01 OVERViEW

HUMAN WEIGHT-LOSS CONTROL SYSTEM (WLS)
-
Aimed to develop a Weight-Loss Control System (WLS) for obese patients to indicate personalized allowable food-mass intake with dynamic input of their weight and real-time digesta flow rate as a measure of dietary intake
-
Analyzed available experimental data and deduced the effect of each diet-variable and other factors on the weight of the body
-
Designed an experiment to be conducted on a subject belonging to the target population to study the bio-dynamics and other factors required to develop a statistical model for the system
-
Simulated and illustrated the performance of the devised controller in comparison with the benchmark system chosen

02 MY CONTRiBUTION
I was offered a research internship opportunity by Dr. P. V. Manivannan of IIT Madras to work on developing a Weight-Loss Control System (WLS) for the application of a Micro-Robotic Capsule, in collaboration with University of Nebraska-Lincoln (UNL), USA. I got an opportunity to visit the laboratory at UNL for about 10 weeks and learn the status-quo of the development of rest of the modules of this novel project led by Dr. Benjamin Terry
Taking this forward, I took this research internship to be wholly worked upon as my Master’s thesis project.
In this regard, I worked at UNL as a Visiting Graduate Research Assistant for a semester to closely work with the other graduate students of the lab involved in this project.
​
Considering the NIH's Body-Weight Simulator/Planner as a reference of open-loop control system, a closed-loop controller for active instructions to the user was aimed.
​
Thereby, I devised the control law for the Weight-Maintenance of a human with the real-time feedback obtained from a Gastro-Intestinal (GI) sensor, using the Model-based Predictive Control (MPC) approach.

03 ACHiEVEMENTS
The simulations run on the transfer function modeling digestive metabolism give the following plots. The plot shown in gives the simulated body weight of a human user for 50 days of the target duration to lose 2 kg for the weight-loss program suggested by the developed system, being called Diet Suggester (DS).



Benchmark System
Body Weight Simulator / Planner that was developed by Hall et al. and has open-access on web, which is an open-loop controller chosen as the benchmark system

Transfer Function
Linearization of weight change dynamics based on mathematical model of digestive metabolism of humans as per the Energy Intake and Expenditure

Adaptive System
Predicts the weight change of a given user for a target weight loss program and suggests the dietary intake as instruction, considering the real-time feedback from GI sensor
​Designed a controller for weight-loss or weight-maintenance system (WMS), a closed-loop control system that receives real-time dietary intake feedback from an in vivo Gastro-Intestinal sensor, and provides personalized diet intake suggestion to the user, by regular monitoring.
-
Reviewed existing literature for the correlations of intake food mass (with and without including water) and corresponding body weight of a human
-
Derived a transfer function model representing the dynamics of human digestive metabolism related to the body weight
-
Developed an open-loop controller of WMS for appropriate diet suggestion and then integrated another controller for closed-loop performance as well using Model-based Predictive Control method
04 LEARNiNGS
DATA PROCESSING
Pre-processing of the data sets of food mass intake and weight of the person at a scale of one day in the time-domain

SYSTEM IDENTIFICATION
Drawing Cross Correlation Function (CCF) of weigh difference and food mass intake, the non-linearity of dynamics was found

MODELLING OF THE SYSTEM
Programming of the linear transfer function of the human metabolism dynamics to derive its response for various types of inputs given

MPC
SIMULATION
Using the Model-Based Predictive Control approach, the transfer function was tested and illustrated for multiple cases