Abstract: Provided are systems, methods, and media for assisting incontinent subjects. An example method includes receiving sensor readings from a sensor to obtain pressure readings from a bladder or bowels of a subject. A machine learning algorithm is executed to determine when incontinence of the subject is expected to occur. The machine learning algorithm establishes threshold(s) at which to notify an autonomous system, the subject, or a caretaker, in which are established based on a correlation between the received sensor readings and a frequency of undeterred accidents of the subject. An electronic warning is transmitted to the autonomous system, the subject, or the caretaker of the subject, indicating that incontinence is expected to occur in response to detecting at least one sensor reading that exceeds a threshold. Feedback is obtained and the threshold is adjusted via the machine learning algorithm based on the received feedback.