Machine Learning Models for Multi Robot Secure Coordination in a Healthcare Facility
Title: Identifying Appropriate Machine Learning Models for Multi Robot Secure Coordination in a Healthcare Facility
Mentor: Dr. Debbie Perouli
Approach: Develop machine learning models on sensor datasets representing data captured through multiple robots in a care facility. Students will gain experience identifying the right data inputs, implementing machine learning models and recognizing critical data from an adversarial point of view.
Summary: The Bureau of Labor Statistics and the Institute of Medicine are among the organizations predicting a significant shortage of Registered Nurses (RNs) in the U.S. in the next decade and onwards. Introducing Robotic Caregivers (RCGs) in senior living facilities and hospitals are among the proposed solutions. Commercial autonomous service robots, such as Relay from Savioke, have already been adopted in such spaces for delivery of blood samples and other items. Successful use of multiple RCGs within a facility requires their secure coordination after processing real time sensor data at a constant rate. The student will investigate the type of data needed for multi-robot coordination and a solution space model that can be identified by a machine learning algorithm. The student will also identify data inputs most critical for the machine learning algorithm that an adversary would desire to manipulate. Mentor’s related work includes a research paper with undergraduate student  and the supervision of a master’s thesis.
Student Research Activities: The REU fellows will perform the following major tasks:
- Investigate the type of data needed for multi-robot coordination.
- Develop machine learning models on sensor datasets.
- Recognize critical data from an adversarial point of view.
- Evaluate the machine learning models using standard measures such as accuracy, precision and recall.
Student Background: Students need to have basic programming skills and introductory knowledge of networking.