 Research
 Open Access
 Published:
Gaussian processes based extended target tracking in polar coordinates with input uncertainty
EURASIP Journal on Advances in Signal Processing volume 2022, Article number: 106 (2022)
Abstract
Gaussian processes (GP) based extended target tracking (ETT) technique has attracted the attention of scientists since it can accurately estimate the kinematic states as well as the contour states of irregularshape targets. Most traditional GP methods consider the ETT problem with a linear measurement model in Cartesian coordinates and ignore the input uncertainty of GP for simplicity. In many applications, however, measurements are generated with highresolution sensors in polar coordinates. More importantly, the ignorance of input uncertainty in GP will exacerbate the tracking performance in these cases. In order to track an irregularshape extended target in polar coordinates with input uncertainty, an improved GPbased probabilistic data association (IGPPDA) algorithm is developed that includes the following enhancements: Firstly, a nonlinear measurement model is used and the unbiased converted measurement technique is invoked for the ETT problem. Secondly, the analytical form of statistical properties of the input uncertainty due to measurements noise and predicted error is derived. Thirdly, an IGPPDA taking into account measurement origin uncertainty as well as input uncertainty is proposed, and three suboptimal implementations are provided. Last, the posterior CramerRao lower bound of ETT with input uncertainty is derived. Simulation results verify the effectiveness of the proposed method.
Introduction
With the rapid development in highresolution sensor technology, extended target tracking (ETT) is attracting more and more attention with applications in military and civilian fields, such as automotive active safety systems [1, 2], advanced driver assistance systems [3, 4], maritime and ground surveillance [5, 6], robotic and control [7] and points cloud processing [8,9,10]. For the problem of ETT, a target is assumed to have multiple random scattering points and hence may yield multiple measurements per scan. The objective of ETT is to estimate the kinematic state along with the contour state of the target of interest [11].
Many ETT methods have been proposed in the last decades, such as Sequential Monte Carlo (SMC) [12, 13], Random Matrix (RM) [14,15,16,17,18], Random Hypersurface Models (RHM) [19,20,21], Gaussian Processes based ETT (GPETT) [6, 22, 23], and so on. Among these works, the GP has the ability to model and learn an unknown radius function of an irregularshape target analytically and can provide benefits toward higherprecision contour estimation [24]. In [22], a GPbased extended Kalman filter (GPEKF) is proposed for ETT in a benign scenario in the absence of clutter. Moreover, the GP model is combined with the labeled multiBernoulli filter for tracking multiple extended targets [25]. In order to track a nonrigid and asymmetric extended target, a Spatiotemporal GPbased ETT technique is proposed in [23]. The above GPETT algorithms verify the contour estimation ability of GP in the benign scenario, ignoring the effect of the measurement origin uncertainty on the tracker. In some applications, e.g., tracking the ships in the ocean, the highresolution sensor receives a lot of measurements not only from targets but also from sea clutter. In this case, the measurement origin uncertainty must be taken into account. To address the problem of ETT in clutter, the GPbased probabilistic data association (GPPDA) is proposed in [26] and the posterior CramerRao lower bound (PCRLB) is derived to evaluate the performance of ETT with measurement origin uncertainty. Besides, the GPPDA is combined with the interacting multiple model (IMM) to simultaneously track a ship and its wake in [6]. As mentioned above, to track extended targets in clutter, the GPPDA approach seems very promising since the fact that it can not only estimate the kinematic state but also give a precise contour state of the irregularshape target.
Most previous GP methods mentioned above assume that the ETT measurement model is linear and the GP input is known accurately for simplicity. This assumption is reasonable in civilian applications but does not hold in many military applications. In maritime surveillance, for example, highresolution maritime radars are used to measure the range, bearing, and Doppler of the target in polar coordinates and cannot provide position measurements directly. In other words, a nonlinear measurement model is more suitable in this case. Moreover, the input uncertainty of GP, caused by poor bearing resolution and imprecise state prediction, will significantly affect the tracking performance of ETT and cannot be ignorable. In the GP literature, there are two categories to deal with the uncertain input [27]. The first category, called heteroscedastic GP models, transforms the input variance into a parameter that varies with the output residual by changing the distribution of the output noise [28,29,30]. However, this approximate method ignores the structure of the uncertain inputs. Another category deals with the uncertain input directly in the prediction phase and generally approximates the mean and variance of the prediction distribution of GP [31,32,33,34]. Further research is developed in [35, 36] where a noiseinput GP model was considered both in the prediction phase and in the training phase.
In this paper, we consider the problem of tracking the irregularshape extended targets in polar coordinates with input uncertainty along with measurement origin uncertainty. This work is an extension of the author’s previous results presented in [26]. Under the GPPDA framework, the measurement origin uncertainty can be effectively solved. In order to modify the effect of input uncertainty on the tracking performance of GPETT, an improved GPbased probabilistic data association (IGPPDA) algorithm is proposed to track an extended target in clutter. The main contributions of this paper are as follows:

The analytical statistical property of the uncertain input is derived:
The unbiased converted measurement (UCM) technique [37] is invoked to transform raw nonlinear measurements into unbiased linear measurements. The uncertain input of GP is assumed to be Gaussian distribution [27] and its statistical property is given analytically using the measurement noise variance, converted bias covariance, and the state predicted covariance.

Three approximation implementations are given to solve the uncertain input:
An IGPPDA algorithm taking into account uncertain input is proposed, where the prediction distribution of GP is modified with the statistical property of uncertain input. Since the analytical solution to the prediction distribution of GP is not available in most cases, three approximation implementations are given.

A new posterior CramerRao lower bound is derived:
In order to evaluate the tracking performance of ETT, a new posterior CramerRao lower bound (PCRLB) of ETT considering not only the measurement origin uncertainty but also the input uncertainty is derived. As a larger covariance of GP prediction distribution due to input uncertainty is included, the new performance bound is more conservative and reasonable than the previous result in [26].
The remainder of this paper is structured as follows. Section 2 introduces the background of standard GP and GP with uncertain input. Section 3 describes the ETT problem in polar coordinates with measurement origin uncertainty. In Sect. 4, the statistical property of the uncertain input and the IGPPDA algorithm is proposed. Section 5 derives a more reasonable PCRLB of ETT. Simulation and conclusions are presented in Sects. 6 and 6, respectively.
Background
The standard GP and the GP with uncertain input are reviewed briefly in this section, that forms the foundation of the proposed method.
Standard Gaussian processes
A Gaussian process is a generalization of the multivariate Gaussian distribution suitable for learning the distribution of an unknown function, and it can make inferences in the functionspace view. Consider a function as follows:
where u is the input of latent function \(f(\cdot )\), y is the noisy output, and \(\epsilon \sim {\mathcal {N}}(0, \sigma _{\epsilon }^{2})\) is a Gaussian distributed noise. Under the GP framework, the latent function \(f(\cdot )\) can be described by the mean function a(u) and covariance function \(\kappa (u,u')\) as follows:
Given a training set \({\mathcal {D}}\) of \(N_{gp}\) observations, \({\mathcal {D}}=\bigl \{(u_i,y_i)\mid i=1,\ldots ,N_{gp}\bigr \}\), where \(u_i\) denotes the \(i^{\mathrm{th}}\) training input and \(y_i\) the corresponding training output. Denote \({\mathbf {u}}=[u_1,\dots ,u_{N_{gp}}]\) and \({\mathbf {y}}=[y_1,\dots ,y_{N_{gp}}]\) as the training input vector and training output vector, respectively. The objective of GP is to use a set of training data \({\mathcal {D}}=[{\mathbf {u}},{\mathbf {y}}]\) to learn the unknown latent function \(f(\cdot )\), or the predictive distribution of the function outputs given the inputs [24].
For a test input \(u_{*}\), since the joint distribution is a Gaussian, the predictive distribution \(p\left( f(u_{*}) \mid u_{*},{\mathcal {D}}\right)\) is also Gaussian with mean \(\mu _{\text {SGP}}({u}_{*})\) and variance \(v_{\text {SGP}}({u}_{*})\) [33]
where
where I is an \(N_{gp} \times N_{gp}\) identity matrix, \(K(\cdot ,\cdot )\) is a matrix of covariance function, and
Herein, the prediction mean \(\mu _{\text {SGP}}(u_*)\) serves as an estimate of the latent function \(f(u_{*})\) with the uncertainty of \(\sqrt{v_{\text {SGP}}(u_{*})}\).
Gaussian processes with uncertain input
Usually, the input of GP \(u_{*}\) is assumed to be accurately known and noisefree. However, this assumption does not hold in some applications. For example, in a discretetime series analysis problem, the current input is estimated from the last step. For a GP model that ignores the cumulative prediction variance, its model is not conservative enough, which makes the prediction accompanied by unrealistic small uncertainty [31]. Assume that the uncertain input of latent function \(u_{*}\) follows a Gaussian distribution with mean \(\mu _{u_{*}}\) and variance \(\Sigma _{u_{*}}\) [32], the prediction distribution can be obtained by
where \(p\bigl (f(u_{*})\mid u_{*},{\mathcal {D}}\bigr )\) is a Gaussian distribution with mean \(\mu _{\text {SGP}}(u_*)\) and variance \(v_{\text {SGP}}(u_*)\) given by (4) and (5), respectively. Some methods are proposed to approximate the above integral of prediction distribution since it is analytically intractable. For example, the numerical approximation method and the Taylor Series approximation method.
In numerical approximation, the general solution of (8) can be defined as [27]:
where \(u_{i}^{*}\) is the ith random or determinate sample from \(p\left( u^{*}\right)\), \(N_t\) denotes the number of sampling.
In Taylor Series approximation, the analytic Gaussian approximation only needs the first two moments of \(p \bigl ( f(u_{*}) \mid \mu _{u_{*}}, \Sigma _{u_{*}}, {\mathcal {D}} \bigr )\). They are obtained by the law of iterated expectation and the law of conditional variances [31]:
where
The predictive variance is given as
where
The input uncertainty does not provide any correction for the prediction expectation of GP. However, it provides a correction term of zeroorder to the variance of prediction, and more details refer to [31].
Dynamic and measurement models
In this section, the GPbased dynamic model and the measurement model of ETT are provided in 2D space, more details see [22, 26]. Herein the measurement model is established in polar coordinates with measurement origin uncertainty.
Consider the following dynamic model of an irregularshape extended target
where \({\mathcal {X}}_{k}=[({\mathbf {x}}_{k}^{{\mathbf {s}}})^\top ,({\mathbf {x}}_{k}^{{\mathbf {f}}})^\top ]^\top\) is the extended target state at time k, \({\mathbf {x}}_{k}^{\mathbf {s}}\) and \({\mathbf {x}}_{k}^{\mathbf {f}}\) denote the kinematic state and contour state, respectively. The term \({F_{k1}}={\text {diag}}\{F_{k1}^{{\mathbf {s}}},F_{k1}^{{\mathbf {f}}}\}\) is state transition matrix and \({\mathbf {v}}_{k1}=[({\mathbf {v}}_{k1}^{{\mathbf {s}}})^\top ,({{\mathbf {v}}}_{{k1}}^{{\mathbf {f}}})^\top ]^\top\) is a zeromean white Gaussian noise with covariance matrix \(Q_{k1}={\text {diag}}\{Q_{k1}^{{\mathbf {s}}},Q_{k1}^{{\mathbf {f}}}\}\). The specific form of state transition matrix and covariance matrix will be given in the simulation. Then, the Eq. (14) can be rewritten as
where the kinematic state \({\mathbf {x}}_{k}^{\mathbf {s}}= [x_k, {\dot{x}}_k, y_k, {\dot{y}}_k, \phi _k, {\dot{\phi }}_k]^\top\) is a sixdimension vector, which is composed of the target centroid position \({\mathbf {x}}_{k}^{{\mathbf {c}}} \triangleq [x_k,y_k]^\top\), the target velocity \([\dot{x_k},\dot{y_k}]^\top\), the heading \(\phi _k\) and the heading rate \({\dot{\phi }}_k\). The contour state is denoted as \({{\mathbf {x}}}_{k}^{{\mathbf {f}}}=[f(\theta _{k,1}^{}), f(\theta _{k,2}),\dots f(\theta _{k,N})]^\top\), where \(\theta _{k,i}\) and \(f(\theta _{k,i})\), \(1\le {i}\le {N}\), are the angle and the unknown radius function of the ith contour point, respectively, and N is the number of contour points. Usually, the prior knowledge of the target contour is unavailable, and the angle of contour point \(\theta _{k,i} \in [0,2\pi )\) can be chosen as \(\theta _{k,i} = {2\pi (i1)}/{N},1 \le i \le N\). In addition, all angles of the contour points are stacked into a vector as \(\Theta \triangleq [ \theta _{k,1}, \ldots , \theta _{k,N} ]^\top\). The unknown radius function \(f(\theta _{k,i})\) can be described by a GP model as follows
Without loss of generality, the mean function is assumed to follow \(a(\theta _k) \sim {\mathcal {N}}(0,\sigma _r^2)\), and the covariance function can be chosen as a modified squared exponential function
to encode the periodicity of angle \(\theta _{k}\) [22], where \(\sigma _f\) and l are the hyperparameters of GP.
An extended target has multiple fluctuating scattering points, which may appear on or within the contour. These scattering points may generate multiple noisy measurements per scan. Previous work usually assumes that the sensor can obtain the positions of scattering points directly and invoke a measurement model in Cartesian coordinates. However, many types of sensors, such as radar and sonar, can only measure the range and bearing of the scattering points in polar coordinates. In this subsection, a generalized nonlinear measurement model with measurement origin uncertainty is presented.
Suppose that there are \(M_{k}\) measurements at time k. In view of the measurement origin uncertainty, the jth measurement, denoted as \(c_{k,j}=[d_{k,j},\beta _{k,j}]^{\top }\), \(1\le {j}\le {M_k}\), may be originated from a scattering point of the target or from clutter, where \(d_{k,j}\) and \(\beta _{k,j}\) denote the range and bearing, respectively. Thus, a generalized nonlinear measurement model is given as follows.
For a clutteroriginated measurement, the clutter \(\mathbf {\chi }_{k,j}\) is assumed to follow a uniform spatial distribution and a Poisson cardinality distribution with parameter \(\lambda _c\). For a targetoriginated measurement, the nonlinear function \({\mathbf {h}}_{k,j}^{\text {pol}}(z_{k,j})\) represents the noisefree measurement and \(w_{k,j}^{\text {pol}} \sim {\mathcal {N}}(0, R^{\text {pol}})\) is the measurement noise in polar coordinates, \(R^{\text {pol}}={\text {diag}}\{\sigma _d^2,\sigma _\beta ^2\}\).
In the above equation, \(z_{k,j}=[z_{k,j}^{\xi },z_{k,j}^{\eta }]^\top\) denotes the xaxis and yaxis position of the scattering point corresponding to the jth measurement (see Fig. 1). For simplicity, the scattering points are assumed to distribute along the contour of the target here. In this case, the position of scattering point \(z_{k,j}\) is given by
where \({\mathbf {h}}_{k,j}^{\text {car}}({\mathcal {X}}_{k})\) denotes the measurement function in the Cartesian coordinates, \({\mathbf {p}}_{k,j}\) is the orientation vector and \(\theta _{k,j}^L\) is the angle of the scattering point \(z_{k,j}\) relative to the target centroid \({\mathbf {x}}_k^{{\mathbf {c}}}\) in local coordinates, more details see [22]. In addition, \(f(\theta _{k,j}^L)\) is a unknown radius function and can be rewritten as follows with the GP model.
where \(\mu _{\text {SGP}}(\theta _{k,j}^{L})\) is the prediction mean of \(p(f(\theta _{k,j}^L)\mid \theta _{k,j}^L,\Theta )\), and \(w_{k,j}^{{\mathbf {f}}} \sim {\mathcal {N}}(0,v_{\text {SGP}}({\theta }_{k,j}^{L}))\) is a zeromean white Gaussian noise of radius function with the following variance
Then, substitute the above equations into (18), the targetoriginated measurement in polar coordinate can be rewritten as
Denote the measurement set as \({\mathcal {C}}_k=\left\{ c_{k,j}\right\} _{j=1}^{M_k}\) at time k. The main objective of ETT is to estimate the kinematic state \({\mathbf {x}}_{k}^{\text {s}}\) and the contour state \({\mathbf {x}}_{k}^{\text {f}}\) of the extended target from a total measurement set \({\mathcal {C}}_{1:k} \triangleq \{C_1,\dots ,C_{k}\}\) up to and including time k.
Methodology
In this section, the statistical properties of the uncertain input of GPETT are derived. Then an improved GPPDA algorithm is proposed accounting for input uncertainty. Also, three approximation implementations are provided since the analytical solution of prediction distribution is unavailable in most cases.
Statistical properties of uncertain input
Under the framework of GPETT [22], the GP input is the angle of scattering point relative to the target centroid in local coordinates and it is given as
where \(z_{k,j}=[z_{k,j}^{\xi },z_{k,j}^{\eta }]^\top\) denotes the position vector of the scattering point corresponding to the jth measurement, \({\mathbf {x}}_k^{{\mathbf {c}}}=[{x}_{k},{y}_{k}]^\top\) denotes the position vector of target centroid, and \(\phi _{k}\) denotes the heading of target. Note that the above variables \(z_{k,j}\), \({\mathbf {x}}_k^{{\mathbf {c}}}\) and \(\phi _{k}\) are all unknown to the tracker.
They are replaced by their observations and estimates in actual processing [22, 26] and hence yield the input uncertainty (see Fig. 2). Therein, the term \(\theta _{k,j}^{\text {G}}=\theta _{k,j}^{{L}}+\phi _{k}\) is the angle of scattering point relative to the target centroid in global coordinates. The true distribution of GP input \(\theta _{k,j}^{L}\) is intractable even if all the above variables are Gaussian distributed. For simplicity, the GP input is assumed to follow a Gaussian distribution [23, 31], i.e., \(\theta _{k,j}^{L} \sim {\mathcal {N}}(\mu _{\theta _{k,j}^{L}},\Sigma _{\theta _{k,j}^{L}})\). The mean of GP input \(\mu _{\theta _{k,j}^{L}}\) is approximated as
where \(c_{k,j}^{{\textbf {u}}}\triangleq [c_{k,j}^{{\mathbf {u}},\xi },c_{k,j}^{{\mathbf {u}},\eta }]^\top\) is the linear measurement converted from \(c_{k,j}\) via UCM technique [37], more details see “Appendix 1”. \([{\hat{x}}_{k\mid k1},{\hat{y}}_{k\mid k1}]^\top\) and \({\hat{\phi }}_{k\mid k1}\) are the position prediction of target centroid and the heading prediction, respectively. They can be predicted from the dynamic model with the state estimate at time \(k1\).
Now the variance of input \(\Sigma _{\theta _{k,j}^{L}}\) can be derived as follows according to the error propagation theory [27].
where \({\tilde{\sigma }}_{c1}^2\), \({\tilde{\sigma }}_{c2}^2\), \({\tilde{\sigma }}_{x x}^2\), \({\tilde{\sigma }}_{y y}^2\) and \({\tilde{\sigma }}_{\phi \phi }^2\) are given as
where eig\([\cdot ]\) denotes the eigendecomposition operator. \(R_{k,j}^{{\mathbf {u}}}\) is the converted measurement noise covariance via UCM technique, see “Appendix 1”. Also \(P_{{\mathcal {I}}} \in {\mathcal {R}}^{3 \times 3}\) is the predicted state covariance matrix whose main diagonal elements are the variances of \({\hat{x}}_{k\mid k1}\), \({\hat{y}}_{k\mid k1}\) and \({\hat{\phi }}_{k\mid k1}\), respectively.
Improved GPPDA
In this subsection, an improved GPPDA method is proposed for the problem of ETT in the clutter with polar measurements. Firstly, the predicted state and its covariance are obtained from the estimated state at the last time. Secondly, the original polar measurements are converted into linear measurements via the UCM technique. And three implementations are presented to approximate the prediction distribution of GP and hence to predict measurements accounting for the input uncertainty. Thirdly, the validation gate is constructed to select the validated measurement. Last, the probabilities of all feasible association events are calculated to update the state and its covariance. The algorithm flowchart is shown in Fig. 3.
Step 1: State prediction
Given the state estimate \(\hat{{\mathcal {X}}}_{k1}\) and the corresponding covariance \({P}_{k1}\) at time \(k1\), the predicted state \(\hat{{\mathcal {X}}}_{k\mid k1}\) and covariance \({P}_{k\mid k1}\) can be calculated as
Thus, one can obtain \({\hat{x}}_{k\mid k1}\), \({\hat{y}}_{k\mid k1}\) and \({\hat{\phi }}_{k\mid k1}\) from \(\hat{{\mathcal {X}}}_{k\mid k1}\), and obtain \(P_{{\mathcal {I}}}\) from \({P}_{k\mid k1}\).
Step 2: Measurement prediction
Given the original polar measurement \(c_{k,j}\) and its variance \(R^{\text {pol}}\), \(1\le {j}\le {M_k}\), where \(M_k\) is the number of measurements at time k. They can be converted from polar to Cartesian coordinates via the UCM technique [37], see “Appendix 1”. Denote the converted measurement and its variance as \(c_{k,j}^{{\mathbf {u}}}\) and \(R_{k,j}^{{\textbf {u}}}\), respectively.
Using \(c_{k,j}^{{\mathbf {u}}}\), \(R_{k,j}^{{{\textbf {u}}}}\), \({\hat{x}}_{k\mid k1}\), \({\hat{y}}_{k\mid k1}\), \({\hat{\phi }}_{k\mid k1}\) and \(P_{{\mathcal {I}}}\), the first two moments of the GP input \({\hat{\theta }}_{k,j}^{L}\), i.e., \(\mu _{\theta _{k,j}^{L}}\) and \(\Sigma _{\theta _{k,j}^{L}}\), can be estimated as described in Sect. 4.1. Thus, the predicted measurement \({{\hat{c}}}_{k\mid k1,j}^{{\mathbf {u}}}\) and the corresponding covariance of the measurement noise \({\hat{R}}_{k,j}^{{\mathbf {u}}}\) are given as
where \(\hat{{\mathbf {p}}}_{k,j}=\left[ \cos (\mu _{\theta _{k,j}^{L}} + {\hat{\phi }}_{k\mid k1}),\sin (\mu _{\theta _{k,j}^{L}} + {\hat{\phi }}_{k\mid k1}) \right] ^\top\) is the orientation vector, and \(\hat{{\mathbf {x}}}_{k\mid k1}^{{\mathbf {c}}}=[{\hat{x}}_{k\mid k1}, {\hat{y}}_{k\mid k1}]^{\top }\) the predicted position of target centroid. Also \(\mu _{\text {IGP}}({\hat{\theta }}_{k,j}^{L})\) and \(v_{\text {IGP}}({\hat{\theta }}_{k,j}^{L})\) denote the mean and covariance of the GP predicted distribution with uncertain input, respectively. According to Sect. 2.2, the analytical solution to the predicted distribution is intractable [32]. In the following, three suboptimal implementations are invoked to approximate \(\mu _{\text {IGP}}({\hat{\theta }}_{k,j}^{L})\) and \(v_{\text {IGP}}({\hat{\theta }}_{k,j}^{L})\).
(1) Sigma sampling (SS) method [38]:
where \({\hat{\theta }}_{k,j}^{L,(i)}\), \(\omega _{k}^{m,(i)}\) and \(\omega _{k}^{c,(i)}\) are the ith determined sampling point, its mean weight and covariance weight, respectively, n is the dimension of \({\hat{\theta }}_{k,j}^{L}\). More details see “Appendix 2”. In addition, \(\mu _{\text {SGP}}({\hat{\theta }}_{k,j}^{L,(i)})\) is the mean of standard GP predicted distribution. According to (4), one has
where \(\hat{{\mathbf {x}}}_{k\mid k1}^{{\mathbf {f}}}\) is the predicted contour state obtained from \(\hat{{\mathcal {X}}}_{k\mid k1}\).
(2) MonteCalor (MC) method [41]:
where \({\hat{\theta }}_{k,j}^{L,(i)}\) is the ith random sampling point drawn from the Gaussian distribution \({\mathcal {N}}(\mu _{\theta _{k,j}^{L}},\Sigma _{\theta _{k,j}^{L}})\) via the MC sampling technique, and \(N_{\text {MC}}\) is the number of sampling points. Similarly, \(\mu _{\text {SGP}}({\hat{\theta }}_{k,j}^{L,(i)})\) can be calculated by (36).
(3) Taylor Series (TS) method [34]:
According to (10) and (12), one has
where \(v_{\text {SGP}}(\mu _{\theta _{k,j}^{L}})\) is covariance of standard GP predicted distribution. According to (5), one has
Substituting the above approximations of \(\mu _{\text {IGP}}({\hat{\theta }}_{k,j}^{L})\) and \(v_{\text {IGP}}({\hat{\theta }}_{k,j}^{L})\) into (32) and (33), one has the predicted measurement \({{\hat{c}}}_{k\mid k1,j}^{{\mathbf {u}}}\) and the corresponding covariance \({\hat{R}}_{k,j}^{{\mathbf {u}}}\).
Step 3: Validation gate
Because an extended target has multiple scattering points and hence multiple measurements per scan, its validation gate can be constructed as a union of multiple validation subgates. Each subgate is centered on one of the predicted measurements \({{\hat{c}}}_{k\mid k1,j}^{{\mathbf {u}}}\). The jth converted measurement \(c_{k,j}^{{\mathbf {u}}}\) is considered as a validated measurement if it falls within the following subgate \(\Omega _{k,j}\):
In the above, g is the gate parameter and \(S_{k,j}\) is the measurement innovation covariance,
where \(H_{k,j}\) is the Jacobian matrix of \({\mathbf {h}}_{k,j}^{\text {car}}(\cdot )\) evaluated at \({\mathcal {X}}_{k\mid k1}\). Assume that there are \(m_{k}\) validated measurements at time k, the overall validation gate can be expressed as \(\Omega _{k}=\bigcup _{j=1}^{m_{k}} \Omega _{k,j}\). More details see [26].
Step 4: Association event enumeration and state update
Given the \(m_{k}\) validated measurements, the total number of association events is \(2^{m_{k}}\). Let \(A_{k,\psi }^{n_\psi }\) denote the \(n_{\psi }\)th association event where \(\psi\) out of \(m_{k}\) validated measurements are originated from targets. The corresponding posterior probability \(\gamma _{k,\psi }^{n_{\psi }}\) can be calculated as:
where c is the normalization constant that is obtained by summing the numerators over all association events \(A_{k,\psi }^{n_{\psi }}\). And the two terms in (44) are expressed as:
where \(V_{k}\), \(P_{G}\), \(P_{D}\), \(\mu _{t}(\cdot )\) and \(\mu _{F}(\cdot )\) are the volume of \(\Omega _{k,j}\), gate probability, detection probability, the probability mass function of the number of targetoriginated measurements and that of the number of clutter measurements, respectively. The output estimate is obtained by summing all conditional estimates as
where \(C_{m_{k}}^{\psi }=m_{k}!/[(m_{k}\psi )!\psi !]\). And its covariance \(P_{k}\) are given by
where \(\hat{{\mathcal {X}}}_{k,\psi }^{n_{\psi }}\) and \(P_{k,\psi }^{n_{\psi }}\) denote the posterior conditional estimate and its covariance on the association event \(A_{k,\psi }^{n_\psi }\), respectively.
Performance analysis
In [26], the PCRLB of ETT in the presence of measurement origin uncertainty is first derived using the GP model. However, the result is somewhat overoptimistic since the input uncertainty of GP is omitted. In this section, we derive a more conservative PCRLB which takes into account the input uncertainty. Note that all formulas are evaluated at the true state of the target, which is only available for simulation cases. Let \(\hat{{\mathcal {X}}}_{k}\) be the unbiased estimate of \({\mathcal {X}}_{k}\) conditioned on the converted measurement set \(c_{1:k}^{{\mathbf {u}}}\). The PCRLB on the covariance \(P_{k}\) is the inverse of the Fisher information matrix (FIM) \(J_{k}\). That is
The FIM can be evaluated recursively by
with the initial value \(J_{0}=(P_{0})^{1}\). The term \(J_{c,k}\) is measurement contribution to the PCRLB and is defined as
where \(p(c_{k}^{{\mathbf {u}}}\mid {\mathcal {X}}_{k})\) is the likelihood function. Assume that the number of targetoriginated measurements is \(\psi\), the form of \(J_{c,k}\) is given by [26]:
where \({\hat{q}}_2\) is the information reduction factor obtained by offline Monte Carlo. More details about \({\hat{q}}_2\) refer to [26, 40]. In addition, \(H_{k,i}^{\text {car}}\) denotes the Jacobian matrix of \({\mathbf {h}}_{k,i}^{\text {car}}({\mathcal {X}}_{k})\) and \({\hat{R}}_{k,i}\) is the predicted covariance of measurement noise. According to (33), the variance of GP with uncertain inputs is lager than standard GP, which leads to a smaller \(J_{c,k}\) and results in a more conservative PCRLB. In addition, the derivation of \({\hat{q}}_2\) accounting for the GP input uncertainty is beyond the scope of this work, and it will be discussed for further research.
Results and discussion
In the simulation, the performance of the proposed methods named IGPPDASS, IGPPDAMC, and IGPPDATS, is evaluated in two scenarios to track a moving ship in benign and clutter cases, respectively. In both cases, the target follows a maneuvering motion and is monitored by a highresolution radar, which is located at the origin in the global coordinates. To verify the effectiveness of the proposed methods, they are compared with the standard GPPDA and with the derived PCRLB. The rootmeansquare error (RMSE), the Mahalanobis distance (MD), and the elapsed time are used as performance metrics. In addition, the effect of measurement noise on the GP input uncertainty and the tracking performance is analyzed. 100 Monte Carlo runs are used on a PC with an i58250U CPU running at 1.60 GHz, and all algorithms are implemented in MATLAB.
ETT in benign case
In the first case, the performance of the proposed methods is tested in a benign environment where no clutter exists. The surveillance duration is 100 s. The ship follows a nearly constant velocity (CV) motion model from \(k=1\) s to \(k=30\) s. Then, it makes a coordinated turn of \(90^\circ\) from \(k=31\) s to \(k=80\) s and returns to a nearly CV model from \(k=81\) s to the end. The initial kinematic state of target is \(X_0=X_{0}^{\text {tru}}+{\bar{X}}_{0}\) with the corresponding covariance \(P_{0}={\text {diag}}\{1 {\text { m}}^2,0.1 {\text { m}}^2/{\text {s}}^2,1 {\text { m}}^2,0.1{\text { m}}^2/{\text {s}}^2,0.1 {\text { rad}}^2/{\text {s}}^2,0.001 {\text { rad}}^2/{\text {s}}^2\}\), where \({\bar{X}}_{0} \sim {\mathcal {N}}(0, P_{0})\) and \(X_{0}^{\text {tru}}=[10 {\text { m}},\sqrt{2} {\text { m/s}}, 10 {\text { m}}, 0 {\text { m/s}}, 0 {\text { rad}}, 0 {\text { rad/s}}]^\top\). The scan period is \(T=1\) s. The process transition matrix and process noise covariance of kinematic state is given by
where \(\otimes\) denotes the Kronecker product and \(I_3\) is a threedimensional identity matrix. The term \(\sigma _q=0.1 {\text { m}} \cdot {\text {s}}^{3/2}\) and \(\sigma _{q^{\phi }}=0.001 {\text { rad}}\cdot {\text {s}}^{3/2}\) denotes the standard deviations of the position and the heading process noise, respectively. In the measurements model, the standard deviation of the measurements is \(\sigma _{d}=0.1 {\text { m}}\) and \(\sigma _{\beta }=0.0001 {\text { rad}}\). The detection probability is \(P_D=0.95\) and the gating probability is \(P_G=0.99\). The volume of the surveillance region is \(V_c=150\times \pi ({\text {m}}\cdot {\text {rad}})\), and the number of the measurements scattering points follows a Poisson distribution with expected cardinality per unit volume \(\lambda _{t}=2 \times 10^{1} {\text { m}}^{2}\).
The extended target contour is assumed to be stationary during the surveillance. The number of contour points is set to \(N=12\), and the contour transition matrix is given as \(F^{\text {f}}=I_{N}\) and \(Q_{k}^{\text {f}}=0\), where \(I_{N}\) is an \(N \times N\) identity matrix. In addition, the initial contour state \(f_0(\theta _{1:N})=2 {\text { m}}\). The hyperparameters of the GP is set to \(\sigma _r=2\),\(\sigma _f=1.5\), and \(l=\pi /8\). The number of random sampling points in IGPPDAMC is \(N_{\text {MC}}=100\), and the parameters in the IGPPDASS are as follows: \(n=1\), \(\varsigma =0\), \(\varpi =2\) and \(\alpha =\sqrt{2}\) [39].
Figure 4 shows the true and estimates contour/trajectory at time \(k=1,25,50,75,100 {\text { s}}\), respectively. It can be seen that all methods achieve satisfied estimate accuracy in the benign case. Figure 5 illustrates the RMSE and PCRLB curves of the contour estimate with four methods. The left subfigure represents the local PCRLB curves from \(k=70\) s to \(k=100\) s, and the right one shows the local RMSE curves between the same interval. It can be seen that the contour estimate RMSE of the proposed methods is smaller than that of GPPDA, and the proposed PCRLB is more conservative and reasonable. Moreover, the average RMSE and PCRLB of the target centroid estimate, the average RMSE and PCRLB of the contour estimate, Mahalanobis distance (MD) [35], and the elapsed time per run of all methods are compared in Table 1. It shows that all proposed methods achieve better performance than GPPDA in RMSE and MD. In addition, the proposed methods have a more conservative PCRLB by taking into account the GP input uncertainty. The elapsed time of the GPPDA, IGPPDASS, and IGPPDATS is comparable and that of IGPPDAMC is longer. Figure 6 shows the error bars of the radius function estimates using different methods at \(k=100\) s. The solid red line represents the true value of the radius function and the error bar shows the twice standard deviations of the radius function distribution. It can be seen that the proposed methods have better posterior means than the standard GPPDA. Also, the confidence intervals of the proposed methods can cover the true value better. In other words, the proposed methods achieve better estimate performance in the firsttwo moment than the standard GPPDA by taking into account the input uncertainty. Among all methods, the IGPPDAMC provides the best coverage of the truth value.
Figure 7 shows the effect of measurement noise on input uncertainty. It can be seen that the input uncertainty \(\Sigma _{\theta ^{L}}\) is greatly affected by the measurement noise \(\sigma _d\) and \(\sigma _{\beta }\). It increases as the measurement noise increases. Table 2 shows the average RMSE and PCRLB of the contour estimates against different \(\sigma _d\) and \(\sigma _{\beta }\) with four methods. It can be seen that as the measurement noise increases, the accuracy of the contour estimation decreases. For the same measurement noise level, the IGPPDAMC method achieves the smallest RMSE. Besides, the proposed methods have more conservative PCRLBs than GPPDA. Similarly, Table 3 shows the centroid estimate RMSE and PCRLB. The result is consistent with the contour estimate, i.e., the proposed methods have smaller RMSE and more conservative PCRLB.
Table 4 compares the contour estimate MD and the elapsed time with different methods. It can be seen that the proposed methods take a significant performance advantage in MD. Among all the methods, the IGPPDAMC is the most timeconsuming since lots of samples are required.
ETT in clutter case
In the second scenario, the methods are evaluated on a maneuvering ETT scenario in clutter. The number of clutter follows a Poisson distribution with expected cardinality per unit volume \(\lambda _{c}=2 \times 10^{4}/({\text {m}}\cdot {\text {rad}})\), and the other parameters as the same as those in the first scenario.
Figure 8 shows the true and estimates contour/trajectories at time \(k=1,25,50,75,100\) s, respectively. Figure 9 shows the RMSE and PCRLB curves of the average contour estimate with four methods. The left subfigure represents the local PCRLB curves from \(k=70\) s to \(k=100\) s, and the right one shows the local RMSE curves in the same interval. Figure 10 shows the error bars of the radius function estimated via different methods at \(k=100\) s. The conclusions are the same as those in the first scenario and are omitted here for brevity.
Tables 5 and 6 show the average RMSE and PCRLB of contour estimate and target centroid estimate via different methods against different measurement noise, respectively. Table 7 compares the contour estimate MD and the elapsed time with different methods against different measurement noise. We can obtain the same conclusions as those in the first scenario.
Tables 8 and 9 show the average RMSE and PCRLB of contour estimate and target centroid estimate against different clutter densities, respectively. Table 10 represents the MD and elapsed time. It can be seen that as the clutter density increases, the accuracy of the contour and centroid estimation decreases, and the computational load increases. For the same clutter level, the IGPPDAMC method achieves the best performance in RMSE and MD at the cost of more computation burden.
Conclusion and future work
In this paper, the IGPPDA method is proposed to track an irregularshape extended target in polar coordinates with measurement origin uncertainty. In the proposed method, the statistical property of input uncertainty is analyzed via the error propagation theory, and then a more reasonable PCRLB of ETT is derived. Three approximation implementations are utilized to estimate the predicted distribution of GP with input uncertainty. Compared with the standard GPPDA method, the IGPPDA method achieves more accurate contour and kinematic estimates in the presence of clutter. The direction of future work includes analyzing the effect of input uncertainty on the IRF of PCRLB and reducing the number of feasible association events in heavy clutter.
Availability of data and materials
Not applicable.
Abbreviations
 GP:

Gaussian Processes
 ETT:

Extended Target Tracking
 GPETT:

GPbased ETT
 GPEKF:

GPbased Extended Kalman Filter
 GPPDA:

GPbased Probabilistic Data Association
 IGPPDA:

Improved GPbased Probabilistic Data Association
 SMC:

Sequential Monte Carlo
 RM:

Random Matrix
 RHM:

Random Hypersurface Models
 PCRLB:

Posterior CarmerRao Lower Bound
 UCM:

Unbiased Converted Measurement
 SS:

Sigma Sampling
 IGPPDASS:

Improved GPPDA using Sigma Sampling method
 MC:

MonteCarlo Sampling
 IGPPDAMC:

Improved GPPDA using MonteCarlo Sampling method
 TS:

Taylor Series method
 IGPPDATS:

Improved GPPDA using Taylor Series method
 FIM:

Fisher Information Matrix
 RMSE:

Root Mean Square Error
 MD:

Mahalanobis Distance
 CV:

Constant Velocity
 IRF:

Information Reduction Factor
 \(f(\cdot )\) :

Latent function
 \({\mathbf {u}}\) :

Input set of GP
 \({u}_{*}\) :

Test input of GP
 \({\mathbf {y}}\) :

Output set of GP
 \(a(\cdot )\) :

Mean function of GP
 \(\kappa (\cdot ,\cdot )\) :

Covariance function of GP
 \(K(\cdot ,\cdot )\) :

Covariance matrix of GP
 \(\mu _{\text {SGP}}(\cdot )\) :

Predictive mean of standard GP
 \(v_{\text {SGP}}(\cdot )\) :

Predictive covariance of standard GP
 \(\mu _{u_*}\) :

Mean of GP input
 \(\Sigma _{u_*}\) :

Covariance of GP input
 \(\mu _{\text {IGP}}(\cdot )\) :

Predictive mean of GP with uncertain input
 \(v_{\text {IGP}}(\cdot )\) :

Predictive covariance of GP with uncertain input
 \({\mathcal {X}}\) :

Extended target state
 \(\Theta\) :

Angle of contour points
 z :

Position of scattering point
 c :

Original measurement
 \(\chi\) :

Clutter
 \(c^{{\mathbf {u}}}\) :

Converted measurement via UCM
 \({\mathbf {h}}^{\text {pol}}(\cdot )\) :

Nonlinear measurement function in polar coordinates
 \({\mathbf {h}}^{\text {car}}(\cdot )\) :

Measurement function in Cartesian coordinates
 \(R^{\text {pol}}\) :

Covariance of the original measurement
 \(R^{{\mathbf {u}}}\) :

Covariance of the converted measurement
 \({\hat{R}}^{{\mathbf {u}}}\) :

Covariance of the predicted measurement
References
Y. Liu, H. Ji, Y. Zhang, Measurement transformation algorithm for extended target tracking. Signal Process. 186, 108129 (2021). https://doi.org/10.1016/j.sigpro.2021.108129
K. Thormann, M. Baum, J. Honer, Extended target tracking using Gaussian processes with highresolution automotive radar, in 21st International Conference on Information Fusion (2018), pp. 1764–1770. https://doi.org/10.23919/ICIF.2018.8455630
Z. Qin, T. Kirubarajan, Y. Liang, Application of an efficient graphbased partitioning algorithm for extended target tracking using GMPHD filter. IEEE Trans. Aerosp. Electron. Syst. 56, 4451–4466 (2020). https://doi.org/10.1109/TAES.2020.2990803
M. Michaelis, P. Berthold, D. Meissner, H.J. Wuensche, Heterogeneous multisensor fusion for extended objects in automotive scenarios using Gaussian processes and a GMPHDfilter. Sens. Data Fusion Trends Solut. Appl. 2017, 1–6 (2017). https://doi.org/10.1109/SDF.2017.8126351
F. Li, I. Shugurov, B. Busam, M. Li, S. Yang, S. Ilic, PolarMesh: a starconvex 3D shape approximation for object pose estimation. IEEE Robot. Autom. Lett. 7, 4416–4423 (2022). https://doi.org/10.1109/LRA.2022.3147880
Y. Guo, Y. Li, A. Xue, R. Tharmarasa, T. Kirubarajan, Simultaneous tracking of a maneuvering ship and its wake using Gaussian processes. Signal Process. 172, 107547 (2020). https://doi.org/10.1016/j.sigpro.2020.107547
S. Yang, M. Baum, Tracking the orientation and axes lengths of an elliptical extended object. IEEE Trans. Signal Process. 67, 4720–4729 (2019). https://doi.org/10.1109/TSP.2019.2929462
X. Jiang, F.R. Yu, T. Song, Z. Ma, Y. Song, D. Zhu, Blockchainenabled crossdomain object detection for autonomous driving: a model sharing approach. IEEE Internet Things J. 7, 3681–3692 (2020)
M. Kumru, E. Ozkan, 3D extended object tracking using recursive Gaussian processes, in 21st International Conference on Information Fusion (2018), pp. 1–8. https://doi.org/10.23919/ICIF.2018.8455480
M. Kumru, E. Ozkan, Threedimensional extended object tracking and shape learning using Gaussian processes. IEEE Trans. Aerosp. Electron. Syst. 57, 2795–2814 (2021). https://doi.org/10.1109/TAES.2021.3067668
K. Granstrom, M. Baum, S. Reuter, Extended object tracking: introduction, overview, and applications. J. Adv. Inf. Fusion 12, 139–174 (2017)
J. Vermaak, N. Ikoma, S. Godsill, Sequential Monte Carlo framework for extended object tracking. IEEE Proc. Radar Sonar Navig. 152, 353–363 (2005). https://doi.org/10.1049/iprsn:20045044
L. Mihaylova, A.Y. Carmi, F. Septier, A. Gning, S.K. Pang, S. Godsill, Overview of Bayesian sequential Monte Carlo methods for group and extended object tracking. Digital Signal Process. 25, 1–16 (2014)
J.W. Koch, Bayesian approach to extended object and cluster tracking using random matrices. IEEE Trans. Aerosp. Electron. Syst. 44, 1042–1059 (2008). https://doi.org/10.1109/TAES.2008.4655362
J. Lan, X.R. Li, Tracking of maneuvering nonellipsoidal extended object or target group using random matrix. IEEE Trans. Signal Process. 62, 2450–2463 (2014). https://doi.org/10.1109/TSP.2014.2309561
J. Lan, X.R. Li, Tracking of extended object or target group using random matrix: new model and approach. IEEE Trans. Aerosp. Electron. Syst. 52, 2973–2989 (2016). https://doi.org/10.1109/TAES.2016.130346
N.J. Bartlett, C. Renton, A.G. Wills, A closedform prediction update for extended target tracking using random matrices. IEEE Trans. Signal Process. 68, 2404–2418 (2020). https://doi.org/10.1109/TSP.2020.2984390
P. Hoher, S. Wirtensohn, T. Baur, J. Reuter, F. Govaers, W. Koch, Extended target tracking with a lidar sensor using random matrices and a virtual measurement model. IEEE Trans. Signal Process. 70, 228–239 (2022)
M. Baum, B. Noack, U.D. Hanebeck, Extended object and group tracking with elliptic random hypersurface models, in 13th International Conference on Information Fusion (2010), pp. 1–8. https://doi.org/10.1109/ICIF.2010.5711854
M. Baum, U.D. Hanebeck, Extended object tracking with random hypersurface models. IEEE Trans. Aerosp. Electron. Syst. 50, 149–159 (2014). https://doi.org/10.1109/TAES.2013.120107
X. Zhang, Z. Yan, Y. Chen, Y. Yuan, A novel particle filter for extended target tracking with random hypersurface model. Appl. Math. Comput. 425, 127081 (2022). https://doi.org/10.1016/j.amc.2022.127081
N. Wahlstrom, E. ozkan, Extended target tracking using Gaussian processes. IEEE Trans. Signal Process. 63, 4165–4178 (2015). https://doi.org/10.1109/TSP.2015.2424194
W. Aftab, R. Hostettler, A. De Freitas, M. Arvaneh, L. Mihaylova, Spatiotemporal Gaussian process models for extended and group object tracking with irregular shapes. IEEE Trans. Veh. Technol. 68, 2137–2151 (2019). https://doi.org/10.1109/TVT.2019.2891006
C.K. Williams, C.E. Rasmussen, Gaussian Processes for Machine Learning, vol. 2 (MIT Press, Cambridge, 2006)
T. Hirscher, A. Scheel, S. Reuter, K. Dietmayer, Multiple extended object tracking using gaussian processes, in 19th International Conference on Information Fusion (FUSION) (2016), pp. 868–875
Y. Guo, Y. Li, R. Tharmarasa, T. Kirubarajan, M. Efe, B. Sarikaya, GPPDA filter for extended target tracking with measurement origin uncertainty. IEEE Trans. Aerosp. Electron. Syst. 55, 1725–1742 (2019). https://doi.org/10.1109/TAES.2018.2875555
J.E. Johnson, V. Laparra, G. CampsValls, Accounting for input noise in Gaussian process parameter retrieval. IEEE Geosci. Remote Sens. Lett. 17, 391–395 (2020). https://doi.org/10.1109/LGRS.2019.2921476
K. Kersting, C. Plagemann, P. Pfaff, W. Burgard, Mostlikely heteroscedastic Gaussian process regression. 227, 393–400 (2007). https://doi.org/10.1145/1273496.1273546
M. LazaroGredilla, M. Titsias, Variational heteroscedastic Gaussian process regression, in Proceedings of the 28th International Conference on Machine Learning (2011), pp. 841–848. https://doi.org/10.5555/3104482.3104588
L. MunozGonzalez, M. LazaroGredilla, A.R. FigueirasVidal, Laplace approximation for divisive Gaussian processes for nonstationary regression. IEEE Trans. Pattern Anal. Mach. Intell. 38, 618–624 (2016). https://doi.org/10.1109/TPAMI.2015.2452914
A. Girard, C. Rasmussen, R. MurraySmith, Gaussian Process Priors with Uncertain Inputs: MultipleStepAhead Prediction (School of Computing Science, 2002)
A. Girard, C. E. Rasmussen, J. Q. Candela, R. MurraySmith, Gaussian process priors with uncertain inputs application to multiplestep ahead time series forecasting, in Proceedings of the 15th International Conference on Neural Information Processing Systems (2002), pp. 545–552. https://doi.org/10.5555/2968618.2968686
J.Q. Candela, A. Girard, C. Rasmussen, Prediction at an uncertain input for Gaussian processes and relevance vector machines application to multiplestep ahead timeseries forecasting. Technical University of Denmark (2003)
A. Girard, R. MurraySmith, Learning a Gaussian Process Model with Uncertain Inputs (Department of Computing Science, University of Glasgow, Glasgow, 2003)
M.P. Deisenroth, M.F. Huber, U.D. Hanebeck, Analytic momentbased Gaussian process filtering, in Proceedings of the 26th Annual International Conference on Machine Learning (2009), pp. 225–232. https://doi.org/10.1145/1553374.1553403
A. Mchutchon, C. Rasmussen, Gaussian process training with input noise. Adv. Neural. Inf. Process. Syst. 24, 1341–1349 (2011). https://doi.org/10.5555/2986459.2986609
M. Longbin, S. Xiaoquan, Z. Yiyu, S.Z. Kang, Y. BarShalom, Unbiased converted measurements for tracking. IEEE Trans. Aerosp. Electron. Syst. 34, 1023–1027 (1998). https://doi.org/10.1109/7.705921
S.J. Julier, J.K. Uhlmann, New extension of the Kalman filter to nonlinear systems. Signal Process. Sens. Fusion Target Recogn. VI 3068, 182–193 (1997). https://doi.org/10.1117/12.280797
E. Wan, R. Van Der Merwe, The unscented Kalman filter for nonlinear estimation, in Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium (2000), pp. 153–158. https://doi.org/10.1109/ASSPCC.2000.882463
C. Jauffret, Y. BarShalom, Track formation with bearing and frequency measurements in clutter. IEEE Trans. Aerosp. Electron. Syst. 26, 999–1010 (1990). https://doi.org/10.1109/7.62252
M. David, Introduction to Monte Carlo Methods. Learning in Graphical Models (1998), pp. 175–204
Acknowledgements
Not applicable.
Author information's
Yunfei Guo received the B.S. degree in electrical engineering from Yanshan University, China, in 2002, and the Ph.D. degree in electrical engineering from Zhejiang University, China, in 2007. From 2014 to 2015, he was a Visiting Professor with McMaster University, Canada. He is currently a Professor with the School of Automation, Hangzhou Dianzi University, Hangzhou. He is also serving as the Deputy Secretary General with the China Information Fusion Society. His research interests include target tracking, information fusion, and Gaussian processes.
Dongsheng Yang received the B.S. and M.S. degrees in automation from Anhui Polytechnic University, Wuhu, China, and Shenyang university of chemical technique, Shenyang, China, in 2018 and 2021, respectively, and he is currently working toward the Ph.D. degree with the School of Automation in Hangzhou Dianzi University. His research interests include extended target tracking and information fusion.
Lei Ren received the B.S. degree in automation in 2019 from Hangzhou Dianzi University, Hangzhou, China, and he is currently working toward the Postgraduate degree with the School of Automation. His research interests include extended target tracking.
Lei Yan received the B.S. degree in automation in 2020 from Hangzhou Dianzi University, Hangzhou, China, and he is currently working toward the Postgraduate degree with the School of Automation. His research interests include target tracking.
Funding
Supported by Zhejiang Provincial Natural Science Foundation of China (Grant No. LZ20F010002), and by Zhejiang Provincial Science and Technology Project (Grant No. 2022C01095).
Author information
Authors and Affiliations
Contributions
YG was in charge of the whole trial; DY wrote the manuscript; LR and LY assisted with sampling and laboratory analyses. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix 1
Unbiased converted measurement
Assume that \(c_{k,j}=[d_{k,j},\beta _{k,j}]\) is the polar measurements at time k with covariance \(R^{\text {pol}}={\text {diag}} \{\sigma _d^2,\sigma _{\beta }^{2}\}\). Using the unbiased converted measurement (UCM) technique [37], the converted measurements \(c_{k,j}^{{\mathbf {u}}}\) can be given as
where \(c_{k,j}^{{\mathbf {u}},\xi }\) and \(c_{k,j}^{{\mathbf {u}},\eta }\) are the xaxis and yaxis position of measurements in Cartesian coordinates, respectively. And the term \(\lambda _\beta =\exp {(\sigma _\beta ^2/2)}\) is the bias compensation factor [37]. The converted covariance \(R_{k,j}^{{\mathbf {u}}}\) is given by
where
Appendix 2
Sigma sampling
Assume that the GP input is a ndimensional variable with distribution \({\hat{\theta }}_{k,j}^{L}\sim {\mathcal {N}}(\mu _{\theta _{k,j}^{L}},\Sigma _{\theta _{k,j}^{L}})\) (dimension n). The Sigma sampling points \({\hat{\theta }}_{k,j}^{L,(i)}\) are given as [38]
The mean weights \(\omega _{k}^{m,(i)}\) and the covariance weights \(\omega _{k}^{c,(0)}\) are given as
where \(\lambda =\alpha ^2(n+\varsigma )n\) is a scaling parameter, \(\alpha\) denotes the spread of the sigma points around the input mean value, \(\varsigma\) is a secondary scaling parameter, and \(\varpi =2\) is optimal for Gaussian distribution [38].
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Guo, Y., Yang, D., Ren, L. et al. Gaussian processes based extended target tracking in polar coordinates with input uncertainty. EURASIP J. Adv. Signal Process. 2022, 106 (2022). https://doi.org/10.1186/s1363402200940w
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s1363402200940w
Keywords
 Extended target tracking
 Gaussian processes
 Input uncertainty
 Nonlinear measurement
 Measurement origin uncertainty