Feedback control for optimizing human wellness

Abstract

For most people, decisions about “wellness” are made by default. Without specific evidence-based guidance, most find it difficult to make specific long-term commitment to achieving and maintaining wellness. Proposed here is an approach to filling the information gap through “Informed spectatoring” using appropriate instrumentation and advice: a type of “feedback control”. Human beings are complex systems having many interoperable subsystems. Neither these subsystems nor their interactions are completely understood. Further, while there are broad principles describing these subsystems and their interactions, the specifics are personal. Taken together, these factors make optimization of the human being by achieving and maintaining Wellness a challenging problem. We describe a cloud-based system intended to provide actionable recommendations to users for improving their wellness. The recommendations are personalized for the user demographic and history. This is accomplished in three steps:1)Assessment: a knowledge-based expert system ingests data from the user, and positions them on an objective, numeric multi-dimensional wellness scale in the general areas of Psychology, Physiology, and Finanaces. This assessment of wellness is informed by the user demographic, history, and a wearable monitoring device.2)Projection: A bank of feedback controllers (pid controllers) is used to estimate user future wellness states by projecting the user wellness state forward in time. This is a prospective, rather than merely retrospective analysis. It is focused on where the user will be rather than merely where the user has been.3)Recommendation: Based upon the assessment of user wellness, and the forward projection of the user wellness, a knowledge-based expert system selects a few areas of wellness that can be addressed in a unified way by simple user actions. An innovation is that the expert system makes these recommendations by automatically crafting a report in colloquial prose that reads like it was written by a human. An operational, cloud-based prototype has been built and tested on simulated data. We describe an upcoming human trial using members of the U.S. Olympic Team as subjects.