Brain-Controlled Wheelchair on Windows Pc

App was created with the goal of creating a way to operate the miniature wheelchair that is as simple and straightforward for the user while staying within the boundaries of the very limited controls the EEG Device affords us. The researchers came up with a sequential operation loop composed of four different modes, each representing a state of the wheelchair. The modes are as follows: standby, command, focus, and running. After both the EEG headset and the Raspberry Pi establish connection with the Android application, the Android application begins fetching the signal quality value, which can
be not detected, poor, medium, or good. The signal quality will be not detected when the user is not wearing the headset, poor if almost no contact is made by the forehead skin with the dry sensor, medium if partial contact is made by the forehead skin with the dry sensor, and good if the dry sensor makes firm contact with the forehead. As can be seen in Chapter 5, the signal quality has a value from 0-255 with 0 being the best and 255 being the worst. The range of values that each signal quality value is based upon has not been revealed by Neurosky. As an added safety precaution, when the signal quality value is not good, a stop command will be sent to the miniature wheelchair, preventing any unwanted motion. Once the signal quality value turns into good, the Android application begins listening for any incoming force blink data from headset. At this point, normal blinks or blinks whose blink strength values are below the threshold value of 90 are discarded. When a force blink or a blink whose blink strength value is above the threshold value of 50 is detected, the Android application begins cycling direction values – forward, left, and right – for 10 seconds with a 2-second interval in between changing the direction value. This 10-second direction- cycle window is known as command mode. During command mode, the Android application listens for two consecutive blinks, otherwise known as a double blink event, from the user. When it detects a blink event, the cycling of directions stops and whatever direction is shown in the cycle at the moment of the double blink event will become the chosen direction. For blinks to be considered consecutive, the time elapsed between two blink events must be equal to or less than 400 milliseconds. When a direction has been chosen, the Android applications shifts to focus mode where it starts listening to any incoming attention data from the EEG headset. Attention values are outputted by the headset once every 1 second and once it goes to 50 or more, the Android application switches to running mode where it sends a command to the Raspberry Pi based on the direction chosen earlier. Each direction has respective command that will be transmitted to and interpreted by the Raspberry Pi residing on the miniature wheelchair. Outside of focus mode, the attention listener process is set to null to reduce the amount of work the Android application has to do simultaneously. Similar to command mode, the user exits running mode by blinking consecutively to go back to standby mode. From then on, the whole operation loop is repeated should the user want to move the miniature wheelchair once again. The speed is kept at a constant throughout operation when the miniature wheelchair is running. This is due to accuracy and control-issues that are innate to the brainwave detection in the NeuroSky’s EEG headset. Because of this, the constant speed can also be thought of as a safety feature for the user.

Brain-Controlled Wheelchair App