Search

The trend in aviation automation demands more mental workload of pilots, in addition to their routine manoeuvring task. For fighter aircraft pilots, the mental workload multifold due to read-back or hear-back with precise weapon handling in combat scenarios. Assessing the pilot cognitive workload is an important aspect, as it influences the pilot performance, and this is an error intolerance environment. This study aims to assess fighter aircraft Pilot’s Cognitive WorkLoad (PCWL) and attention when they exposed to a dynamic workload environment; by monitoring the neuronal activities. Here the spectral Electroencephalographic (EEG) features were extracted to assess the dynamic workload (normal, moderate, high, and very high workload), and attention monitored by National Aeronautics and Space Administration-Task Load Index (NASA-TLX) and it serves as a validation for cognitive findings. Also, brain source localization obtained from EEG inverse calculation technique, Standardized Low-Resolution brain Electromagnetic Tomography (sLORETA) performed to understand the coherence of workload and neuronal activity engagement concerning brain lobes. Statistical significance (p

EEG signal is one of the most complex signals having the lowest amplitude which makes it challenging for analysis in real-time. The different waveforms like alpha, beta, theta and delta were studied and selected features were related with the consciousness levels. The consciousness levels detection is useful for estimating the subjects’ performance in certain selected tasks which requires high alertness. This estimation was performed by analyzing signal properties of the EEG using features extracted through discrete wavelet transform with a moving window of 10 seconds with 90% overlap. The EEG signal is decomposed in to wavelets and the average energy and power of the coefficients related to the EEG bands is taken as the features. The data is collected from standard EEG machine from the volunteers as per the protocol. C3 and C4 locations (unipolar) of the standard 10-20 electrode system were selected. The central region of the brain is most optimal location for the consciousness levels detection. The estimation of the data using Discrete Wavelet Transform (DWT) energy, power features provided better accuracy when the central regions were chosen. An accuracy of 99% was achieved when the algorithm was implemented using a classifier based on linear kernel support vector machines (SVM).