Co-Adaptive Delta-Assist: Distilled Tiny Transformer with Control-Barrier Safety for Real-Time Upper-Limb Lift Compensation
Bibliographic record
Abstract
Co-Adaptive Delta-Assist is introduced as an edgeresident controller that issues small delta-angle setpoints to commodity servos, in which an INT8 Tiny-Transformer is combined with control-barrier-function (CBF) safety. On microcontroller-class hardware, sub-millisecond inference is achieved at <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{0. 6 2\ m s}$</tex> with 128 kB memory. In a within-subject, counterbalanced study (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{N}=16$</tex>, age <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=28.3 \pm 7.1 \mathrm{y}, 56 \%$</tex> females; tasks: cup-lift, shelf-reach), the learned policy was found to outperform LPF- and Kalman-based assistance: tracking RMSE was reduced by 26 %, spectral-arc length was increased by 0.09, jerk was reduced by 38 %, and task time was shortened by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 4 \%}$</tex> (all <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p<.01$</tex>; large Cliff's <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\delta$</tex>). Session energy was decreased by 12 % relative to Kalman, while barrier margins (angle, velocity, current) remained positive and no adverse events were observed. A one-minute on-device personalization was shown to further improve smoothness/effort by 12 % without widening safety envelopes. These results indicate that commodity hardware together with an edge-resident distilled policy and CBF safety constitutes a practical path for clinic-to-home upper-limb lift assistance.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 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 teacher head, 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".