Learned trafficability for UGVs: inferring geometry from imagery
Bibliographic record
Abstract
Unmanned ground vehicles (UGV) operating in outdoor environments must traverse unstructured terrain. This terrain is diverse in nature and contains natural obstacles such as rocks, brushes, berms, and low lying wet areas. Outdoor terrain is not static as it varies on a seasonal basis due to the life cycle associated with natural vegetation. Additionally, outdoor terrain may change appearance due to variations in lighting conditions that result from the Sun's relative position and from weather conditions such as clouds, fog or rain. This environmental diversity has long caused researchers considerable grief, as developing a classical terrain classification algorithm has proven to be a very difficult if not an impossible task. Researchers have skirted this problem by relying upon ranging sensors and constructing 2 ½D or, more recently, 3D world representations. Although geometrical representations have been used extensively, the low data rates associated with laser rangefinders, the unreliability of stereo vision, and the interaction between geometry and orientation estimation errors have limited the lookahead distance, thereby reducing the maximum attainable vehicle speeds. Learning from experience, in a more human like manner, promises to reduce or alleviate many of the issues posed by unstructured outdoor terrain. Defence R&D Canada (DRDC) "Learned Trafficability" program researches learning from experience. The paper presents DRDC's progress in extending a 2 ½D world representation using vision and learning from experience.
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 | 0.000 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".