Table 2.

Brain-computer interface (BCI) technologies for adaptive assistive mobility devices.

Brain signals and auxiliary sensors	Classifier for feature extraction	Output command	Contributions	Drawbacks
P300 (laser scanner) [74]	Stepwise linear discriminant analysis	A predefined set of locations and stops	High accuracy, no training required, and autonomous navigation after successful selection	Low information transfer rate, predefined paths, limited testing scenarios, and possible fatigue after long focus period of the eye on the target stimulus
P300 (odometer, barcode scanner, and a proximity sensor) [38]	Support vector machine	A predefined set of locations and stops	Same as Rebsamen et al [38]	Same as Rebsamen et al [38] and a modified environment requires an update of the guiding path
MIa-based mu rhythm and the P300 [39]	One versus the rest common spatial patterns transformation matrix	Left, right, accelerate, and decelerate	Improved performance	Limited testing scenarios and possible fatigue after long focus period of the eye on the target stimulus
MI-based BCI (10 sonar sensors and 2 webcams) [41]	Gaussian classifier	Left, right, and keep moving forward	Spontaneous and shared control	Limited testing scenarios, requires extensive training, and limited classes (typically three)
Steady-state visual evoked potentials (camera and adaptive fuzzy controller) [42]	Frequency recognition algorithm based on multivariable synchronization index	Left, right, upwards, and downwards	Teleoperation control of an exoskeleton using a brain-machine interface	Possible fatigue after a long focus period of the eye on a target stimulus and a significant reduction in recognition accuracy for inexperienced subjects
Steady-state somatosensory evoked potential [40]	Regularized linear discriminant analysis	Turn left, turn right, and move forward	Spontaneous, first of its kind, and addressed the possible fatigue after a long focus period of the eye on the target stimulus	Only healthy subjects were used, with limited testing scenarios (two)

^aMI: motor imagery.