This thesis project is titled "Control System Design of Intermittent Pneumatic Compression (IPC) Device with PID Method and Safe Experimentation Dynamics (SED) Optimization for Venous Insufficiency Rehabilitation".
Files attached are MATLAB code scripts including:
- SED_code_2_LS.mlx: main live script for PID tuning and optimization algorithm
- objectiveFunction.mlx: separate function to store objective function with rise time, settling time, and objective function parameters
- h.mlx: separate sub function of SED algorithm
Simulink files include:
- fullmodel_IPC_A_bintang_ps_regulated.slx: full IPC device model with 3 main parts (pneumatic circuit, bladder, contact characterization) connected to PID blocks
Requires:
- MATLAB R2020a
- Simulink 10.0
The main objective of the project is to obtain SED optimization results of a PID-controlled IPC device in simulation. The model simulates the inflating-deflating action of a pump and bladder attached to limbs with a theorized stiffness coefficient. Objective function values summarize the overall performance of tuned PID parameters, with a focus on minimizing rise time (Tr), settling time (Ts), and overshoot (Ovt). Results from the optimization are compared to results from manual tuning. The optimized model was also used to simulate different inflating-deflating cycle times.
Below is the Simulink model of a single-bladder system:
Comparison of performance on step response (reference (blue), without PID (red), manual tuning (yellow), after SED optimization (purple)):
The PID-SED method achieved best values of rise time (Tr) 2.2091 seconds, settling time (Ts) 5.3218 seconds, and Overshoot 4.7055 mmHg on the step signal experiment, meanwhile manual tuning did not reach the target values (reference).
The SED optimization was applied to the simulation with different target and supply pressures (30 - 262.5 mmHg and 150 - 1500 mmHg respetively) and returned performance value ranges:
- Rise time (Tr): 1.3224 – 3.0831 seconds
- Settling time (Ts): 5.3218 – 9.5292 seconds
- Overshoot: 0 – 12.1734 mmHg
The optimized model was then used to simulate fast and slow inflating-deflating cycles (2-2 seconds and 10-10-10 seconds respectively) usually found in commercial IPC devices. The graph below shows a more stable output for the slow cycle, meaning the model is more suited for slower and longer periods of inflation and deflation. A comparison of performance on low and high supply pressures (yellow and red) is also shown:
