首页|Recent Findings in Robotics Described by Researchers from Brandenburg University of Technology (Advancing passive BCIs: a feasibility study of two temporal deri vative features and effect size-based feature selection in continuous online ... )
Recent Findings in Robotics Described by Researchers from Brandenburg University of Technology (Advancing passive BCIs: a feasibility study of two temporal deri vative features and effect size-based feature selection in continuous online ... )
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Data detailed on robotics have been pr esented. According to news reporting originating from Cottbus, Germany, by NewsR x correspondents, research stated, "The emerging integration of Brain- Computer I nterfaces (BCIs) in human-robot collaboration holds promise for dynamic adaptive interaction. The use of electroencephalogram (EEG)-measured error-related poten tials (ErrPs) for online error detection in assistive devices offers a practical method for improving the reliability of such devices." Our news correspondents obtained a quote from the research from Brandenburg Univ ersity of Technology: "However, continuous online error detection faces challeng es such as developing efficient and lightweight classification techniques for qu ick predictions, reducing false alarms from artifacts, and dealing with the non- stationarity of EEG signals. Further research is essential to address the comple xities of continuous classification in online sessions. With this study, we demo nstrated a comprehensive approach for continuous online EEG-based machine error detection, which emerged as the winner of a competition at the 32nd Internationa l Joint Conference on Artificial Intelligence. The competition consisted of two stages: an offline stage for model development using pre-recorded, labeled EEG d ata, and an online stage 3 months after the offline stage, where these models we re tested live on continuously streamed EEG data to detect errors in orthosis mo vements in real time. Our approach incorporates two temporalderivative features with an effect size-based feature selection technique for model training, toget her with a lightweight noise filtering method for online sessions without recali bration of the model. The model trained in the offline stage not only resulted i n a high average cross-validation accuracy of 89.9% across all par ticipants, but also demonstrated remarkable performance during the online sessio n 3 months after the initial data collection without further calibration, mainta ining a low overall false alarm rate of 1.7% and swift response ca pabilities. Our research makes two significant contributions to the field. First ly, it demonstrates the feasibility of integrating two temporal derivative featu res with an effect size-based feature selection strategy, particularly in online EEG-based BCIs."
Brandenburg University of TechnologyCo ttbusGermanyEuropeEmerging TechnologiesMachine LearningRobotRobotics
NETL
NSTL
