A two-stage grasp detection method for sequential robotic grasping in stacking scenarios

Jing Zhang; Baoqun Yin; Yu Zhong; Qiang Wei; Jia Zhao; Hazrat Bilal; Jing Zhang; Baoqun Yin; Yu Zhong; Qiang Wei; Jia Zhao; Hazrat Bilal

doi:10.3934/mbe.2024152

Mathematical Biosciences and Engineering

2024, Volume 21, Issue 2: 3448-3472. doi: 10.3934/mbe.2024152

Previous Article Next Article

Research article Special Issues

A two-stage grasp detection method for sequential robotic grasping in stacking scenarios

1.
Department of Automation, University of Science and Technology of China, Hefei 230027, China
2.
School of Information Engineering, Southwest University of Science and Technology, Mianyang 621010, China
3.
The 14th Research Institute of China Electronics Technology Group Corporation, Nanjing 210039, China

Academic Editor: Jorge Bernardino

Received: 18 October 2023 Revised: 09 January 2024 Accepted: 18 January 2024 Published: 05 February 2024

Dexterous grasping is essential for the fine manipulation tasks of intelligent robots; however, its application in stacking scenarios remains a challenge. In this study, we aimed to propose a two-phase approach for grasp detection of sequential robotic grasping, specifically for application in stacking scenarios. In the initial phase, a rotated-YOLOv3 (R-YOLOv3) model was designed to efficiently detect the category and position of the top-layer object, facilitating the detection of stacked objects. Subsequently, a stacked scenario dataset with only the top-level objects annotated was built for training and testing the R-YOLOv3 network. In the next phase, a G-ResNet50 model was developed to enhance grasping accuracy by finding the most suitable pose for grasping the uppermost object in various stacking scenarios. Ultimately, a robot was directed to successfully execute the task of sequentially grasping the stacked objects. The proposed methodology demonstrated the average grasping prediction success rate of 96.60% as observed in the Cornell grasping dataset. The results of the 280 real-world grasping experiments, conducted in stacked scenarios, revealed that the robot achieved a maximum grasping success rate of 95.00%, with an average handling grasping success rate of 83.93%. The experimental findings demonstrated the efficacy and competitiveness of the proposed approach in successfully executing grasping tasks within complex multi-object stacked environments.

Keywords:

Citation: Jing Zhang, Baoqun Yin, Yu Zhong, Qiang Wei, Jia Zhao, Hazrat Bilal. A two-stage grasp detection method for sequential robotic grasping in stacking scenarios[J]. Mathematical Biosciences and Engineering, 2024, 21(2): 3448-3472. doi: 10.3934/mbe.2024152

Related Papers:

[1]	Ardiansyah, Fauziyyah Ariffa, Rizki Maryam Astuti, Wahyudi David, Dody Dwi Handoko, Slamet Budijanto, Hitoshi Shirakawa . Non-volatile compounds and blood pressure-lowering activity of Inpari 30 and Cempo Ireng fermented and non-fermented rice bran. AIMS Agriculture and Food, 2021, 6(1): 337-359. doi: 10.3934/agrfood.2021021
[2]	Santi Noviasari, Feri Kusnandar, Agus Setiyono, Slamet Budijanto . Antioxidant activity and inhibition of α-amylase and α-glucosidase in fermented black rice bran-based analog rice. AIMS Agriculture and Food, 2022, 7(1): 61-72. doi: 10.3934/agrfood.2022004
[3]	Budi Suarti, Sukarno, Ardiansyah, Slamet Budijanto . Bio-active compounds, their antioxidant activities, and the physicochemical and pasting properties of both pigmented and non-pigmented fermented de-husked rice flour. AIMS Agriculture and Food, 2021, 6(1): 49-64. doi: 10.3934/agrfood.2021004
[4]	Diana Lo, Andreas Romulo, Jia-Ying Lin, Yuh-Tai Wang, Christofora Hanny Wijaya, Ming-Chang Wu . Effect of different fermentation conditions on antioxidant capacity and isoflavones content of soy tempeh. AIMS Agriculture and Food, 2022, 7(3): 567-579. doi: 10.3934/agrfood.2022035
[5]	Chiao Ying Huang, Gerald L. Riskowski, Jennifer Chang, Ching Hsuan Lin, Jinn Tsyy Lai, Audrey Chingzu Chang . Pecan shell by-products—phenolic compound contents and antimicrobial properties. AIMS Agriculture and Food, 2020, 5(2): 218-232. doi: 10.3934/agrfood.2020.2.218
[6]	Pattarabhorn Pakaweerachat, Worasaung Klinthong, Kazuhisa Ohtaguchi, Teerin Chysirichote . Simultaneous isoquercitin and gallic acid production of Aspergillus niger on Triphala byproduct under solid state fermentation in packed-bed bioreactor. AIMS Agriculture and Food, 2023, 8(2): 359-373. doi: 10.3934/agrfood.2023020
[7]	Zina Guermazi, Cinzia Benincasa . Olive pomace as spreadable pulp: A new product for human consumption. AIMS Agriculture and Food, 2018, 3(4): 441-454. doi: 10.3934/agrfood.2018.4.441
[8]	Teti Estiasih, Jatmiko Eko Witoyo, Khofifah Putri Wulandari, Fadhillah Dwi Juniati, Widiastuti Setyaningsih, Hanifah Nuryani Lioe, Miguel Palma, Kgs Ahmadi, Hamidie Ronald Daniel Ray, Elya Mufidah . Stability comparison of conventional and foam-mat red and purple dried roselle calyces powder as a function of pH. AIMS Agriculture and Food, 2025, 10(1): 177-198. doi: 10.3934/agrfood.2025010
[9]	Dahye Kim, Kyung Dong Lee, Ki Choon Choi . Role of LAB in silage fermentation: Effect on nutritional quality and organic acid production—An overview. AIMS Agriculture and Food, 2021, 6(1): 216-234. doi: 10.3934/agrfood.2021014
[10]	Samsul Rizal, Maria Erna Kustyawati, Murhadi, Udin Hasanudin, Subeki . The effect of inoculum types on microbial growth, β-glucan formation and antioxidant activity during tempe fermentation. AIMS Agriculture and Food, 2022, 7(2): 370-386. doi: 10.3934/agrfood.2022024

Abstract

1. Introduction

Fractional differential equations have gained growing attention in recent years and many monographs have appeared ^[1,2]. The most common definitions of fractional calculus (differentiation and integration) are for Caputo, Riemann-Liouville, and Grunwald-Letnikov derivatives ^[3,4,5,6,7]. Compared with these two types of definitions, the Hadamard fractional calculus, which was first introduced in 1892 by Hadamard ^[8], did not receive much attention. The kernel of the integrand in the definition of the fractional Hadamard derivative contains a logarithmic function with an arbitrary exponent different from the Riemann-Liouville fractional derivatives. Recently, the Hadamard derivative and Hadamard-type fractional differential equations have been useful in practical problems related to mechanics and engineering, such as fracture analysis or both planar and three-dimensional elasticities ^[9]. Kilbas discussed Hadamard-type fractional differential equations in different spaces ^[10]. Recently, Ma and Li described the properties of Hadamard calculus ^[11] and they also proposed the definite conditions for Hadamard-type fractional differential equations.

The Caputo-Hadamard (C-H) fractional derivative is a kind of fractional derivative that is useful in describing abnormal diffusion processes, especially ultra-slow diffusion. Gohar et al. ^[12] studied the existence and uniqueness of the solution to Caputo-Hadamard fractional differential equations and the corresponding continuation theorem. Wang et al. ^[13] investigated the stability of the zero solution of a class of nonlinear Hadamard-type fractional differential systems by utilizing a new fractional comparison principle. Belbali et al. ^[14] discussed the existence, uniqueness, and stability of solutions for a nonlinear fractional differential system consisting of a nonlinear Caputo-Hadamard fractional initial value problem. Aljoudi et al. ^[15] studied a coupled system of Caputo-Hadamard-type sequential fractional differential equations supplemented with nonlocal boundary conditions involving Hadamard fractional integrals. Dhaniya et al. ^[16] established the existence, uniqueness, and Hyers-Ulam stability of the solution to the nonlinear Langevin fractional differential equation that involves the C-H and Caputo fractional operators, with nonperiodic and nonlocal integral boundary conditions. Beyenea et al. ^[17] established sufficient conditions for the existence and uniqueness of solutions to nonlinear Caputo-Hadamard fractional differential equations involving Hadamard integrals and unbounded delays. He et al. ^[18] considered the Hadamard and the Caputo-Hadamard fractional derivatives and the stability of related systems without and with delay.

Due to the complex form of C-H fractional operators, one often needs to find a suitable numerical scheme to approximate it, which greatly improves the efficiency of the actual calculation process. The studies on numerical methods for nonlinear C-H fractional differential equations are still in their early stages. Gohar ^[19] studied finite difference methods for fractional differential equations with C-H derivatives and investigated the smoothness properties of the solution. Li et al. ^[20] obtained the analytical solution to a certain linear fractional partial differential equation with the C-H fractional derivative by introducing a new modified Laplace transform, and derived a numerical algorithm for such kinds of equations. Fan et al. ^[21] proposed three kinds of numerical formulas for approximating the C-H fractional derivatives, which are called $L1-2$ formula, $L2-1_\sigma$ formula, and H2N2 formula.

Most numerical methods for solving fractional differential equations are based on local difference schemes. Compared with the previous works, the main contribution of this paper is to extend the results in ^[22,23] by constructing and analyzing a nonlocal spectral collocation method for the following system of fractional pantograph delay differential equations:

$\begin{equation} \begin{cases} {}_\ell^{CH} D_t^\rho X_1(t) = g_1(t, X_1(t), \ldots, X_M(t), X_1(qt), \ldots, X_M(qt)), \ t \in I, \\ {}_\ell^{CH} D_t^\rho X_2(t) = g_2(t, X_1(t), \ldots, X_M(t), X_1(qt), \ldots, X_M(qt)), \ t \in I, \\ \vdots\\ {}_\ell^{CH} D_t^\rho X_M(t) = g_M(t, X_1(t), \ldots, X_M(t), X_1(qt), \ldots, X_M(qt)), \ t \in I, \\ X_i(t) = \bar X_{i}(t), \ for \ qt \leq \ell \ i = 1, 2, \ldots, M, \quad \quad \ell \in (0, t), \ \rho \in (0, 1), q\in(0, 1), \end{cases} \end{equation}$

(1.1)

where $g_i: I \times \mathbb{R}^M \to \mathbb{R}$ are given continuous functions, $I = (\ell, e \ell)$ , and the C-H derivative ${}_\ell^{CH} D_t^\rho$ of order $0 < \rho < 1$ is given by (2.2).

The outline of this paper is as follows: In Section 2, we introduce some necessary definitions and preliminaries. In Section 3, we construct the spectral collocation scheme. In Section 4, we provide some auxiliary lemmas. The convergence analysis is discussed in Section 5. The numerical results are provided in Section 6.

2. Preliminaries

In this section, some relevant properties of the C-H fractional calculus and the logarithmic Jacobi (log J) approximation are presented.

Definition 2.1. The C-H fractional integral with order $\rho > 0$ is defined as ^[24]

$\begin{equation} {}_{\ell}J_{ z}^{\rho} X (z) = \frac{1}{\Gamma(\rho)} \int_{\ell}^{z}\kappa^{\rho-1}(z, w) X(w) \frac{d w}{w}, \; z > \ell > 0, \end{equation}$

(2.1)

where $\kappa(z, w) = \log(\frac{z}{w}).$

Definition 2.2. The C-H fractional differential operator of order $0 < \rho < 1$ is given as ^[1]

$\begin{equation} {}_\ell^{CH} D_z^\rho X(z) = \frac{1}{{\Gamma (1 - \rho )}}\int_\ell^z {\kappa^{ - \rho }(z, w)\, X'(w)dw}. \end{equation}$

(2.2)

Definition 2.3. Let $\rho, \eta > -1,$ $I: = [\ell, \ell \, e],$ and $\ell > 0$ . The log J functions of order $p$ are given by ^[23]

$\begin{equation} \begin{split} \mathscr{P}^{\rho , \eta , \ell }_{p}(z)& = \mathscr{P}^{\rho , \eta }_{p}\left(\kappa^2(z, \ell) -1\right) \quad (\eta, \rho > -1, \ \ell > 0, \ \forall z \in I)\\ & = \frac{\Gamma (p+\rho +1)}{p!\Gamma (p+\rho +\eta +1)}\sum _{k = 0}^{p}\left( {\begin{array}{c}p\\ k\end{array}}\right) \frac{\Gamma (p+k+\rho +\eta +1)}{\Gamma (k+\rho +1)}\left(\kappa(z, \ell)-1\right)^{k}, \end{split} \end{equation}$

(2.3)

where $\mathscr{P}^{\rho, \eta }_{p}(z)$ is the Jacobi polynomial and it is defined as

$\begin{equation*} \mathscr{P}^{\rho , \eta }_{p}(z) = \frac{\Gamma (p+\rho +1)}{\Gamma (p+1+\rho +\eta )p!}\sum \limits_{k = 0}^{p}\left( {\begin{array}{c}p\\ k\end{array}}\right) \frac{\Gamma (p+k+\rho +\eta +1)}{\Gamma (k+\rho +1)}\left (\frac{z-1}{2}\right )^{k}. \end{equation*}$

We define the space of logarithmic functions of order $s$ by

$P^{\log }_{s}( \Omega): = span\left\{1, \kappa(z, \ell), \kappa(z, \ell)^{2 }, \ldots , \kappa(z, \ell)^{s } \right\},$

where $\Omega = [\ell, +\infty), \ \ell > 0.$ Let

$\begin{equation} \chi ^{\rho , \eta , \ell }(z): = z^{-1}\kappa(z, \ell)^{\eta }\left (1-\kappa(z, \ell)\right)^{\rho }. \end{equation}$

(2.4)

We denote by $L^2_{\chi^{\rho, \eta, \ell }}(I)$ the weighted $L^2$ space with the following inner product and norm:

$\begin{equation} \left( {X , \phi } \right)_{\chi ^{\rho , \eta , \ell } } = \int_I {X (z)\phi (z)\chi ^{\rho , \eta , \ell }(z) dz}, \qquad \left\| X \right\|_{\chi ^{\rho , \eta , \ell } } = \left( {X , X } \right)_{\chi ^{\rho , \eta , \ell } }^{1/2} . \end{equation}$

(2.5)

One of the most important properties of the log J polynomials is that they are mutually orthogonal in $L^2_{\chi^{\rho, \eta, \ell }}(I)$ , i.e.,

$\begin{equation} \begin{split} & \left( { \mathscr{P}^{\rho , \eta , \ell }_{m}(z), \mathscr{P}^{\rho , \eta , \ell }_{j}(z)} \right)_{\chi ^{\rho , \eta , \ell } } = 0, \quad \forall j\neq m, \\ & \left\| \mathscr{P}^{\rho , \eta , \ell }_{j}(z) \right\|_{\chi ^{\rho , \eta , \ell } } = \widehat{\theta}_{j}^{\rho , \eta } = \frac{\Gamma (j+\rho +1)\Gamma (j+\eta +1)}{(2j+\rho +\eta +1)j!\Gamma (j+\rho +\eta +1)}. \end{split} \end{equation}$

(2.6)

We define the following first-order differential operator:

$\begin{equation} D^{1}_{\log }\phi(z) = \frac{d}{d \kappa(z, \ell)}\phi(z) = z\phi'(z), \end{equation}$

(2.7)

and an induction leads to

$\begin{equation} D^{k}_{\log }\phi(z) = \overbrace{D^{1}_{\log }\cdot D^{1}_{\log }\cdots D^{1}_{\log }}^{k}\phi(z). \end{equation}$

(2.8)

We also define the non-uniformly weighted log J Sobolev space as

$\begin{equation*} B^{i, \ell }_{\rho , \eta }(I): = \big \{\phi: D^{j}_{\log }\phi\in L^{2}_{ \chi ^{\rho +j, \eta +j, \ell }}(I), \, 0\le j\le i\big \}, \; \; \; \; \; i\in \mathbb {N}, \end{equation*}$

with

$\begin{equation*} \begin{split} (\psi, \phi)_{B^{i, \ell }_{\rho , \eta }} = & {} \sum _{k = 0}^{i}(D^{k}_{\log }\psi, D^{k}_{\log }\phi)_{ \chi ^{\rho +k, \eta +k, \ell }}, \; \; \Vert \phi\Vert _{B^{i, \ell }_{\rho , \eta }} = (\phi, \phi)^{1/2}_{B^{i, \ell }_{\rho , \eta }}, \nonumber \\|\phi|_{B^{i, \ell }_{\rho , \eta }} = & {} \Vert D^{i}_{\log }\phi\Vert _{ \chi ^{\rho +i, \eta +i, \ell }}.\end{split} \end{equation*}$

For the usual shifted-weighted Jacobi Sobolev space, we define

$\begin{equation*} B^{i}_{\rho , \eta }(\Lambda): = \big \{\phi: \partial ^{j}_{z}\phi\in L^{2}_{ \chi ^{\rho +j, \eta +j}}(\Lambda), 0\le j\le i\big \}, \; \; \; \; \; i\in \mathbb {N}, \end{equation*}$

where $\chi ^{\rho, \eta } = (-z+1)^{\rho }z^{\eta}$ with $z \in \Lambda = [0, 1]$ is the classical Jacobi weight function.

Assume that $x_{0} < x_{1} < \cdots < x_{M-1} < x_{M}$ in $I$ are the roots of $\mathscr{P}^{\rho, \eta, \ell }_{M+1}(x).$ Let $z(x) = \log\frac{x}{\ell}$ . Then $z_{j}: = z(x_{j}) = \log \frac{ x_{j}}{\ell}, \; \; 0\le j\le M$ , are zeros of $\mathscr{P}^{\rho, \eta}_{M+1}(x),$ and $\{ \chi _{i}\}_{i = 0}^{M}$ are the corresponding weights.

The log J-Gauss quadrature enjoys the exactness

$\begin{equation} \int_I {X (z)\chi ^{\rho , \eta , \ell } (z)dz} = \sum\limits_{i = 0}^M {X (z_i )\chi _i }, \quad \forall X(z) \in P^{\log }_{2M+1}. \end{equation}$

(2.9)

Hence,

$\begin{equation} \sum\limits_{k = 0}^{M} \mathscr{P}^{\rho , \eta , \ell }_{q}(z_{k}) \mathscr{P}^{\rho , \eta , \ell }_{j}(z_{k}) \chi_{k} = \widehat{\theta}_{q}^{\rho , \eta } \delta _{q, j}, \quad \forall \ 0 \leq q+j \leq 2M+1. \end{equation}$

(2.10)

For any $X(\ell e^z) \in C(I)$ , the log J-Gauss interpolation operator $I^{\rho, \eta, \ell }_{z, M}:C(I)\longrightarrow P^{\log }_{M}$ is determined uniquely by

$\begin{equation} I^{\rho , \eta, \ell }_{z, M} X(z_{q}) = X(z_{q}), \quad 0\leq q \leq M. \end{equation}$

(2.11)

From the above condition, we have $I^{\rho, \eta, \ell }_{z, M}X = X$ for all $X \in P^{\log }_{M}$ . On the other hand, since $I^{\rho, \eta, \ell }_{z, M}X \in P^{\log }_{M}$ , we can write

$\begin{equation} \begin{split} I^{\rho , \eta, \ell }_{z, M }X(x)& = \sum\limits_{i = 0}^{M} \widehat{X}_{i}^{\rho , \eta, \ell } \mathscr{P}^{\rho , \eta , \ell }_{i} (x), \quad \widehat{X}_{i}^{\rho , \eta, \ell } = \frac{1}{\widehat{\theta}_{i}^{\rho , \eta}} \sum\limits_{j = 0}^{M} X (x_j) \mathscr{P}^{\rho , \eta , \ell }_{i} (x_j) \chi_{j}, \ \forall X \in P^{\log }_{M}( I). \end{split} \end{equation}$

(2.12)

The $L^\infty$ (I) space is the set of all measurable functions that are essentially bounded. That is, functions $g$ that are bounded almost everywhere on a set of finite measures. The essential supermom norm is used to define the norm of this space and is given as

$\left\| g \right\|_\infty = \mathop {ess\ sup}\limits_{x \in I} \left| {g(x)} \right|.$

Definition 2.4. Let $\textbf{A}(z) = \left(a_{ij}(z)\right)_{m \times n}$ be an $(m \times n)$ matrix function with $z \in I.$ We consider the non-negative real-valued function

$\begin{equation} \left|\mathbf{A}(z) \right| = \sum\limits_{i = 1}^{m}\sum\limits_{j = 1}^{n} \left|a_{ij}(z)\right|, \end{equation}$

(2.13)

and the norms

$\begin{equation} \begin{split} &\left\|\textbf{A} \right\|_{\chi ^{\rho , \eta , \ell }}: = \left(\int_I \left|\textbf{A}(z) \right|^2 \chi ^{\rho , \eta , \ell } dz\right)^{1/2}, \\ &\left\|\textbf{A} \right\|_\infty: = ess \sup\limits_{z\in I} \left|\textbf{A}(z)\right|. \end{split} \end{equation}$

(2.14)

Proposition 2.1. It holds for any $\psi(\ell\, e^x)\in B^{m}_{\rho, \eta }(\Lambda),$ $m\ge 1$ and $M+1 \ge m \ge q \ge 0$

$\begin{equation} \Vert D^{q}_{\log }\big (\psi-I^{\rho , \eta, \ell} _{M }\psi\big )\Vert _{ \chi ^{\rho +q, \eta +q, \ell }} \le c\sqrt{\frac{(1+M-m)!}{M!}}M^{q-(1+m)/2}\Vert \partial _{x}^{m}\big \{\psi(\ell\, e^x)\big \}\Vert _{ \chi ^{\rho +m, \eta +m}}, \end{equation}$

(2.15)

and it takes the form

$\begin{equation} \Vert D^{q}_{\log }\big (\psi-I^{\rho , \eta, \ell} _{M }\psi\big )\Vert _{ \chi ^{\rho +q, \eta +q, \ell }} \le cM^{q-m}\Vert \partial _{x}^{m}\big \{\psi(\ell\, e^x)\big \}\Vert _{ \chi ^{\rho +m, \eta +m}}, \ c\approx 1, \quad \mathit{\text{for fixed}} \quad m \quad \mathit{\text{and}} \quad M\gg 1. \end{equation}$

(2.16)

In the case of $q = 0, 1$ , we can write

$\begin{equation} \Vert \psi-I^{\rho , \eta , \ell }_{M }\psi\Vert _{ \chi ^{\rho , \eta , \ell }}\le cM^{-m}\Vert \partial _{x}^{m}\big \{\psi(\ell\, e^x)\big \}\Vert _{ \chi ^{\rho +m, \eta +m}}, \end{equation}$

(2.17)

$\begin{equation} \Vert \partial _{x}(\psi-I^{\rho , \eta, \ell }_{M }\psi)\Vert _{\tilde{ \chi }^{\rho , \eta , \ell }}\le cM^{1-m}\Vert \partial _{x}^{m}\big \{\psi(\ell\, e^x)\big \}\Vert _{ \chi ^{\rho +m, \eta +m}}, \end{equation}$

(2.18)

where $\tilde{ \chi }^{\rho, \eta, \ell } = x\left (1-\log\left(\frac{x}{\ell}\right)\right)^{\rho +1 }\left(\log\left(\frac{x}{\ell}\right)\right)^{\eta +1}$ .

Lemma 2.1. ^[23] For any $\rho, \eta \in (-1, -\frac{1}{2})$ and for all $\psi(x)\in B^{1, \ell }_{\rho, \eta }(I),$ $\psi(\xi) = 0$ for some $\xi \in I$ , it holds

$\begin{equation} \Vert \psi\Vert _{\infty }\le \sqrt{2}\Vert \partial _{x}\psi\Vert ^{1/2}_{\tilde{ \chi }^{\rho , \eta , \ell }} \Vert \psi\Vert ^{1/2}_{ \chi ^{\rho , \eta , \ell }}. \end{equation}$

(2.19)

Proposition 2.2. ^[23] For $\rho, \eta \in (-1, -\dfrac{1}{2}]$ ,

$\begin{equation} \Vert \psi-I_{M }^{\rho , \eta, \ell }\psi\Vert _{\infty }\le c M^{1/2-m}\Vert \partial ^{m}_{x}\psi(\ell\, e^x)\Vert _{ \chi ^{\rho +m, \eta +m}}, \; \; \forall \psi(\ell\, e^x)\in B^{m}_{\rho , \eta }(\Lambda), \ m\ge 1. \end{equation}$

(2.20)

Lemma 2.2. ^[23]

$\begin{equation} \Vert I^{\rho , \eta, \ell }_{M }\Vert _{\infty } : = \max _{x\in I}\sum \limits_{j = 0}^{M}|h_{j }^{\rho , \eta, \ell }(x)| = {\left\{ \begin{array}{ll} O(\log M), \; \; \; -1 < \rho , \eta \le -\frac{1}{2}, \\ O(M^{\mu +\frac{1}{2}}), \; \; \; \mu = \max (\rho , \eta ), \ otherwise, \\ \end{array}\right. } \end{equation}$

(2.21)

where $\{h^{\rho, \eta, \ell }_{j }(x)\}_{j = 0}^{M}$ are the logarithmic Lagrange interpolation functions that are related to $\mathscr{P}^{\rho, \eta, \ell }_{M+1}(x)$ .

3. Non-polynomial spectral collocation scheme

To begin with, we rewrite the differential equation (1.1) in the following equivalent compact integral form:

$\begin{equation} Z (t) = Z _\ell + \frac{1}{{\Gamma (\rho )}}\int_\ell^t {\left( {\kappa(t, s)} \right)^{\rho - 1} \textbf{Q}\left( {s, Z (s), Z (q\, s)} \right)\frac{{ds}}{s}}, \ t\in (\ell, e \ell], \end{equation}$

(3.1)

where

$\begin{array}{l} Z (t) = \left[ {X _1(t) , X _2(t) , \ldots , X _M(t) } \right]^T, \\ Z _\ell = \left[ {\bar X_1(\ell) , \bar X_2(\ell) , \ldots , \bar X_M(\ell) } \right]^T, \\ Z (q\, t) = \left\{ {\begin{array}{*{20}c} { \left[ {X _1(q\, t) , X _2(q\, t) , \ldots , X _M (q\, t)} \right]^T, \quad if \quad q\, t > \ell, } \\ \\ { \left[ {\bar X _1(q\, t) , \bar X _2(q\, t) , \ldots , \bar X _M (q\, t)} \right]^T, \quad if \quad q\, t\leq \ell, } \\ \end{array}} \right. \\ \textbf{Q}(t) = \left[ {g_1 , g_2 , \ldots , g_M } \right]^T. \\ \end{array}$

In the following, we will make some useful transformations, which in turn are the basis for the numerical solution scheme and its numerical analysis. In order to convert the integral interval $\left(\ell, t\right)$ to $I$ , we consider

$\kappa(s, \ell) = \kappa(t, \ell)\kappa(r, \ell),$

$s = s(t, r) = \ell \left( {\frac{r}{\ell}} \right)^{\kappa(t, \ell)}.$

Hence, the system(3.1) becomes

$\begin{equation} \begin{split} Z(t) & = Z _\ell + \frac{{\left( {\kappa(t, \ell)} \right)^{\rho } }}{{\Gamma (\rho )}}\int_I {\left( {1 - \kappa(r, \ell)} \right)^{\rho - 1} \textbf {G}(s(t, r), Z (s(t, r)), Z (q\, s(t, r)))\frac{{dr}}{r}}.\\ \end{split} \end{equation}$

(3.2)

The non-polynomial spectral collocation scheme for (3.2) is to find $X_{m, N} (t) \in P^{\log }_{N}(I), \ m = 1, 2, \ldots, M$ such that

$\begin{equation} \begin{split} Z _N(t) = Z _\ell + \frac{1}{{\Gamma (\rho )}}I_{t, N}^{0, 0, \ell } {\left( {\kappa(t, \ell)} \right)^\rho }\int_I {r^{-1}\left( {1 - \kappa(r, \ell)} \right)^{\rho - 1} I_{r, N}^{\rho - 1, 0, \ell } \textbf{Q}(s(t, r), Z _N(s(t, r)), Z _N(q\, s(t, r)))dr}, \end{split} \end{equation}$

(3.3)

where

$Z_N (t) = \left[ {X _{1, N} , X _{2, N} , \ldots , X _{M, N} } \right]^T,$

and $I_{z, N}^{\rho, \eta, \ell }$ the log J-Gauss interpolation operator in the $z$ -direction. For simplicity, we will consider the trial functions as

$\begin{equation} X_{m, N} (t) = \sum\limits_{i = 0}^N {X_{m, i} \mathscr{P}^{0 , 0 , \ell }_{i}(t)}, \quad m = 1, \dots , M. \end{equation}$

(3.4)

Also, we can use the following approximation:

$\begin{equation} \begin{split} & I_{t, N}^{0, 0, \ell } I_{r, N}^{\rho -1 , 0, \ell } \left( {\kappa(t, \ell)} \right)^\rho g_m\left( {s(t, r), Z _N (s(t, r)), Z _N (q\, s(t, r))} \right) \\ = &\sum\limits_{i = 0}^N {\sum\limits_{j = 0}^N {v_{m, i, j} \mathscr{P}^{0 , 0 , \ell }_{i}(t)\mathscr{P}^{\rho -1 , 0 , \ell }_{j}(r)} }, \quad m = 1, \dots , M . \end{split} \end{equation}$

(3.5)

A straightforward calculation by using (3.5) and (2.6) gives

$\begin{equation} \begin{split} &\frac{1}{{\Gamma (\rho )}}I_{t, N}^{0, 0, \ell }\left[ {\left( {\kappa(t, \ell)} \right)^\rho }\int_I {r^{-1}\left( {1 - \kappa(r, \ell)} \right)^{\rho - 1} I_{r, N}^{\rho - 1, 0, \ell } g_m(s(t, r), Z _N(s(t, r)), Z _N(qs(t, r)))dr}\right]\\ = & \frac{1}{{\Gamma (\rho )}}\sum\limits_{i = 0}^N {\sum\limits_{j = 0}^N {v_{m, i, j} \mathscr{P}^{0 , 0 , \ell }_{i}(t)\int_I r^{-1}\left( {1 - \kappa(r, \ell)} \right)^{\rho - 1}\mathscr{P}^{\rho -1 , 0 , \ell }_{j}(r)}dr }\\ = &\frac{1}{{\Gamma (\rho +1 )}}\sum\limits_{i = 0}^N v_{m, i, 0} \mathscr{P}^{0 , 0 , \ell }_{i}(t), \quad m = 1, \dots , M. \end{split} \end{equation}$

(3.6)

Let $\left\{\chi _p^{\rho, \eta, \ell }, x _p^{\rho, \eta, \ell }\right\}_{p = 0}^N$ be the weights and the nodes of Gauss-type logarithmic Jacobi interpolation. A direct application of (3.5) and (2.12) yields

$\begin{equation} \begin{split} v_{m, i, 0} = &\rho \, (2i+1) \sum\limits_{p = 0}^N \sum\limits_{q = 0}^N \left( \kappa(t_p^{0, 0, \ell }, \ell) \right)^\rho \mathscr{P}_i^{0, 0, \ell } \left( {t_p^{0, 0, \ell } } \right)\\ &\times g_m\left( {s\left( {t_p^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right), Z _N \left( {s\left( {t_p^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right)} \right), Z _N \left( {q s\left( {t_p^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right)} \right)} \right)\chi _p^{0, 0, \ell } \chi _q^{\rho - 1, 0, \ell } . \end{split} \end{equation}$

(3.7)

Hence, we deduce that

$\begin{equation} \sum\limits_{i = 0}^N {X_{m, i} \mathscr{P}^{0 , 0 , \ell }_{i}(t)} = X_\ell\mathscr{P}^{0 , 0 , \ell }_{0}(t)+\frac{1}{{\Gamma (\rho +1 )}}\sum\limits_{i = 0}^N v_{m, i, 0} \mathscr{P}^{0 , 0 , \ell }_{i}(t). \end{equation}$

(3.8)

We compared the coefficients of (3.8) to get

$\begin{equation} \begin{split} X_{m, 0}& = Z_\ell + \frac{{v_{m, 0, 0} }}{{\Gamma \left( {\rho + 1} \right)}}, \\ X_{m, i} & = \frac{{v_{m, i, 0} }}{{\Gamma \left( {\rho + 1} \right)}}, \quad 1 \le i \le N, \quad m = 1, \dots , M, \end{split} \end{equation}$

(3.9)

where $Z_\ell$ is the vector of initial values defined in (3.1).

4. Auxiliary lemmas

Here, we derive the rate of convergence of the scheme (3.3) in the $L_{\chi ^{0, 0, \ell }}^2$ -norm. Accordingly, we introduce some lemmas.

Let $r _i^{\rho, \eta, \ell }$ be the log J-Gauss nodes in $I,$ and $s _i^{\rho, \eta, \ell } = s \left({x, r _i^{\rho, \eta, \ell } } \right)$ . The mapped log J-Gauss interpolation operator ${}_x \widetilde{\mathcal{I}}_{s, N}^{\rho, \eta, \ell }: C(\ell, x) \longrightarrow P^{\log }_{N} (\ell, x)$ is defined by

$\begin{equation} {}_x \widetilde{\mathcal{I}}_{s , N}^{\rho , \eta, \ell } u\left( {s _i^{\rho , \eta, \ell } } \right) = u\left( {s _i^{\rho , \eta, \ell } } \right), \quad 0 \le i \le N. \end{equation}$

(4.1)

Hence,

$\begin{equation} {}_x\widetilde{\mathcal{I}}_{s , N}^{\rho , \eta, \ell } u\left( {s _i^{\rho , \eta, \ell } } \right) = u\left( {s _i^{\rho , \eta, \ell } } \right) = u\left( {s (x, r _i^{\rho , \eta, \ell } )} \right) = \mathcal{I}_{r , N}^{\rho , \eta, \ell } u\left( {s (x, r _i^{\rho , \eta, \ell } )} \right), \end{equation}$

(4.2)

and

$\begin{equation} {}_x\widetilde{\mathcal{I}}_{s , N}^{\rho , \eta, \ell } u\left( s \right) = \left. {\mathcal{I}_{r, N}^{\rho , \eta, \ell } u\left( {s (x, r)} \right)} \right|_{\kappa(r, \ell) = \frac{{\kappa(s, \ell)}}{{\kappa(x, \ell)}}}. \end{equation}$

(4.3)

Moreover, the following results can be easily derived:

$\begin{equation} \begin{split} \int_{\ell}^{x} s^{-1}\left( {\kappa(x, s)} \right)^{\rho - 1} {}_x \widetilde{\mathcal{I}}_{s, N}^{\rho -1, 0, \ell } X(s)ds & = \left( {\kappa(x, \ell)} \right)^\rho \int_{I} r^{-1}\left( {1 - \kappa(r, \ell)} \right)^{\rho - 1} \mathcal{I}_{r, N}^{\rho -1, 0, \ell } X \left(s\left(x, r\right)\right)dr\\ & = \left( {\kappa(x, \ell)} \right)^\rho \sum\limits_{j = 0}^{N} X \left(s(x, r_{j}^{\rho -1, 0, \ell })\right) \chi_{j}^{\rho -1, 0, \ell }\\ & = \left( {\kappa(x, \ell)} \right)^\rho \sum\limits_{j = 0}^{N} X \left(s_{j}^{\rho -1, 0, \ell }\right) \chi_{j}^{\rho -1, 0, \ell }. \end{split} \end{equation}$

(4.4)

Similarly,

$\begin{equation} \int_{\ell}^{x} s^{-1}\left( {\kappa(x, s)} \right)^{\rho - 1} \left({}_x \widetilde{\mathcal{I}}_{s, N}^{\rho -1, 0, \ell } X(s)\right)^2ds = \left( {\kappa(x, \ell)} \right)^\rho \sum\limits_{j = 0}^{N} X^{2} \left(s_{j}^{\rho -1, 0, \ell }\right) \chi_{j}^{\rho -1, 0, \ell }. \end{equation}$

(4.5)

Then, for any $1 \leq s \leq N + 1$ , we have

$\begin{equation} \begin{split} &\int_{\ell}^x {s^{-1}\left( {\kappa(x, s)} \right)^{\rho - 1}\left| {\left( \mathcal{I}- {}_x \widetilde{\mathcal{I}}_{s , N}^{\rho -1, 0, \ell } \right) X\left( s \right)} \right|^2 ds }\\ = & \left( {\kappa(x, \ell)} \right)^\rho \int_{I} {r^{-1}\left( {1 - \kappa(r, \ell)} \right)^{\rho - 1} \left| {\left( \mathcal{I} - \mathcal{I}_{r, N}^{\rho -1, 0, \ell }\right) X \left( {s (x, r)} \right)} \right|^2 dr}\\ \le& cN^{ - 2m} \left( {\kappa(x, \ell)} \right)^\rho \int_{ I} {r^{-1}\left (1-\kappa(r, \ell)\right)^{\rho +m-1 }\left(\kappa(r, \ell)\right)^{m } \left| {D_{\log, r}^m X\left( {s (x, r)} \right)} \right|^2 dr}\\ = & cN^{ - 2m} \int_{ \ell}^x {s^{-1}\left(\kappa(x, s)\right)^{\rho + m -1} \left(\kappa(s, \ell)\right)^m \left| {D_{\log, s} ^m X(s )} \right|^2 ds }, \end{split} \end{equation}$

(4.6)

where $\mathcal{I}$ is the identity operator.

Lemma 4.1. The following estimate holds for the error function $e_N(x) = Z (x)-Z _N(x)$ :

$\begin{equation} \left\| e_N \right\|_{\chi ^{0 , 0 , \ell }} \leq \sum\limits_{j = 1}^{3} \left\| \Xi_{j} \right\|_{\chi ^{0 , 0 , \ell }}, \end{equation}$

(4.7)

where

$\begin{equation*} \begin{split} \Xi_1 & = Z (x)- \mathcal{I}^{0, 0, \ell }_{x, N} Z (x), \\ \Xi_2 & = \mathcal{I}^{0, 0, \ell }_{x, N} \int_{\ell}^{x} \boldsymbol{R}(x, s) (\mathcal{I}-{}_x \widetilde{\mathcal{I}}_{s, N}^{\rho -1, 0, \ell }) \quad \textbf{Q}(s, Z (s), Z (q\, s)) ds, \\ \Xi_3 & = \mathcal{I}^{0, 0, \ell }_{x, N} \int_{\ell}^{x} \boldsymbol{R}(x, s) \quad {}_x \widetilde{\mathcal{I}}_{s, N}^{\rho -1, 0, \ell }(\textbf{Q}(s, Z (s), Z (qs))-\textbf{Q}(s, Z _N(s), Z _N(qs))) ds, \end{split} \end{equation*}$

and $\boldsymbol{R}(x, s) = (R_{ij})$ with $R_{ij} = \frac{s^{-1} \left({\kappa(x, s)} \right)^{\rho - 1}}{\Gamma(\rho)} \delta_{ij}, \ i, \ j = 1, \ldots, M.$

Proof.

$\begin{equation} \left\| e_N \right\|_{\chi ^{0 , 0 , \ell }} \leq \left\| Z -\mathcal{I}^{0, 0, \ell }_{x, N} Z \right\|_{\chi ^{0 , 0 , \ell }} + \left\| \mathcal{I}^{0, 0, \ell }_{x, N} Z- Z_N \right\|_{\chi ^{0 , 0 , \ell }}. \end{equation}$

(4.8)

It is clear from (3.1) that

$\begin{equation} \mathcal{I}^{0, 0, \ell }_{x, N} Z (x) = Z _\ell + \frac{1}{{\Gamma (\rho )}} \mathcal{I}^{0, 0, \ell }_{x, N} \int_\ell^x {s^{-1}\left( {\kappa(x, s)} \right)^{\rho - 1} \textbf{Q}\left( {s, Z (s), Z (qs)} \right)ds}, \end{equation}$

(4.9)

and

$\begin{equation} Z _N(x) = Z _\ell + \frac{1}{{\Gamma (\rho )}} \mathcal{I}^{0, 0, \ell }_{x, N} \int_\ell^x {s^{-1}\left( {\kappa(x, s)} \right)^{\rho - 1} {}_x \widetilde{\mathcal{I}}_{s, N}^{\rho -1, 0, \ell } \textbf{Q}\left( {s, Z _N(s), Z _N(qs)} \right)ds}. \end{equation}$

(4.10)

Subtracting (4.9) from (4.10) yields

(4.11)

which has the form:

$\begin{equation} \begin{split} & \mathcal{I}^{0, 0, \ell }_{x, N} Z (x)- Z _N(x)\\ = & \frac{1}{{\Gamma (\rho )}} \mathcal{I}^{0, 0, \ell }_{x, N} \int_\ell^x {s^{-1}\left( {\kappa(x, s)} \right)^{\rho - 1} \left(\mathcal{I}-{}_x \widetilde{\mathcal{I}}_{s, N}^{\rho -1, 0, \ell }\right) \textbf{Q}\left( {s, Z (s), Z (qs)} \right)ds}\\ &+ \frac{1}{{\Gamma (\rho )}} \mathcal{I}^{0, 0, \ell }_{x, N} \int_\ell^x {s^{-1}\left( {\kappa(x, s)} \right)^{\rho - 1} {}_x \widetilde{\mathcal{I}}_{s, N}^{\rho -1, 0, \ell }\left(\textbf{Q}\left( {s, Z (s), Z (qs)} \right)-\textbf{Q}\left( {s, Z _N(s), Z _N(qs)} \right)\right)ds}. \end{split} \end{equation}$

(4.12)

□

5. Convergence analysis

5.1. Convergence analysis in $L_{ \chi ^{0, 0, \ell }}^{2}$ -norm

Theorem 5.1. Let $Z(x)$ be the solutions of the systems (3.1) and (3.3), respectively. Then we have the following estimate:

$\begin{equation} \left\| Z-Z_N \right\|_{\chi ^{0 , 0 , \ell }} \leq c N^{-m} \left(\Vert D^{m}_{\log } Z \Vert ^{2}_{ \chi ^{m, m, \ell }}+\Vert D^{m}_{\log } \textbf{Q}\left(x, Z (x), Z (q\, x)\right)\Vert ^{2}_{ \chi ^{\rho + m - 1, m, \ell }}\right), \end{equation}$

(5.1)

where $1\leq m \leq N+1$ and $N\geq 1$ .

Proof. Using Proposition 2.1, we get

$\begin{equation} \left\|\Xi_1 \right\|_{\chi ^{0 , 0 , \ell }} = \left\| Z -\mathcal{I}^{0 , 0 , \ell }_{x, N} Z \right\|_{\chi ^{0 , 0 , \ell }}\leq c N^{-m}\Vert D^{m}_{\log } Z \Vert ^{2}_{ \chi ^{m, m, \ell }} \leq c N^{-m} \left\| \partial_{x}^{m} Z (\ell\, e^{x}) \right\|_{\chi^{m, m}}. \end{equation}$

(5.2)

Using the log J-Gauss integration formula, gives

$\begin{equation*} \begin{split} \left\|\Xi_2 \right\|_{\chi ^{0 , 0 , \ell }}& = \left\| {\mathcal{I}_{x, N}^{0, 0, \ell } \int_\ell^x \textbf{R}(x, s) \left( {\mathcal{I} - {}_x\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) \textbf{Q}\left( {s, Z (s), Z (qs)} \right)ds} \right\|_{\chi ^{0, 0, \ell } } \\ & = \left\| \sum\limits_{k = 1}^M {\mathcal{I}_{x, N}^{0, 0, \ell } \int_\ell^x R_{kk}(x, s) \left( {\mathcal{I} - {}_x\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) g_{k}\left( {s, Z (s), Z (q\, s)} \right)ds} \right\|_{\chi ^{0, 0, \ell } } \\ & = \left[ \int_I \chi ^{0, 0, \ell } \left( \sum\limits_{k = 1}^M {\mathcal{I}_{x, N}^{0, 0, \ell } \int_\ell^x R_{kk}(x, s) \left( {\mathcal{I} - {}_x\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) g_{k}\left( {s, Z (s), Z (q\, s)} \right)ds}\right) dx \right]^{1/2}\\ & = \left[ \sum\limits_{j = 0}^N \chi_j ^{0, 0, \ell } \left( \sum\limits_{k = 1}^M { \int_\ell^{x_j^{0, 0, \ell } } R_{kk}(x_j^{0, 0, \ell }, s) \left( {\mathcal{I} - {}_{x_j^{0, 0, \ell }}\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) g_{k}\left( {s, Z (s), Z (q\, s)} \right) ds}\right)^2 \right]^{1/2}\\ & \leq \left[ \sum\limits_{j = 0}^N \chi_j ^{0, 0, \ell } \sum\limits_{k = 1}^M \left( { \int_\ell^{x_j^{0, 0, \ell } } R_{kk}(x_j^{0, 0, \ell }, s) \left( {\mathcal{I} - {}_{x_j^{0, 0, \ell }}\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) g_{k}\left( {s, Z (s), Z (q\, s)} \right) ds}\right)^2 \sum\limits_{k = 1}^M (1)^2 \right]^{1/2}.\\ \end{split} \end{equation*}$

Using Cauchy-Schwarz inequality leads to the following estimate:

$\begin{equation} \begin{split} \left\|\Xi_2 \right\|_{\chi ^{0 , 0 , \ell }} &\leq C \left[ {\sum\limits_{j = 0}^N \sum\limits_{k = 1}^M {\chi _j^{0, 0, \ell } \int_\ell^{x_j^{0, 0, \ell } } {R_{kk} (x_j^{0, 0, \ell }, s) \ ds} \int_\ell^{x_j^{0, 0, \ell } } {R_{kk}(x_j^{0, 0, \ell }, s) \left| {\left( {\mathcal{I} - {}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) g_k\left( {s, Z (s), Z (q\, s)} \right)} \right|^2 ds} } } \right]^{1/2} \\ &\leq C \left[ {\sum\limits_{j = 0}^N \sum\limits_{k = 1}^M {\chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^\rho \int_\ell^{x_j^{0, 0, \ell } } {s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left| {\left( {\mathcal{I} - {}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) g_k\left( {s, Z (s), Z (q\, s)} \right)} \right|^2 ds} } } \right]^{1/2} \\ &\leq C \left( \sum\limits_{j = 0}^N \chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^\rho \right)^{1/2} \left(\sum\limits_{k = 1}^M\int_\ell^{x_j^{0, 0, \ell } } s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left| {\left( {\mathcal{I} - {}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) g_k\left( {s, Z (s), Z (q\, s)} \right)} \right|^2 ds \right)^{1/2} \\ &\leq C \left( \sum\limits_{j = 0}^N \chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^\rho \right)^{1/2} \left(\int_\ell^{x_j^{0, 0, \ell } } s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left| {\left( {\mathcal{I} - {}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } } \right) \textbf{Q}\left( {s, Z (s), Z (qs)} \right)} \right|^2 ds \right)^{1/2} \\ &\leq c N^{-m} \left[ {\sum\limits_{j = 0}^N {\chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^\rho \int_\ell^{x_j^{0, 0, \ell } } {s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho + m - 1} \left( {\kappa(s, \ell)} \right)^m \left| {D_{\log , s}^m \textbf{Q}\left( {s, Z (s), Z (qs)} \right)} \right|^2 ds} } } \right]^{1/2} \\ &\leq c N^{-m}\Vert D^{m}_{\log } \textbf{Q}\left(\cdot, Z (\cdot), Z (q\, \cdot)\right)\Vert ^{2}_{ \chi ^{\rho + m - 1, m, \ell }}. \end{split} \end{equation}$

(5.3)

An estimate for the term $\left\|E_3 \right\|_{\chi ^{0, 0, \ell }}$ can be obtained by using the log J-Gauss integration formula, to give

$\begin{equation*} \begin{split} \left\|\Xi_3\right\|_{\chi ^{0 , 0 , \ell }} ^2 = &\left\| \textbf{R}(x, s) \mathcal{I}^{0, 0, \ell }_{x, N} \int_{\ell}^{x} {}_x \widetilde{\mathcal{I}}_{s, N}^{\rho -1, 0, \ell }(\textbf{Q}(s, Z (s), Z (q\, s))-\textbf{Q}(s, Z _N(s), Z _N(q\, s))) ds\right\|_{\chi ^{0 , 0 , \ell }}^2\\ = &\frac{1}{{\Gamma^2 \left( \rho \right)}} \int_I \chi ^{0, 0, \ell } \times \left(\sum\limits_{k = 1}^M \mathcal{I}_{x, N}^{0, 0, \ell } \int_\ell^{x} s^{ - 1} \left( \kappa(x, s) \right)^{\rho - 1} {}_{x }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } \left(g_k \left( {s, Z (s), Z (q\, s)} \right)-g_k\left( {s, Z _N(s), Z _N(q\, s)} \right)\right)ds \right)^2 dx \\ = &\frac{1}{{\Gamma^2 \left( \rho \right)}} \sum\limits_{j = 0}^N \chi _j^{0, 0, \ell } \times\left( \int_\ell^{x_j^{0, 0, \ell }} s^{ - 1} \left( \kappa(x, s) \right)^{\rho - 1} \sum\limits_{k = 1}^M {}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } \left(g_k \left( {s, Z (s), Z (q\, s)} \right)-g_k\left( {s, Z _N(s), Z _N(q\, s)} \right)\right)ds \right)^2 .\\ \end{split} \end{equation*}$

Using the Cauchy-Schwarz inequality, we get

$\begin{equation} \begin{split} & \left\|\Xi_3 \right\|_{\chi ^{0 , 0 , \ell }}^2\\\leq& \frac{1}{{\Gamma^2 \left( \rho \right)}} \sum\limits_{j = 0}^N \chi _j^{0, 0, \ell } \int_\ell^{x_j^{0, 0, \ell }} s^{ - 1} \left( \kappa(x, s) \right)^{\rho - 1} ds \\& \times \int_\ell^{x_j^{0, 0, \ell }} s^{ - 1} \left( \kappa(x, s) \right)^{\rho - 1} \left( \left| \sum\limits_{k = 1}^M {}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } \left(g_k \left( {s, Z (s), Z (q\, s)} \right)-g_k\left( {s, Z _N(s), Z _N(q\, s)} \right)\right) \right| \right)^2 ds \\ \leq &\frac{1}{{\Gamma^2 \left( \rho \right)}} \sum\limits_{j = 0}^N \chi _j^{0, 0, \ell } \left( {\log \frac{{x }}{\ell}} \right)^{\rho } \int_\ell^{x_j^{0, 0, \ell }} s^{ - 1} \left( \kappa(x, s) \right)^{\rho - 1} \\ &\times\left( \sum\limits_{k = 1}^M \left| {}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } \left(g_k \left( {s, Z (s), Z (q\, s)} \right)-g_k\left( {s, Z _N(s), Z _N(q\, s)} \right)\right) \right| \right)^2 ds , \\ \end{split} \end{equation}$

(5.4)

and using the logarithmic Jacobi-Gauss quadrature formula (4.4), we obtain

$\begin{equation} \begin{split} \left\|\Xi_3 \right\|_{\chi ^{0 , 0 , \ell }} \leq & \frac{1}{{\Gamma \left( {\rho + 1} \right)}}\left[ \sum\limits_{j = 0}^N \rho \chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^{2\rho }\right.\\ & \times \left. \sum\limits_{q = 0}^N \chi _q^{\rho - 1, 0, \ell } \left( \sum\limits_{k = 1}^M \left| {g_k \left( {s\left( {x_j^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right), Z \left( {s\left( {x_j^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right)} \right), Z \left( {q\, s\left( {x_j^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right)} \right)} \right)}\right. \right. \right. \\ &-\left. \left. \left.{g_k\left( {s\left( {x_j^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right), Z _N \left( {s\left( {x_j^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right)} \right), Z _N \left( {q\, s\left( {x_j^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right)} \right)} \right)} \right| \right)^2 \right]^{1/2}. \end{split} \end{equation}$

(5.5)

Using the Lipschitz condition, we obtain

$\begin{equation} \begin{split} \left\| \Xi_3 \right\|_{\chi ^{0 , 0 , \ell }}\leq \frac{L}{{\Gamma \left( {\rho + 1} \right)}} \times\left[ {\sum\limits_{j = 0}^N {\rho \chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^{2\rho } \sum\limits_{q = 0}^N \left( \sum\limits_{i = 1}^M \chi _q^{\rho - 1, 0, \ell } \left| {X_i\left( {s\left( {x_j^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right)} \right) - X_{N, i} \left( {s\left( {x_j^{0, 0, \ell } , r_q^{\rho - 1, 0, \ell } } \right)} \right)} \right| \right) ^2 } } \right]^{1/2} , \end{split} \end{equation}$

(5.6)

using (4.5), we get

$\begin{equation} \begin{split} \left\| \Xi_3 \right\|_{\chi ^{0 , 0 , \ell }}\leq & \frac{L}{{\Gamma \left( {\rho + 1} \right)}} \times \left[ {\sum\limits_{j = 0}^N \rho \chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^\rho \int_\ell^{x_j^{0, 0, \ell } } {s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left( \sum\limits_{i = 1}^M\left| {{}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } \left( {X_i(s) - X_{N, i} (s)} \right)} \right| \right)^2 ds} } \right]^{1/2}. \\ \left\|E_3 \right\|_{\chi ^{0 , 0 , \ell }}\leq& \frac{L}{{\Gamma \left( {\rho + 1} \right)}}\left( {\sum\limits_{j = 0}^N {\rho \chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^\rho } } \right)^{1/2} \\ &\times \mathop {\max }\limits_{0 \le j \le N} \left( {\int_\ell^{x_j^{0, 0, \ell } } {s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left( \sum\limits_{i = 1}^M \left| {{}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } \left( {X_i(s) - X_{i, N} (s)} \right)} \right| \right)^2 ds} } \right)^{1/2}. \end{split} \end{equation}$

(5.7)

For any $x_j^{0, 0, \ell }\in I$ . Let $f(\rho) = \left(\kappa(x_j^{0, 0, \ell }, \ell) \right)^\rho$ . We note that $f(\rho)$ is a convex function of $\rho$ . Hence, by Jensen's inequality for all $\rho \in (0, 1)$ ,

$\begin{equation*} f(\rho ) = (1 - \rho )f(0) + \rho f(1). \end{equation*}$

The above inequality yields

$\begin{equation} \begin{split} \rho \sum\limits_{j = 0}^N {\chi _j^{0, 0, \ell } \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)^\rho } &\le \rho \sum\limits_{j = 0}^N {\chi _j^{0, 0, \ell } \left[ {1 - \rho + \rho \left( \kappa(x_j^{0, 0, \ell }, \ell) \right)} \right]}\\ & \le \rho \left[ {1 - \rho + \rho \int_I {s^{ - 1} \left( {\log \frac{x}{a}} \right)dx} } \right]\leq \rho \left( {1 - \frac{\rho }{2}} \right) \leq \frac{1}{2}. \end{split} \end{equation}$

(5.8)

Hence, by using the above inequality, the triangle inequality, (4.6) and (4.5), we deduce that

$\begin{equation} \begin{split} \left\|\Xi_3 \right\|_{\chi ^{0 , 0 , \ell }}&\leq \frac{L}{{\sqrt{2}\, \Gamma \left( {\rho + 1} \right)}} \mathop {\max }\limits_{0 \le j \le N} \left( {\int_\ell^{x_j^{0, 0, \ell } } {s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left( \sum\limits_{i = 1}^M \left| {}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } \left( {X_i(s) - X_{N, i} (s)} \right) \right| \right)^2 ds} } \right)^{1/2}\\ &\leq \frac{L}{{\sqrt 2 \ \Gamma \left( {\rho + 1} \right)}} \times \mathop {\max }\limits_{0 \le j \le N} \left[ {\left( {\int_\ell^{x_j^{0, 0, \ell } } {s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left( \sum\limits_{i = 1}^M\left| {{}_{x_j^{0, 0, \ell } }\widetilde{\mathcal{I}}_{s, N}^{\rho - 1, 0, \ell } X_i(s)-X_i(s)} \right| \right)^2 ds} } \right)^{1/2}}\right. \\ &\quad + \left.{\left( \int_\ell^{x_j^{0, 0, \ell } } {s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left( \sum\limits_{i = 1}^M\left| {X_i(s) - X_{N, i} (s)} \right|\right)^2 ds} \right)^{1/2} } \right]\\ &\leq cN^{ - m} \mathop {\max }\limits_{0 \le j \le N} \left(\int_{ \ell}^{x_j} \left(\kappa(s, \ell)\right)^m \left( \sum\limits_{i = 1}^M \left| {D_{\log, s} ^m X_i(s )} \right| \right)^2 ds \right)^{1/2} \\ & \quad +\frac{L}{{\sqrt 2 \ \Gamma \left( {\rho + 1} \right)}} \times \mathop {\max }\limits_{0 \le j \le N} {\left( \int_\ell^{x_j^{0, 0, \ell } } {s^{ - 1} \left( \kappa(x_j^{0, 0, \ell }, s) \right)^{\rho - 1} \left( \sum\limits_{i = 1}^M\left| {X_i(s) - X_{N, i} (s)} \right|\right)^2 ds} \right)^{1/2} } \\ & \leq c N^{-m}\Vert D^{m}_{\log }Z \Vert ^{2}_{ \chi ^{m, m, \ell }}+ \frac{L}{{\sqrt 2 \Gamma \left( {\rho + 1} \right)}} \Vert e_N \Vert ^{2}_{ \chi ^{m, m, \ell }}. \end{split} \end{equation}$

(5.9)

Hence, a combination of (5.2), (5.3), (5.9) and the Lipschitz constant $L < \Gamma(\rho +1)$ leads to the desired result. □

6. Numerical results

In order to illustrate the significance of our key findings, we provide two numerical examples in this section.

Example 6.1. We consider the following initial value problem:

$\begin{equation} {}_1^{CH} D_t^\rho X(t) = g(x), \quad X(1) = 0, \ t \in (1, e), \ \rho \in (0, 1]. \end{equation}$

(6.1)

Table 1 shows a comparison of the maximum absolute errors that are obtained from the method that we have presented and those given in ^[22] and ^[21]. The numerical results depict that, by using the method proposed in this paper, higher accuracy is achieved.

Table 1. A comparison between the maximum absolute errors of presented method and methods in ^[22] and ^[21] with

$\rho = 0.5$ for Example 6.1.

$N$	Error	Error ^[22]	$N$ ^[21]	Error ^[21]
$9$	$5.3874 \times 10^{-9}$	$9.1283 \times 10^{-6}$	$20$	$1.2500\times 10^{-3}$
$10$	$2.7917 \times 10^{-9}$	$2.2831 \times 10^{-6}$	$40$	$2.8647\times 10^{-4}$
$11$	$1.5313 \times 10^{-9}$	$7.1562 \times 10^{-7}$	$80$	$6.6144\times 10^{-5}$
$12$	$8.8111 \times 10^{-10}$	$2.6679 \times 10^{-7}$	$160$	$1.5345\times 10^{-5}$

| Show Table

DownLoad: CSV

To investigate numerically the stability of the spectral collocation scheme, we consider the initial value problem (6.1) and the following problems, whose right-hand side, the initial value, and the order of the differential operator suffer perturbations.

$\begin{equation} {}_1^{CH} D_t^\rho Y(t) = g(x)+\varepsilon_g, \; Y(1) = 0, \; t \in (1, e), \; \rho = 0.5. \end{equation}$

(6.2)

$\begin{equation} {}_1^{CH} D_t^{\rho+\varepsilon_{\rho}} Y(t) = g(x), \; Y(1) = 0, \; t \in (1, e), \; \rho = 0.5, \; \varepsilon_{\rho} \in (-0.5, 0.5). \end{equation}$

(6.3)

$\begin{equation} {}_1^{CH} D_t^\rho Y(t) = g(x), \; Y(1) = \varepsilon_{Y_0}, \; t \in (1, e), \; \rho = 0.5. \end{equation}$

(6.4)

The maximum absolute errors $\left| {X_N - Y_N } \right|$ , where $X_N$ is the numerical solution of problem (6.1) and $Y_N$ is the numerical solution of the perturbed problems (6.2), (6.3), and (6.4), are displayed in , , and , respectively. We observe that $\| {X_N - Y_N } \|_{\infty} = {\rm O}(\varepsilon_{g})$ , $\| {X_N - Y_N } \|_{\infty} = {\rm O}(\varepsilon_{\rho})$ , and $\| {X_N - Y_N } \|_{\infty} = {\rm O}(\varepsilon_{y_{0}})$ , respectively, independently of $N$ .

Table 2. Maximum of the absolute errors,

$\left| {X_N - Y_N } \right|$ , where

$X_N$ is the numerical solution of problem (6.1) and

$Y_N$ is the numerical solution of the perturbed problem (6.2) with several values of

$\varepsilon_g$ .

$N$	$\varepsilon_g=0.1$	$\varepsilon_g=0.01$	$\varepsilon_g=0.001$
$5$	$1.130 \times 10^{-1}$	$1.131 \times 10^{-2}$	$1.131\times 10^{-3}$
$10$	$1.127 \times 10^{-1}$	$1.127 \times 10^{-2}$	$1.127\times 10^{-3}$
$15$	$1.128 \times 10^{-1}$	$1.128 \times 10^{-2}$	$1.128\times 10^{-3}$

| Show Table

DownLoad: CSV

Table 3. Maximum of the absolute errors,

$\left| {X_N - Y_N } \right|$ , where

$X_N$ is the numerical solution of problem (6.1) and

$Y_N$ is the numerical solution of the perturbed problem (6.3) with several values of

$\varepsilon_{\rho}$ .

$N$	$\varepsilon_{\rho}=0.1$	$\varepsilon_{\rho}=0.01$	$\varepsilon_{\rho}=0.001$
$5$	$1.192 \times 10^{-1}$	$1.249 \times 10^{-2}$	$1.255\times 10^{-3}$
$10$	$1.192 \times 10^{-1}$	$1.249 \times 10^{-2}$	$1.255\times 10^{-3}$
$15$	$1.192 \times 10^{-1}$	$1.249 \times 10^{-2}$	$1.255\times 10^{-3}$

| Show Table

DownLoad: CSV

Table 4. Maximum of the absolute errors,

$\left| {X_N - Y_N } \right|$ , where

$X_N$ is the numerical solution of problem (6.1) and

$Y_N$ is the numerical solution of the perturbed problem (6.4) with several values of

$\varepsilon_{Y_0}$ .

$N$	$\varepsilon_{Y_0}=0.1$	$\varepsilon_{Y_0}=0.01$	$\varepsilon_{Y_0}=0.001$
$5$	$1.000 \times 10^{-1}$	$1.000 \times 10^{-2}$	$1.000\times 10^{-3}$
$10$	$1.000 \times 10^{-1}$	$1.000 \times 10^{-2}$	$1.000\times 10^{-3}$
$15$	$1.000 \times 10^{-1}$	$1.000 \times 10^{-2}$	$1.000\times 10^{-3}$

| Show Table

DownLoad: CSV

Example 6.2. We consider the following coupled system:

$\begin{equation} \begin{split} &{}_1^{CH} D_t^\rho X_1(t) = X_2^2(q t) + g_1(t) , \quad \rho \in (0, 1), \\ &{}_1^{CH} D_t^\rho X_2(t) = X_1^2(q t) + g_2(t) , \quad \rho \in (0, 1). \end{split} \end{equation}$

(6.5)

For this problem, the exact solution is given as

$X_1(t) = \left( {\log t} \right)^5+ 2\left( {\log t} \right)^3,$

$X_2(t) = - \left( {\log t} \right)^4 +2\left( {\log t} \right)^3.$

We employ the proposed method to solve this problem with various $N$ and $\rho$ values. In , we list the errors for different values of $N$ and $\rho$ . The numerical results show the convergence of the scheme, which confirms our error analysis.

Table 5. The errors with the fractional orders

$\rho = 0.2, \, 0.4, \, 0.6, \, 0.8,$ and

$q = 3/4$ for Example 6.2.

The errors for $X_1$
$N$	$\rho=0.2$	$\rho=0.4$	$\rho=0.6$	$\rho=0.8$
$5$	$6.8 \times 10^{-8}$	$4.6 \times 10^{-7}$	$1.3\times 10^{-6}$	$2.1\times 10^{-6}$
$10$	$6.7 \times 10^{-10}$	$6.0 \times 10^{-9}$	$2.4\times 10^{-8}$	$5.0\times 10^{-8}$
$15$	$3.9 \times 10^{-11}$	$4.2 \times 10^{-10}$	$1.9\times 10^{-9}$	$4.9\times 10^{-9}$
$20$	$5.0 \times 10^{-12}$	$6.0 \times 10^{-11}$	$3.1\times 10^{-10}$	$8.8\times 10^{-10}$
$25$	$9.9 \times 10^{-13}$	$1.3 \times 10^{-11}$	$7.5\times 10^{-11}$	$2.3\times 10^{-10}$
$30$	$2.6 \times 10^{-13}$	$7.2 \times 10^{-13}$	$2.3\times 10^{-11}$	$7.6\times 10^{-11}$
The errors for $X_2$
$N$	$\rho=0.2$	$\rho=0.4$	$\rho=0.6$	$\rho=0.8$
$5$	$5.7 \times 10^{-8}$	$4.3 \times 10^{-7}$	$1.3\times 10^{-6}$	$2.1\times 10^{-6}$
$10$	$5.3 \times 10^{-10}$	$5.3 \times 10^{-9}$	$2.2\times 10^{-8}$	$4.9\times 10^{-8}$
$15$	$3.0 \times 10^{-11}$	$3.6 \times 10^{-10}$	$1.8\times 10^{-9}$	$4.7\times 10^{-9}$
$20$	$3.8 \times 10^{-12}$	$5.2 \times 10^{-11}$	$2.9\times 10^{-10}$	$8.4\times 10^{-10}$
$25$	$7.6 \times 10^{-13}$	$1.1 \times 10^{-11}$	$7.0\times 10^{-11}$	$2.2\times 10^{-10}$
$30$	$2.0 \times 10^{-13}$	$3.21 \times 10^{-12}$	$2.1\times 10^{-11}$	$7.2\times 10^{-11}$

| Show Table

DownLoad: CSV

7. Conclusions

We provided a collocation spectral scheme for nonlinear systems of fractional pantograph delay differential equations. We constructed a mapped Jacobi spectral collocation scheme, described its effective implementation, and derived its convergence analysis. In addition, we provided a numerical example to support our theoretical analysis. The numerical results demonstrate the accuracy and effectiveness of the proposed scheme. We also conclude that the described technique produces very accurate results, even when employing a small number of base functions. Preserving some important mathematical properties and physical structures, such as existence, positivity preservation, the maximum principle, long-time behavior, and singular solutions, may be considered in future work ^[25,26].

Use of AI tools declaration

The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

Acknowledgments

This work was supported and funded by the Deanship of Scientific Research at Imam Mohammad Ibn Saud Islamic University (IMSIU) (grant number IMSIU-RP23095).

Conflict of interest

The authors assert that they do not have any known competing financial interests or personal relationships that could have influenced the work reported in this paper.

References

[1]	Y. Liu, Z. Li, H. Liu, Z. Kan, Skill transfer learning for autonomous robots and human-robot cooperation: A survey, Rob. Auton. Syst., 128 (2020), 103515. https://doi.org/10.1016/j.robot.2020.103515 doi: 10.1016/j.robot.2020.103515
[2]	J. Luo, W. Liu, W. Qi, J. Hu, J. Chen, C. Yang, A vision-based virtual fixture with robot learning for teleoperation, Rob. Auton. Syst., 164 (2023), 104414. https://doi.org/10.1016/j.robot.2023.104414 doi: 10.1016/j.robot.2023.104414
[3]	Y Liu, Z. Li, H. Liu, Z. Kan, B. Xu, Bioinspired embodiment for intelligent sensing and dexterity in fine manipulation: A survey, IEEE Trans. Ind. Inf., 16 (2020), 4308–4321. https://doi.org/10.1109/TⅡ.2020.2971643 doi: 10.1109/TⅡ.2020.2971643
[4]	A. Bicchi, V. Kumar, Robotic grasping and contact: A review, in IEEE International Conference on Robotics and Automation, 1 (2020), 348–353. https://doi.org/10.1109/ROBOT.2000.844081
[5]	A. T. Miller, S. Knoop, H. I. Christensen, P. K. Allen, Automatic grasp planning using shape primitives, in 2003 IEEE International Conference on Robotics and Automation, 2 (2003), 1824–1829. https://doi.org/10.1109/ROBOT.2003.1241860
[6]	G. P. Slota, M. S. Suh, M. L. Latash, V. M. Zatsiorsky, Stability control of grasping objects with different locations of center of mass and rotational inertia, J. Mot. Behav., 44 (2012), 169–178. https://doi.org/10.1080/00222895.2012.665101 doi: 10.1080/00222895.2012.665101
[7]	J. Bohg, A. Morales, T. Asfour, D. Kragic, Data-driven grasp synthesis-A survey, IEEE Trans. Rob., 30 (2014), 289–309. https://doi.org/10.1109/TRO.2013.2289018 doi: 10.1109/TRO.2013.2289018
[8]	J. Redmon, A. Angelova, Real-time grasp detection using convolutional neural networks, in 2015 IEEE International Conference on Robotics and Automation (ICRA), (2015), 1316–1322. https://doi.org/10.1109/ICRA.2015.7139361
[9]	R. Xu, F. Chu, P. A. Vela, GKNet: Grasp keypoint network for grasp candidates detection, Int. J. Rob. Res., 41 (2022), 361–389. https://doi.org/10.1177/02783649211069569 doi: 10.1177/02783649211069569
[10]	H. Cheng, Y. Wang, M. Q. Meng, A robot grasping system with single-stage anchor-free deep grasp detector, IEEE Trans. Instrum. Meas., 71 (2022), 1–12. https://doi.org/10.1109/TIM.2022.3165825 doi: 10.1109/TIM.2022.3165825
[11]	Y. Wu, F. Zhang, Y. Fu, Real-time robotic multigrasp detection using anchor-free fully convolutional grasp detector, IEEE Trans. Ind. Electron., 69 (2022), 13171–13181. https://doi.org/10.1109/TIE.2021.3135629 doi: 10.1109/TIE.2021.3135629
[12]	G. Zuo, J. Tong, H. Liu, W. Chen, J. Li, Graph-based visual manipulation relationship reasoning network for robotic grasping, Front. Neurorobot., 15 (2021), 719731. https://doi.org/10.3389/fnbot.2021.719731 doi: 10.3389/fnbot.2021.719731
[13]	J. Ge, L. Mao, J. Shi, Y. Jiang, Fusion-Mask-RCNN: Visual robotic grasping in cluttered scenes, Multimedia Tools Appl., (2023), 1–21. https://doi.org/10.1007/s11042-023-16365-y doi: 10.1007/s11042-023-16365-y
[14]	Y. Li, F. Guo, M. Zhang, S. Suo, Q. An, J. Li, et al., A novel deep learning-based pose estimation method for robotic grasping of axisymmetric bodies in industrial stacked scenarios, Machines, 10 (2022), 1141. https://doi.org/10.3390/machines10121141 doi: 10.3390/machines10121141
[15]	L. François, S. Bruno, C. Philippe, C. Gosselin, A model-based scooping grasp for the autonomous picking of unknown objects with a two-fingered gripper, Rob. Auton. Syst., 106 (2018), 14–25. https://doi.org/10.1016/j.robot.2018.04.003 doi: 10.1016/j.robot.2018.04.003
[16]	N. S. Pollard, Closure and quality equivalence for efficient synthesis of grasps from examples, Int. J. Rob. Res., 23 (2004), 595–613. https://doi.org/10.1177/0278364904044402 doi: 10.1177/0278364904044402
[17]	M. Abdeetedal, M. R. Kermani, Grasp synthesis for purposeful fracturing of object, Rob. Auton. Syst., 105 (2018), 47–58. https://doi.org/10.1016/j.robot.2018.03.003 doi: 10.1016/j.robot.2018.03.003
[18]	A. Saxena, J. Driemeyer, A. Y. Ng, Robotic grasping of novel objects using vision, Int. J. Rob. Res., 27 (2008), 157–173. https://doi.org/10.1177/0278364907087172 doi: 10.1177/0278364907087172
[19]	Y. Jiang, S. Moseson, A. Saxena. Efficient grasping from RGBD images: Learning using a new rectangle representation, in 2011 IEEE International Conference on Robotics and Automation, (2011), 3304–3311. https://doi.org/10.1109/ICRA.2011.5980145
[20]	I. Lenz, H. Lee, A. Saxena, Deep learning for detecting robotic grasps, preprint, arXiv: 1301.3592.
[21]	Y. Song, L. Gao, X. Li, W. Shen, A novel robotic grasp detection method based on region proposal networks, Rob. Comput.-Integr. Manuf., 65 (2020), 101963. https://doi.org/10.1016/j.rcim.2020.101963 doi: 10.1016/j.rcim.2020.101963
[22]	D. Morrison, P. Corke, J. Leitner, Learning robust, real-time, reactive robotic grasping, Int. J. Rob. Res., 39 (2020), 183–201. https://doi.org/10.1177/0278364919859066 doi: 10.1177/0278364919859066
[23]	J. Mahler, J. Liang, S. Niyaz, M. Laskey, R. Doan, X. Liu, et al., Dex-net 2.0: Deep learning to plan robust grasps with synthetic point clouds and analytic grasp metrics, preprint, arXiv: 1703.09312.
[24]	H. Zhu, Y. Li, F. Bai, W. Chen, X. Li, J. Ma, et al., Grasping detection network with uncertainty estimation for confidence-driven semi-supervised domain adaptation, in 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), (2020), 9608–9613. https://doi.org/10.1109/IROS45743.2020.9341056
[25]	S. Yu, D. Zhai, Y. Xia, H. Wu, J. Liao, SE-ResUNet: A novel robotic grasp detection method, IEEE Rob. Autom.. Lett., 7 (2022), 5238–5245. https://doi.org/10.1109/LRA.2022.3145064 doi: 10.1109/LRA.2022.3145064
[26]	Q. Zhang, X. Sun, Bilateral cross-modal fusion network for robot grasp detection, Sensors, 23 (2023), 3340. https://doi.org/10.3390/s23063340 doi: 10.3390/s23063340
[27]	D. Guo, F. Sun, H. Liu, T. Kong, B. Fang, N. Xi, A hybrid deep architecture for robotic grasp detection, in 2017 IEEE International Conference on Robotics and Automation (ICRA), (2017), 1609–1614. https://doi.org/10.1109/ICRA.2017.7989191
[28]	Y. Huang, D. Liu, Z. Liu, K. Wang, Q. Wang, J. Tan, A novel robotic grasping method for moving objects based on multi-agent deep reinforcement learning, Rob. Comput.-Integr. Manuf., 86 (2024), 102644. https://doi.org/10.1016/j.rcim.2023.102644 doi: 10.1016/j.rcim.2023.102644
[29]	S. Kumra, C. Kanan, Robotic grasp detection using deep convolutional neural networks, in 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), (2017), 769–776. https://doi.org/10.1109/IROS.2017.8202237
[30]	F Chu, R. Xu, P. A. Vela, Real-world multiobject, multigrasp detection, IEEE Rob. Autom. Lett., 3 (2018), 3355–3362. https://doi.org/10.1109/LRA.2018.2852777 doi: 10.1109/LRA.2018.2852777
[31]	J. Ge, J. Shi, Z. Zhou, Z. Wang, Q. Qian, A grasping posture estimation method based on 3D detection network, Comput. Electr. Eng., 100 (2022), 107896. https://doi.org/10.1016/j.compeleceng. 2022.107896 doi: 10.1016/j.compeleceng.2022.107896
[32]	H. Zhang, X. Lan, S. Bai, L. Wan, C. Yang, N. Zheng, A multi-task convolutional neural network for autonomous robotic grasping in object stacking scenes, in 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), (2019), 6435–6442. https://doi.org/10.1109/IROS40897.2019.8967977
[33]	Y. Lin, L. Zeng, Z. Dong, X. Fu, A vision-guided robotic grasping method for stacking scenes based on deep learning, in 2019 IEEE 3rd Advanced Information Management, Communicates, Electronic and Automation Control Conference (IMCEC), (2019), 91–96. https://doi.org/10.1109/IMCEC46724.2019.8983819
[34]	C. Lu, R. Krishna, M. Bernstein, L. Fei-Fei, Visual relationship detection with language priors, in European Conference on Computer Vision, (2016), 852–869. https://doi.org/10.1007/978-3-319-46448-0_51
[35]	A. Zeng, S. Song, K. Yu, E. Donlon, F. R. Hogan, M. Bauza, et al., Robotic pick-and-place of novel objects in clutter with multi-affordance grasping and cross-domain image matching, Int. J. Rob. Res., 41 (2022), 690–705. http://doi.org/10.1177/0278364919868017 doi: 10.1177/0278364919868017
[36]	G. Wu, W. Chen, H. Cheng, W. Zuo, D. Zhang, J. You, Multi-object grasping detection with hierarchical feature fusion, IEEE Access, 7 (2019), 43884–43894. https://doi.org/10.1109/ACCESS.2019.2908281 doi: 10.1109/ACCESS.2019.2908281
[37]	W. Hu, C. Wang, F. Liu, X. Peng, P. Sun, J. Tan, A grasps-generation-and-selection convolutional neural network for a digital twin of intelligent robotic grasping, Rob. Comput.-Integr. Manuf., 77 (2022), 102371. https://doi.org/10.1016/j.rcim.2022.102371 doi: 10.1016/j.rcim.2022.102371
[38]	S. Duan, G. Tian, Z. Wang, S. Liu, C. Feng, A semantic robotic grasping framework based on multi-task learning in stacking scenes, Eng. Appl. Artif. Intell., 121 (2023), 106059. https://doi.org/10.1016/j.engappai.2023.106059 doi: 10.1016/j.engappai.2023.106059
[39]	S. Yu, D. Zhai, Y. Xia, EGNet: Efficient robotic grasp detection network, IEEE Trans. Ind. Electron., 70 (2023), 4058–4067. https://doi.org/10.1109/TIE.2022.3174274 doi: 10.1109/TIE.2022.3174274
[40]	X. Li, X. Zhang, X. Zhou, I. Chen, UPG: 3D vision-based prediction framework for robotic grasping in multi-object scenes, Knowl.-Based Syst., 270 (2023), 110491. https://doi.org/10.1016/j.knosys.2023.110491 doi: 10.1016/j.knosys.2023.110491
[41]	J. P. C. de Souza, L. F. Rocha, P. M. Oliveira, A. P. Moreira, J. Boaventura-Cunha, Robotic grasping: from wrench space heuristics to deep learning policies, Rob. Comput.-Integr. Manuf., 71 (2021), 102176. https://doi.org/10.1016/j.rcim.2021.102176 doi: 10.1016/j.rcim.2021.102176
[42]	J. Redmon, A. Farhadi, YOLOv3: An incremental improvement. Preprint, arXiv: 1804.02767.
[43]	S. Ren, K. He, R. Girshick, J. Sun, Faster R-CNN: Towards real-time object detection with region proposal networks, IEEE Trans. Pattern Anal. Mach. Intell., 39 (2017), 1137–1149. https://doi.org/10.1109/TPAMI.2016.2577031 doi: 10.1109/TPAMI.2016.2577031

This article has been cited by:

1.	Riza Andriani, Toto Subroto, Safri Ishmayana, Dikdik Kurnia, Enhancement Methods of Antioxidant Capacity in Rice Bran: A Review, 2022, 11, 2304-8158, 2994, 10.3390/foods11192994
2.	Sucheta Khubber, Francisco J Marti-Quijal, Igor Tomasevic, Fabienne Remize, Francisco J Barba, Lactic acid fermentation as a useful strategy to recover antimicrobial and antioxidant compounds from food and by-products, 2022, 43, 22147993, 189, 10.1016/j.cofs.2021.11.013
3.	Mohamed Ghamry, Wei Zhao, Li Li, Impact of Lactobacillus apis on the antioxidant activity, phytic acid degradation, nutraceutical value and flavor properties of fermented wheat bran, compared to Saccharomyces cerevisiae and Lactobacillus plantarum, 2023, 163, 09639969, 112142, 10.1016/j.foodres.2022.112142
4.	Hla Myo, Nara Nantarat, Nuntawat Khat-Udomkiri, Changes in Bioactive Compounds of Coffee Pulp through Fermentation-Based Biotransformation Using Lactobacillus plantarum TISTR 543 and Its Antioxidant Activities, 2021, 7, 2311-5637, 292, 10.3390/fermentation7040292
5.	Bhagavathi Sundaram Sivamaruthi, Karthikeyan Alagarsamy, Subramanian Thangaleela, Muruganantham Bharathi, Periyanaina Kesika, Chaiyavat Chaiyasut, Composition, Microbiota, Mechanisms, and Anti-Obesity Properties of Rice Bran, 2023, 12, 2304-8158, 1300, 10.3390/foods12061300
6.	Thornthan SAWANGWAN, Daleena KAJADMAN, Ratchanon KULCHANANIMIT, Determination of prebiotic properties of rice bran extract, 2024, 43, 2186-3342, 222, 10.12938/bmfh.2023-090
7.	Xiao Lian, Mingyu Shi, Qinlu Lin, Ying Liang, Lingyu Zhang, Research progress of probiotics and fermented feed effects on pork quality, 2024, 3, 2770-2081, 83, 10.1002/fbe2.12082
8.	Arnau Vilas-Franquesa, Melania Casertano, Anna Tresserra-Rimbau, Anna Vallverdú-Queralt, Cristian Torres-León, Recent advances in bio-based extraction processes for the recovery of bound phenolics from agro-industrial by-products and their biological activity, 2024, 64, 1040-8398, 10643, 10.1080/10408398.2023.2227261
9.	Mehdi Akbari, Seyed Hadi Razavi, Faramarz Khodaiyan, Jesús Blesa, María J. Esteve, Fermented corn bran: A by-product with improved total phenolic content and antioxidant activity, 2023, 184, 00236438, 115090, 10.1016/j.lwt.2023.115090
10.	Merve Tomas, Yuxi Wen, Wei Liao, Lizhu Zhang, Chao Zhao, David Julian McClements, Elifsu Nemli, Mustafa Bener, Resat Apak, Esra Capanoglu, Recent progress in promoting the bioavailability of polyphenols in plant-based foods, 2024, 1040-8398, 1, 10.1080/10408398.2024.2336051
11.	Qinggele Borjihan, Zhan Yang, Kailonng Liu, Chengcong Yang, Xiaoyuan Zhang, Guoqiang Yao, Comparative genomics and metabolite analysis of Lactiplantibacillus plantarum strains in dairy and vegetable fermentation environments, 2025, 65, 22124292, 106058, 10.1016/j.fbio.2025.106058

Reader Comments

Your name:*

Email:*
© 2024 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)