Constrained, Curiosity-driven Trajectory Optimization for Learned Quadrotor Control
Bibliographic record
Abstract
Quadrotors have become widely used robotic platforms in recent years, but developing controllers can be challenging for difficult-to-model scenarios such as wind or ground effect. Simultaneously, model-based reinforcement learning (MBRL) has been applied to many robot platforms due to the ability to quickly learn a dynamics model and use the model for planning and control via data-driven methods. MBRL has been augmented with probabilistic modelling in order to address model bias concerns. Recent methods have used the variance from model predictions to drive exploration during learning, inspired by curiosity. In this thesis, we show that directly applying these methods to control quadrotors via thrust and rotation rate inputs can lead to stability concerns. We propose using ideas from safety-critical control to alleviate these stability concerns. Our method applies control barrier function (CBF) constraints during planning using an augmented Lagrangian approach. The resulting algorithm restricts exploration to areas of the state space where predictive model uncertainty is within tolerable limits, so predictions are more trustworthy. However, we identify some practical issues which affect optimization feasibility. We propose additional measures to alleviate these issues to enable stable, curiosity-driven exploration for quadrotors.
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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".