Enhancement of impact synchronous modal analysis with brain-computer interface

Novel Impact Synchronous Modal Analysis (ISMA) suffers from inefficient operation. The Automated Phase Controlled Impact Device (APCID), a fully automated device, was developed to efficiently perform ISMA, however, the actuator, support structure and power supply of the APCID make it large, heavy, and unsuitable for commercial applications. The APCID can be replaced with manual operation while still using its controls but by nature there is randomness in human behaviour, which can greatly reduce the effectiveness of the APCID control scheme. A smart semi-automated device for imparting impacts is developed in this study, which uses Brain-Computer Interface (BCI) to predict impact time prior to impact. Brainwaves are measured using a portable, wireless and low-cost Electroencephalogram (EEG) device. Using brainwaves, a Machine Learning (ML) model is developed to predict the impact time. The ML model gave a Mean Absolute Percentage Error (MAPE) of 7.5% and 8% in evaluation (offline testing) and in real-time testing, respectively, while predicting impact time prior to impact using brainwaves. When integrated with the control of APCID to perform ISMA, the ML model gave a MAPE of 8.3% in real-time ISMA while predicting impact time prior to impact and adjusting the APCID control for the upcoming impact accordingly. To demonstrate the effectiveness of the EEG ML model in performing ISMA, modal testing was performed at 2 different operating speeds. The study concludes by comparing the developed ISMA method with other ISMA methods. The BCI based device developed in this study for performing ISMA outranks other ISMA methods due to its performance, efficiency and practicality.