lds::poisson::System #

Poisson System type.
#include <lds_poisson_sys.h>

Public Functions #

	Name
	System() =default Constructs a new System.
	System(std::size_t n_u, std::size_t n_x, std::size_t n_y, data_t dt, data_t p0 =kDefaultP0, data_t q0 =kDefaultQ0) Constructs a new Poisson System.
virtual const Vector &	Simulate(const Vector & u_tm1) override Simulate system measurement.

Protected Functions #

	Name
virtual void	h() override System output function.
virtual Vector	h_(Vector x) override System output function: stateless.
virtual void	RecurseKe() override Recursively recalculate estimator gain (Ke)

Additional inherited members #

Public Functions inherited from lds::System

	Name
virtual	~System()
void	Filter(const Vector & u_tm1, const Vector & z) Filter data to produce causal state estimates.
void	f(const Vector & u, bool do_add_noise =false) system dynamics function
size_t	n_u() const Get number of inputs.
size_t	n_x() const Get number of states.
size_t	n_y() const Get number of outputs.
data_t	dt() const Get sample period.
const Vector &	x() const Get current state.
const Matrix &	P() const Get covariance of state estimate.
const Vector &	m() const Get current process disturbance/bias.
const Matrix &	P_m() const Get covariance of process disturbance estimate.
const Vector &	cx() const Get C*x.
const Vector &	y() const Get output.
const Vector &	x0() const Get initial state.
const Vector &	m0() const Get initial disturbance.
const Matrix &	A() const Get state matrix.
const Matrix &	B() const Get input matrix.
const Vector &	g() const Get input gain/conversion factor.
const Matrix &	C() const Get output matrix.
const Vector &	d() const Get output bias.
const Matrix &	Ke() const Get estimator gain.
const Matrix &	Ke_m() const Get estimator gain for process disturbance (m)
const Matrix &	Q() Get process noise covariance.
const Matrix &	Q_m() Get process noise covariance of disturbance evoluation.
const Matrix &	P0() Get covariance of initial state.
const Matrix &	P0_m() Get covariance of initial process disturbance.
void	set_A(const Matrix & A) Set state matrix.
void	set_B(const Matrix & B) Set input matrix.
void	set_m(const Vector & m, bool do_force_assign =false) Set process disturbance.
void	set_g(const Vector & g) Set input gain.
void	set_Q(const Matrix & Q) Set process noise covariance.
void	set_Q_m(const Matrix & Q_m) Set process noise covariance of disturbance evoluation.
void	set_x0(const Vector & x0) Set initial state.
void	set_P0(const Matrix & P0) Set covariance of initial state.
void	set_P0_m(const Matrix & P0_m) Set covariance of initial process disturbance.
void	set_C(const Matrix & C) Set output matrix.
void	set_d(const Vector & d) Set output bias.
void	set_x(const Vector & x) Set state of system.
void	Reset() Reset system variables.
std::vector< UniformMatrixList< kMatFreeDim2 > >	nstep_pred_block(UniformMatrixList< kMatFreeDim2 > u, UniformMatrixList< kMatFreeDim2 > z, size_t n_pred =1)
void	Print() Print system variables to stdout.

Protected Functions inherited from lds::System

	Name
void	InitVars(data_t p0 =kDefaultP0, data_t q0 =kDefaultQ0)

Public Attributes inherited from lds::System

	Name
bool	do_adapt_m whether to adaptively estimate disturbance m

Protected Attributes inherited from lds::System

	Name
std::size_t	n_x_ number of states
std::size_t	n_u_ number of inputs
std::size_t	n_y_ number of outputs
data_t	dt_ sample period
Vector	x_ state
Matrix	P_ covariance of state estimate
Vector	m_ process disturbance
Matrix	P_m_ covariance of disturbance estimate
Vector	cx_ C*x.
Vector	y_ output
Vector	z_ measurement
Vector	x0_ initial state
Matrix	P0_ covariance of initial state estimate
Vector	m0_ initial process disturbance
Matrix	P0_m_ covariance of initial disturbance est.
Matrix	A_ state matrix
Matrix	B_ input matrix
Vector	g_ input gain
Matrix	Q_ covariance of process noise
Matrix	Q_m_ covariance of disturbance random walk
Matrix	C_ output matrix
Vector	d_ output bias
Matrix	Ke_ estimator gain
Matrix	Ke_m_ estimator gain for process disturbance

Public Function Details #

System #

System() =default

System #

System(
    std::size_t n_u,
    std::size_t n_x,
    std::size_t n_y,
    data_t dt,
    data_t p0 =kDefaultP0,
    data_t q0 =kDefaultQ0
)

Parameters:

n_u number of inputs
n_x number of states
n_y number of outputs
dt sample period
p0 [optional] initial diagonal elements of state estimate covariance (P)
q0 [optional] initial diagonal elements of process noise covariance (Q)

Simulate #

virtual const Vector & Simulate(
    const Vector & u_tm1
) override

Parameters:

u_tm1 input at t-1

Return: z measurement

Reimplements: lds::System::Simulate

Simulate system and produce measurement

Protected Function Details #

h #

inline virtual void h() override

Reimplements: lds::System::h

h_ #

inline virtual Vector h_(
    Vector x
) override

Reimplements: lds::System::h_

RecurseKe #

virtual void RecurseKe() override

Reimplements: lds::System::RecurseKe

Recursively recalculate estimator gain (Ke).

References:

Smith AC, Brown EN. (2003) Estimating a State-Space Model from Point Process Observations. Neural Computation 15.

Eden UT, …, Brown EN. (2004) Dynamic Analysis of Neural Encoding by Point Process Adaptive Filtering Neural Computation 16.

Updated on 3 April 2025 at 13:48:30 EDT