PIDControl 1.4.2

Arduino process-control library covering the complete control chain from signal conditioning to closed-loop PID control.

Spanish documentation is available in README_ES.md.

Control workflow

PIDControl is organized around a simple process-control workflow:

Signal acquisition
       |
       v
Signal filtering
       |
       v
Process identification
       |
       v
Controller tuning
       |
       v
Closed-loop control

The library can also be used only as a PID controller without using the filtering, identification or tuning modules.

1. Signal filtering

PIDControlFilters provides reusable filters for conditioning the process variable before it reaches the controller or the identification algorithms.

Available filters:

• MovingAverageFilter<N> for random measurement noise.
• MedianFilter<N> for isolated spikes and outliers.
• LowPassFilter for continuous analog noise.
• ComplementaryFilter for combining two estimates of the same variable.

Example:

#include <PIDControlFilters.h>
using namespace PIDControlFilters;

MedianFilter<5> median;
LowPassFilter lowPass(1.0);

double pvMedian = median.Update(PVRaw);
double PV = lowPass.Update(pvMedian, dt);

A median filter followed by a first-order low-pass filter is useful when the signal contains both isolated spikes and continuous noise.

2. Process identification

PIDControl provides two non-blocking identification methods.

StepTest

StepTest identifies a first-order plus dead-time model and returns:

• Kp process gain
• T0 dead time
• Tp process time constant

FOPDTModel model = stepTest.GetModel();

Engineering ranges can also be defined to calculate normalized process gain:

stepTest.SetOutputLimits(0, 60);
stepTest.SetPVLimits(15, 45);

stepTest.SetOutputRange(0, 100);
stepTest.SetPVRange(0, 80);

KpNormalized = Kp * OPspan / PVspan

The safety limits used during the test are deliberately separated from the full engineering ranges.

RelayTest

RelayTest performs a non-blocking relay experiment and returns:

• Ku ultimate gain
• Tu ultimate period

Compact API example:

RelayTestConfig cfg;
cfg.amplitude = 10.0;
cfg.hysteresis = 0.5;
cfg.cycles = 3;
cfg.useCurrentReference = true;
cfg.maxTestTimeMinutes = 15.0;
cfg.outputMin = 20.0;
cfg.outputMax = 60.0;
cfg.pvMin = 15.0;
cfg.pvMax = 48.0;

relay.Start(cfg);

3. Controller tuning

PIDTuning converts the identification results into PI or PID tuning proposals.

The library never auto-applies a proposed tuning. The user must apply it explicitly with PIDTuning::Apply() or SetTunings().

StepTest -> FOPDT -> PI/PID

FOPDTModel model = stepTest.GetModel();
PIDTuning tuning(model);

PIDTuningResult pi  = tuning.IMC_PI();      // default Tf = 2*T0
PIDTuningResult pid = tuning.Lambda_PID();  // default Tf = 2*T0

Both tuning methods use Tf as the tuning parameter that determines the desired closed-loop response speed.

Available presets for Tf:

AGGRESSIVE = T0
NORMAL     = 2*T0
ROBUST     = 3*T0

RelayTest -> Ku/Tu -> Tyreus-Luyben

RelayTestResult r = relay.GetResult();

PIDTuningResult pi =
    PIDTuning::TyreusLuybenPI(r.Ku, r.Tu);

PIDTuningResult pid =
    PIDTuning::TyreusLuybenPID(r.Ku, r.Tu);

4. Closed-loop control

PIDControl supports:

• PID, PI-D and I-PD structures
• MAN/AUTO operation
• DIRECT and REVERSE control action
• PV Tracking
• bumpless transfer
• conditional-integration anti-windup
• configurable output limits
• configurable sample time

A minimal controller can be created with:

PIDControl pid(&PV, &OP, &SP, Kc, Ki, Kd);

The default configuration is:

Mode            MAN
Structure       PI_D
Action          REVERSE
PV Tracking     ON
Output limits   0..100
Sample time     100 ms

A complete declaration can also be used when the configuration must be explicit. Using a named variable for PV Tracking makes the final argument easier to understand:

const bool pvTracking = true;

PIDControl pid(
    &PV, &OP, &SP,
    Kc, Ki, Kd,
    PIDType::PI_D,
    PIDAction::REVERSE,
    pvTracking
);

In this declaration:

PIDType::PI_D          controller structure
PIDAction::REVERSE     control action
pvTracking = true      PV Tracking enabled

Typical loop:

void loop()
{
    PV = ReadProcessVariable();

    pid.Compute();

    WriteOutput(OP);
}

PIDControl can be used directly from this point without using the filtering, identification or tuning modules if suitable Kc, Ki and Kd values are already known.

TCLab reference sketch

extras/PIDControl_TCLab_UNO_R4_Test_v1.4.2.ino demonstrates the complete chain on Arduino UNO R4 WiFi + TCLab.

The serial interface provides compact identification commands:

STEP(10,2,UP,POS,0.3,60,0.10,0.4,30,0,60,15,45,0,80,0,100)

RELAY(10,0.5,3,CURRENT,15,20,60,15,48)

Stored identification results can be reprinted while they remain in memory:

IDENT
IDENT STEP
IDENT RELAY

See docs/TCLAB_TEST.md for the full command interface.

Modules

• PIDControlFilters
• StepTest
• RelayTest
• PIDTuning
• PIDControl

Documentation

• docs/FILTERS.md
• docs/STEP_TEST.md
• docs/RELAY_TEST.md
• docs/PID_TUNING.md
• docs/API.md
• docs/TCLAB_TEST.md
• docs/TERMINOLOGIA.md

License

PIDControl is released under the MIT License.

Copyright (c) 2026 Garikoitz Martinez.