Go control-system toolbox for modeling, analyzing, transforming, and designing continuous-time and discrete-time linear time-invariant models.

The state-space model is the fundamental representation. Transfer-function, zero-pole-gain, frequency-response data, model-array, generalized-model, and tunable-block workflows convert to or build on that core representation where needed. See docs/codebase-interface-diagram.md for the current interface map.

go get github.com/jamestjsp/controlsys

Note: This package depends on a gonum fork for additional LAPACK routines. Because replace directives do not propagate to downstream modules, applications that import controlsys must add this to their own go.mod:

replace gonum.org/v1/gonum => github.com/jamestjsp/gonum v0.17.3-fork

This package is intended to be usable in production control and estimation code, with the usual caveat that numerical software still needs application-specific validation.

• Pin both controlsys and the required gonum fork to explicit versions.
• Validate mission-critical models against an external reference, especially for ill-conditioned realizations and delay-heavy systems.
• System values are mutable. Use Copy before sharing a model across goroutines that may mutate names, delays, notes, or other receiver state. Use Validate after direct field edits.
• The repository CI runs go fix ./..., go vet ./..., go test -v -count=1 -race ./..., and a downstream consumer import check; those are the recommended baseline checks for downstream integrations.
• Public API and mutation semantics are tracked in docs/api-mutation-audit.md.

• Model interfaces: state-space, transfer function, zero-pole-gain (ZPK), frequency-response data (FRD), model arrays, generalized models, and tunable blocks
• Frequency response: Bode, Nyquist, Nichols, singular values
• Stability and response analysis: gain/phase margins, disk margins, bandwidth, damping, root locus, passivity, step-response metrics
• Control design: LQR, LQE (Kalman), LQI, LQG, H2 synthesis, H-infinity synthesis, pole placement, Ackermann placement, Riccati solvers (CARE/DARE)
• PID and fixed-structure tuning: PID/PID2 controller models, standard/parallel forms, Pidtune, tunable blocks, generalized closed-loop models, tuning goals, Systune, and Looptune
• State estimation: Extended Kalman Filter (EKF) for nonlinear systems
• System identification: Eigensystem Realization Algorithm (ERA) and frequency-response estimation from I/O data
• Nonlinear systems: Jacobian linearization around operating points; Smith predictor for time-delay plants
• Model arrays and physical assembly: compatible model grids for parameter sweeps and port-checked physical component assembly
• Model reduction & decomposition: controllability/observability staircase, balanced realization, balanced truncation, stable/unstable and modal separation, modal truncation
• System norms & covariance: H2/H-infinity norms, Hankel singular values, state covariance
• Interconnection: series, parallel, feedback, safe feedback, append, block diagonal, named/indexed connect, FRD interconnections, sum blocks, and LFT
• Time-domain: step, impulse, initial condition, arbitrary input (lsim), discrete simulation
• Discretization: ZOH, FOH, Tustin (bilinear), matched pole-zero, discrete-to-discrete resampling
• Transport delays: input/output/internal delays, Pade and Thiran approximations, LFT representation
• Transmission zeros & poles via staircase decomposition
• Gramians: controllability and observability

package main

import (
	"fmt"

	"github.com/jamestjsp/controlsys"
	"gonum.org/v1/gonum/mat"
)

func main() {
	// Double integrator: x'' = u
	A := mat.NewDense(2, 2, []float64{0, 1, 0, 0})
	B := mat.NewDense(2, 1, []float64{0, 1})
	C := mat.NewDense(1, 2, []float64{1, 0})
	D := mat.NewDense(1, 1, []float64{0})

	sys, _ := controlsys.New(A, B, C, D, 0)

	poles, _ := sys.Poles()
	fmt.Println("Poles:", poles)
	stable, _ := sys.IsStable()
	fmt.Println("Stable:", stable)

	tf, _ := sys.TransferFunction(nil)
	fmt.Println("Transfer function:", tf.TF)
}

MIT