Motivation Making the world more accessible for differently abled people and empowering them easily move about and interact with the environment. Also, giving everyone else a more novel way to interface with technology.
How it works The user wears Cerebro on his head. It picks up signals such as EMG, Muscle moments, alpha waves etc continuously. It also picks up an approximate location of the user in an environment. When the user goes near an IOT-enabled device, Cerebro presents him with a unique of inputs to the device. For example, if the user approaches a door, cerebro gives the user to open or close the door, or if the user approaches a retracting ramp, it allows him to open or close the ramp. The user can seamlessly transition between using different IOT based devices.
How we made it For this prototype, we used the OpenBCI kit, which included a sensor harness that sits on top of a person's head and a bunch of sensors that read muscle moments. We also used Bluetooth Low Energy Beacon to detect proximities to devices. All of this was controlled using the Qualcomm DragonBoard 410C as the central controller. The Brain Sensor Harness continuously emits readings wirelessly to the central controller. The central controller currently uses heuristics to figure out if the user made an actual input. The BLE proximity sensor lets the controller know where the user is and which device he is closest to. With this information, the controller does the appropriate action based on the context (Such as opening or closing the door).
Impact and Future work We believe Cerebro has a real opportunity to help people who are differently abled and cannot interact with the environment as it is currently designed. The quick adoption of IOT, which makes elements of the environment “smart” also helps Cerebro be immediately usefull. The current limitation of systems like OpenBCI is that its very hard to measure many actual inputs at any given point in time due to heavy noise in the signal. Its “degree of freedom” is severely limited because of this. Cerebro allows multiplexing multiple types of inputs based on context, allowing for greater freedom. Its possible to use machine learning to make Cerebro adapt to variations among different people and continuously learn.