous driving, have attracted large interest recently and been a significant research field. The performance of existing State-Of-The-Art LiDAR SLAM systems has been proven to produce accurate odometry estimation on autonomous driving datasets. These datasets are usually collected by vehicles equipped with various sensors under favorable weather conditions. However, challenging weather conditions such as rain and snow are still great obstacles because the rainfalls or snowflakes which are not static will cause noisy points for LiDAR perception and the assumption that the surrounding environment is static will be broken. Specifically, adverse weather will introduce noisy points which have physical structures and could be detected by LiDAR. Meanwhile, these noisy points would tightly surround the LiDAR and block other objects. This will lead to serious deficiencies in environmental structures and introduce more difficulties to pose estimation and loop closure, finally increasing the error of pose estimation and reducing the accuracy of LiDAR SLAM algorithms. Considering that noisy points usually lack the inherent structures exhibited in clean points, we propose a novel denoising framework for point clouds generated from lidar sensors that eliminate stochastic noisy points in a down sampling and super resolution manner to address this issue. Specifically, we first investigate to which degree the performance of the State-Of-The-Art lidar SLAM approaches will decrease when exposed to adverse weather conditions and then implement the denoising framework by combining the DS module with SR module which is based on the U-Net and trained under certain super-solution datasets. The accuracy and robustness of our framework were validated on the Oxford RobotCar dataset and the Canadian Adverse Driving Conditions dataset.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".