|
Briefing
|
- Brain-Controlled Robot Correction – Researchers from MIT and Boston University developed way for humans to correct robots in their tasks through mind, letting them intervene when robots make mistakes
- How It Works – User wearing electroencephalography (EEG) cap, which records brain activity, picks up “error-related potentials or brain signals generated when humans sense mistake, allowing robots to course correct actions in object sorting task
- Secondary Mechanism – Can also detect secondary signals from humans when robots fail to respond to initial correction, with plans to improve accuracy to over 90% over time
- More Complex Tasks – Plans to move robot from selecting between two choices to more complex multiple choice tasks
- Applications – Include brain-controlled robots and prosthesis, and self-driving cars sensitive to passenger’s fears or brain signals
|
|
Accelerator
|
|
|
Sector
|
Information Technology
|
|
Organization
|
Boston University, MIT Computer Science and Artificial Intelligence Laboratory (CSAIL)
|
|
Source
|
|
|
Original Publication Date
|
March 6, 2017
|
