基于多模感知数据的驾驶员房颤检测方法及系统研究

Driver Atrial Fibrillation Detection Method and System Based on Multimode Perception Data

设计了一种针对驾驶环境的房颤自动检测方向盘。该方案利用安装在方向盘上的传感器对驾驶员的心率数据进行采集,将数据上传至云平台后,云平台利用基于深度互学习和判别滤波群算法的神经网络对心率信号进行分类识别,并将识别结果反馈至车辆终端,如果检测到司机有房颤现象,车辆终端则会及时发出警报,协同制动或者提供医疗介入。所提算法在真实数据集上表现优异,能够精确分别出正常心率和房颤。通过在模拟驾驶环境中的实验,所设计的方案可以在不对驾驶员操作造成任何影响的情况下,低延时、高准确性地对驾驶员心率状态进行判断。

An automatic atrial fibrillation detection steering wheel for driving environment is designed.The scheme uses sensors installed on the steering wheel to collect the driver’s heart rate data.After the data is uploaded to the cloud platform,the neural network based on deep mutual learning and discriminant filter group algorithm is used to classify and recognize the heart rate signal,and the recognition results are fed back to the vehicle terminal.If the driver has atrial fibrillation,the vehicle terminal will issue an alarm in time,co-bra-king or providing medical intervention.The proposed algorithm performs well on real data sets and can accurately distinguish normal heart rate from atrial fibrillation.Through the experiment in the simulated driving environment,the designed scheme can judge the driver’s heart rate state with low delay and high accuracy without causing any influence on the driver’s operation.