Optimising Automatic Calibration of Electric Muscle Stimulation
Authors: Graeme Gange, Jarrod Knibbe3759-3767
AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluated calibration on a simulated arm/electrode array system. We conducted 200 tests for each array/target configuration, different random seeds and permuted action numbering. The tests are conducted with a budget of 200 queries for each target. ... We report the number of queries evaluated during each test, and how successful we are at achieving target patterns. ... Following the simulated technical evaluation, we replicate our evaluation on the second author |
| Researcher Affiliation | Academia | Graeme Gange,1 Jarrod Knibbe2 1Monash University, Victoria, Australia 2The University of Melbourne, Victoria, Australia graeme.gange@monash.edu, jarrod.knibbe@unimelb.edu.au |
| Pseudocode | Yes | Algorithm 1 Pseudocode for Quick Xplain ... Algorithm 2 The aggregation phase... Algorithm 3 Given a candidate query... Algorithm 4 SEEK is called... Algorithm 5 Choosing a new subset to explore... |
| Open Source Code | Yes | Code is available at https://bitbucket.org/gkgange/stim-cal |
| Open Datasets | No | The paper uses a 'simulated arm/electrode array system' and a 'simplified response model' for its experiments, and does not provide concrete access information or refer to an established public dataset. |
| Dataset Splits | No | The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology). |
| Hardware Specification | Yes | Computational experiments were performed on an AMD Ryzen 5 3600 with 16Gb memory running Ubuntu 18.04. |
| Software Dependencies | No | The paper mentions using an 'LCG solver' and references 'geas (Gange et al. 2020)', but does not provide specific version numbers for any software components. |
| Experiment Setup | Yes | We calibrated simulated fixed arrays with 7 and 12 electrodes on each side... Each selected pair distributes 1 unit of energy uniformly along a line between its endpoints. A joint is considered flexed (or extended) if all corresponding muscles are receive at least 0.2 units of energy. ... We set a budget of 50 queries per target. We fix the stimulation parameters to 17V, 100Hz, and 300µs. We consider a joint actuated if it demonstrates > 30 angular change. |