US 12,437,872 B2
Autonomous medical screening and recognition robots, systems and method of identifying a disease, condition, or injury
June Lee, North Bethesda, MD (US); Thomas J Berti, Lusby, MD (US); and Yan Li, Bethesda, MD (US)
Assigned to Planned Systems International, Inc., Arlington, VA (US); and National Society of Medical Scientists, Bethesda, MD (US)
Filed by Planned Systems International, Inc., Arlington, VA (US)
Filed on Sep. 30, 2022, as Appl. No. 17/957,694.
Application 17/957,694 is a continuation in part of application No. 17/894,764, filed on Aug. 24, 2022.
Application 17/957,694 is a continuation in part of application No. 17/870,621, filed on Jul. 21, 2022.
Application 17/957,694 is a continuation in part of application No. 17/836,264, filed on Jun. 9, 2022, granted, now 12,283,378.
Claims priority of provisional application 63/390,816, filed on Jul. 20, 2022.
Prior Publication US 2023/0402179 A1, Dec. 14, 2023
Int. Cl. G16H 50/20 (2018.01); G16H 10/60 (2018.01); G16H 20/00 (2018.01); G16H 40/67 (2018.01)
CPC G16H 50/20 (2018.01) [G16H 10/60 (2018.01); G16H 20/00 (2018.01); G16H 40/67 (2018.01)] 20 Claims
OG exemplary drawing
 
1. A screening system for autonomous screening and diagnosis using a screening robot, comprising:
one or more processors; and
a memory communicably coupled to the one or more processors and storing instructions that, when executed by the one or more processors, cause the one or more processors to:
generate, using a learning model, a diagnosis for a patient according to health information acquired from at least the screening robot, the health information including sensor data about the patient and perceptions derived from the sensor data, wherein the instructions to generate the diagnosis using the learning model include instructions to generate the diagnosis as an initial probabilistic diagnosis, represented as a Q-value distribution, by a Deep Q-Learning (DOL) agent trained on multimodal medical data, including genomic profiles, wherein the DOL agent, leveraging predictive analytics, estimates future health risks and potential medical conditions to derive the initial probabilistic diagnosis;
responsive to determining that the diagnosis is incomplete according to the Q-value distribution, generate a request for additional information that the DOL agent identifies based on action response pairs associated with unobserved evidence for the patient, wherein the instructions to generate the request includes selecting, from a set of available robots including the screening robot, a respective robot of the set of available robots to perform dynamic clinical screening and acquire the additional information;
update the diagnosis according to the additional information; and
provide the diagnosis to facilitate treatment of the patient by controlling at least one treatment robot to perform a therapy on the patient according to the diagnosis.