| This Paper:Browse 106 Download 0 |
码上扫一扫！ |
| Online control for pressure regulation of oxygen mask based on neural network |
| LiganZhao1,QinglinSun1,HaoSun1,JinTao2,ZengqiangChen1 |
|
|
| (1 College of Artificial Intelligence, Nankai University, Tianjin 300350, China;2 Silo AI, Helsinki 00100, Finland) |
|
| 摘要: |
| The aviation oxygen mask, which has a small volume of less than 1L and strong air tightness, imposes extremely high
requirements on control performance of the oxygen regulator. Based on analyses of the operation principle of oxygen supply
system, the dynamic model is established through the combination of mechanism analysis and experimental data. Considering
that the traditional fixed-parameter controllers are difficult to meet the control requirements with changes in pulmonary
ventilation, this paper presents an online feedback controller based on neural network compensation (NNC), with connection
weights that can be updated without pre-training. Then mathematical simulations at different respiratory parameters, such
as respiratory rate, are performed to verify the superior lower inspiratory resistance of controller with NNC. In terms of
hardware, an embedded AI control platform is to complete the experimental verification. Furthermore, the work may have
downward compatibility to achieve stable oxygen supply in civil fields, such as medical ventilators, high-altitude expeditions. |
| 关键词: Neural network · Online control · Pressure regulation · AI platform |
| DOI：https://doi.org/10.1007/s11768-024-00222-w |
|
| 基金项目:This work was supported by the National Natural Science Foundation of China (Nos. 61973172, 62003177, 62003175 and 61973175). |
|
| Online control for pressure regulation of oxygen mask based on neural network |
| Ligan Zhao1,Qinglin Sun1,Hao Sun1,Jin Tao2,Zengqiang Chen1 |
| (1 College of Artificial Intelligence, Nankai University, Tianjin 300350, China;2 Silo AI, Helsinki 00100, Finland) |
| Abstract: |
| The aviation oxygen mask, which has a small volume of less than 1L and strong air tightness, imposes extremely high
requirements on control performance of the oxygen regulator. Based on analyses of the operation principle of oxygen supply
system, the dynamic model is established through the combination of mechanism analysis and experimental data. Considering
that the traditional fixed-parameter controllers are difficult to meet the control requirements with changes in pulmonary
ventilation, this paper presents an online feedback controller based on neural network compensation (NNC), with connection
weights that can be updated without pre-training. Then mathematical simulations at different respiratory parameters, such
as respiratory rate, are performed to verify the superior lower inspiratory resistance of controller with NNC. In terms of
hardware, an embedded AI control platform is to complete the experimental verification. Furthermore, the work may have
downward compatibility to achieve stable oxygen supply in civil fields, such as medical ventilators, high-altitude expeditions. |
| Key words: Neural network · Online control · Pressure regulation · AI platform |
