Providing healthcare via autonomous, self-learning, and self-evolutionary processes
Inventors
Lee, June • Shaffer, Gary L. • Berti, Thomas J • Li, Yan
Assignees
National Society Of Medical Scientists Inc • Planned Systems International Inc
Publication Number
US-12283378-B2
Publication Date
2025-04-22
Expiration Date
2042-06-09
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
System, methods, and other embodiments described herein relate to autonomous assessment and treatment of a patient. In one embodiment, a method includes, responsive to acquiring, in a managing device, sensor data characterizing a condition of the patient, determining a diagnosis for the condition according to a correlation of the sensor data with a subset of markers. The method includes selecting, using a treatment model, a treatment algorithm from a set of treatment algorithms for performing therapeutic delivery using a robotic device. The method includes causing the robotic device to perform the therapeutic delivery according to the treatment algorithm selected by the treatment model.
Core Innovation
The invention relates to a system and methods for autonomous assessment and treatment of patients using robotic devices, managed by a central management system equipped with machine learning models. The system acquires sensor data characterizing a patient's condition, determines diagnoses based on correlations with markers, selects suitable monitoring and treatment algorithms via learning models, and causes robotic devices to autonomously perform therapeutic delivery accordingly.
The core problem being addressed is the complexity and rising costs of healthcare, including the increasing scarcity of caregivers, potential errors due to overworked personnel, and the need for improved accuracy and quality in patient care. The invention seeks to alleviate these issues by implementing autonomous, self-learning robotic systems that monitor, diagnose, and deliver treatments to improve healthcare efficiency and outcomes without manual intervention.
The system employs a management system that orchestrates multiple autonomous robots, which may include specialized or multifunctional medical robots, to perform monitoring and treatment tasks. The management system utilizes reinforcement learning to select and adapt algorithms for clinical screening, diagnosis, and therapeutic delivery, dynamically adjusting treatments and monitoring based on patient feedback. The robotic devices and management system thus synergize to create an evolving, intelligent healthcare delivery process that continually improves through iterative data acquisition and learning.
Claims Coverage
The patent claims cover a management system and associated methods for autonomously monitoring, diagnosing, and treating patients using robotic devices with machine-learning models that select and adapt monitoring and treatment algorithms.
System for autonomous patient assessment and treatment using machine-learning models
The management system acquires sensor data characterizing a patient's condition, determines a diagnosis by correlating the sensor data with a subset of markers, selects a monitoring algorithm via a monitoring model to focus on the patient’s condition, selects a treatment algorithm and robotic device from a set based on the diagnosis and sensor data using a treatment model, and causes the robotic device to autonomously perform therapeutic delivery accordingly.
Adaptive treatment algorithm selection based on patient response
The treatment model applies to the diagnosis and patient data to predict which treatment algorithms will resolve the condition and learns which treatments to apply based on patient response and feedback.
Robotic device selection and treatment adaptation
Selection of the robotic device is based on availability of robotic equipment; the treatment algorithm adapts dynamically according to feedback observed in the patient; treatment algorithms include pharmaceutical and surgical procedures; markers correspond to physiological attributes correlating with different conditions.
Dynamic monitoring algorithm selection
The monitoring algorithm is selected by considering available sensors of multiple robotic devices according to the patient’s condition and dynamically adapted based on changes over iterative monitoring cycles.
Iterative diagnosis and feedback-based treatment adaptation
The system receives sensor data iteratively from robotic triage devices, updates diagnosis with observed patient changes, and receives feedback from robotic devices about treatment performance to adapt subsequent diagnosis and treatment selections.
Reinforcement learning-based training of models
The treatment and monitoring models are trained using reinforcement learning policies based on feedback during robotic autonomous treatment, continually improving their accuracy and effectiveness.
The claims collectively describe an autonomous healthcare management system that dynamically acquires patient data, diagnoses conditions, selects algorithms and robots for monitoring and treatment, and uses machine learning, particularly reinforcement learning, to adapt and improve clinical decisions and therapeutic delivery performed by robotic devices.
Stated Advantages
Improved quality of healthcare through consistent, precise, and accurate autonomous robotic treatment delivery.
Reduced medical errors by leveraging intelligent autonomous robots that supplement or replace overextended caregivers.
Enhanced adherence to treatment protocols and faster, more accurate clinical decision making.
Cost reduction in healthcare by reducing dependency on human caregivers and enabling automated diagnosis and treatment.
Continuous self-learning and evolution of algorithms and models increase treatment effectiveness and efficiency over time through feedback.
Improved access and convenience for patients through autonomous robotic delivery of healthcare services.
Documented Applications
Autonomous healthcare delivery and assessment including triage, monitoring, diagnosis and treatment of patients.
Use of multifunctional and specialized robotic devices for surgical procedures, pharmaceutical delivery, imaging, laboratory tests, preventive medicine, and rehabilitation.
Implementation in cloud-based healthcare environments managing swarms or teams of robots cooperating to provide tailored patient care.
Application in managing patients with multiple traumatic injuries requiring complex monitoring and therapeutic interventions.
Learning and evolving therapeutic protocols for various medical conditions, including hemorrhagic shock and associated coagulopathies, through continuous robotic feedback and reinforcement learning.
Interested in licensing this patent?