Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179599
P. Kumar, Prerna Saxena
We propose a novel compact antipodal elliptically tapered slot antenna loaded with left-handed metamaterial over 20-40 GHz for millimetre wave applications. The conventional antipodal tapered slot antenna is modified by incorporating a novel transition between the feed and taper section of the antenna to stabilize the radiation pattern in E-plane and to reduce the beam tilting. We employ a semi-dodecagon shaped dielectric lens with unit cells of a novel broadband metamaterial placed on it to improve the gain of the proposed antenna. The proposed antenna is designed on Rogers RT/Duroid 5880 substrate with dielectric constant of 2.2. The antenna is compact with a volume of $ 30times 49times$ 0.254 $mathrm{m}mathrm{m}^{3}$. The antenna reflection coefficient $(|mathrm{S}_{11}|lt -10mathrm{d}mathrm{B})$ shows good impedance matching over a wide frequency range of 20-40GHz. We exploit the left-handedness of the proposed broadband metamaterial structure to guide the electromagnetic waves along end-fire which results in gain enhancement up to 13.78 dB at the highest frequency (40GHz). The proposed antenna exhibits high gain, small area and stable radiation pattern as compared to other state-of-the-art designs.
{"title":"High Gain Metamaterial Loaded Antipodal Tapered Slot Antenna for Millimeter Wave Applications","authors":"P. Kumar, Prerna Saxena","doi":"10.1109/SPCOM50965.2020.9179599","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179599","url":null,"abstract":"We propose a novel compact antipodal elliptically tapered slot antenna loaded with left-handed metamaterial over 20-40 GHz for millimetre wave applications. The conventional antipodal tapered slot antenna is modified by incorporating a novel transition between the feed and taper section of the antenna to stabilize the radiation pattern in E-plane and to reduce the beam tilting. We employ a semi-dodecagon shaped dielectric lens with unit cells of a novel broadband metamaterial placed on it to improve the gain of the proposed antenna. The proposed antenna is designed on Rogers RT/Duroid 5880 substrate with dielectric constant of 2.2. The antenna is compact with a volume of $ 30times 49times$ 0.254 $mathrm{m}mathrm{m}^{3}$. The antenna reflection coefficient $(|mathrm{S}_{11}|lt -10mathrm{d}mathrm{B})$ shows good impedance matching over a wide frequency range of 20-40GHz. We exploit the left-handedness of the proposed broadband metamaterial structure to guide the electromagnetic waves along end-fire which results in gain enhancement up to 13.78 dB at the highest frequency (40GHz). The proposed antenna exhibits high gain, small area and stable radiation pattern as compared to other state-of-the-art designs.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130124912","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179602
Swamy Chandra Prakash, Suranjita Ganguly, P. Yadav, M. Raghavan, K. S. Sridharan
The current study aims to evaluate the bimanual trainer, ArmAble, using electrophysiology and kinematic data from a healthy cohort, that can help in creating a reliable rehabilitation schema. We use muscle activation patterns recorded through electromyography in healthy subjects, in order to understand the effect on synergies and activation patterns while using a bimanual trainer. We recorded electromyography from six muscles on either side (including four anti-gravity muscles) and kinematic data, while the subject uses the bimanual trainer to understand the muscular activation in the upper limbs. Experimental conditions included different complexity of reaching tasks and different inclinations. We computed the muscle output as quantified by RMS values and intermuscular coherence, which denotes common cortical drive and coordination between muscles. While inclination did not have a significant effect on RMS, there was a marginal yet non-significant effect on IMC. Whereas the complexity of the reaching task did affect the RMS, while it did not affect IMC. We discuss these results in the context of game design principles for neuro-rehabilitation.
{"title":"Evaluation of a gamified upper-arm bimanual trainer for stroke patients - A healthy cohort study","authors":"Swamy Chandra Prakash, Suranjita Ganguly, P. Yadav, M. Raghavan, K. S. Sridharan","doi":"10.1109/SPCOM50965.2020.9179602","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179602","url":null,"abstract":"The current study aims to evaluate the bimanual trainer, ArmAble, using electrophysiology and kinematic data from a healthy cohort, that can help in creating a reliable rehabilitation schema. We use muscle activation patterns recorded through electromyography in healthy subjects, in order to understand the effect on synergies and activation patterns while using a bimanual trainer. We recorded electromyography from six muscles on either side (including four anti-gravity muscles) and kinematic data, while the subject uses the bimanual trainer to understand the muscular activation in the upper limbs. Experimental conditions included different complexity of reaching tasks and different inclinations. We computed the muscle output as quantified by RMS values and intermuscular coherence, which denotes common cortical drive and coordination between muscles. While inclination did not have a significant effect on RMS, there was a marginal yet non-significant effect on IMC. Whereas the complexity of the reaching task did affect the RMS, while it did not affect IMC. We discuss these results in the context of game design principles for neuro-rehabilitation.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129485813","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179522
Sagar Kavaiya, Dhaval K. Patel, Y. Guan, Sumei Sun, Yoong Choon Chang, J. Lim
In this paper, we study the secrecy problem for a relay-based vehicular network. We assume that the legitimate transmitter, legitimate receiver, and eavesdropper are equipped with a single antenna. By considering various initial positions of the relay, we obtain the statistical knowledge of the received signal-to-noise ratio over $alpha - eta - k - mu$ fading channel under vehicle mobility. Further, we derive an exact closed form expression for outage probability and secrecy outage probability utilizing the amplify-and-forward relay protocol for a two-lane high way scenario. Monte-Carlo simulations are performed to validate the accuracy of the derived analytical expressions.
本文研究了基于中继的车辆网络的保密问题。我们假设合法的发射器、合法的接收器和窃听者配备了一个天线。通过考虑继电器的不同初始位置，得到了车辆移动情况下$alpha - eta - k - mu$衰落信道接收信噪比的统计知识。此外，我们利用放大转发中继协议，推导出两车道高速公路场景下的停电概率和保密停电概率的精确封闭表达式。通过蒙特卡罗仿真验证了所得解析表达式的准确性。
{"title":"On Physical Layer Security over α - η -κ -μ Fading for Relay based Vehicular Networks","authors":"Sagar Kavaiya, Dhaval K. Patel, Y. Guan, Sumei Sun, Yoong Choon Chang, J. Lim","doi":"10.1109/SPCOM50965.2020.9179522","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179522","url":null,"abstract":"In this paper, we study the secrecy problem for a relay-based vehicular network. We assume that the legitimate transmitter, legitimate receiver, and eavesdropper are equipped with a single antenna. By considering various initial positions of the relay, we obtain the statistical knowledge of the received signal-to-noise ratio over $alpha - eta - k - mu$ fading channel under vehicle mobility. Further, we derive an exact closed form expression for outage probability and secrecy outage probability utilizing the amplify-and-forward relay protocol for a two-lane high way scenario. Monte-Carlo simulations are performed to validate the accuracy of the derived analytical expressions.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122560198","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179593
M. C. Coşkun, H. Pfister
Successive cancellation list decoding of polar codes provides very good performance for short to moderate block lengths. However, the list size required to approach the performance of maximum-likelihood decoding is still not well understood theoretically. This work identifies information-theoretic quantities that are closely related to this required list size. It also provides a natural approximation for these quantities that can be computed efficiently even for very long codes. Simulation results are provided for the binary erasure channel as well as the binary-input additive white Gaussian noise channel.
{"title":"Bounds on the List Size of Successive Cancellation List Decoding","authors":"M. C. Coşkun, H. Pfister","doi":"10.1109/SPCOM50965.2020.9179593","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179593","url":null,"abstract":"Successive cancellation list decoding of polar codes provides very good performance for short to moderate block lengths. However, the list size required to approach the performance of maximum-likelihood decoding is still not well understood theoretically. This work identifies information-theoretic quantities that are closely related to this required list size. It also provides a natural approximation for these quantities that can be computed efficiently even for very long codes. Simulation results are provided for the binary erasure channel as well as the binary-input additive white Gaussian noise channel.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122830199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179541
Sachin Kadam, G. Kasbekar
We consider the problem of estimation of the node cardinality of each node type in a heterogeneous wireless network with T types of nodes deployed over a large region, where $Tgeq 2$ is an integer. A mobile base station (MBS), such as that mounted on an unmanned aerial vehicle (UAV), is used in such cases since a single static base station is not sufficient to cover such a large region. The MBS moves around in the region, and makes multiple stops, and at the last stop, it is able to estimate the node cardinalities for the entire region. In this paper, we propose two schemes, viz., HSRC-MI and HSRC-M2, to rapidly estimate the number of nodes of each type. Both the schemes have two phases and they are performed at each stop. We prove that the node cardinality estimates computed using our proposed schemes equal and hence are as accurate as the estimates that would have been obtained if a well known estimation protocol designed for homogeneous networks in prior work is separately executed T times. Using simulations, we show that the numbers of slots required by the proposed schemes, viz., HSRC-MI and HSRCM2, for computing the node cardinality estimates are significantly less than the number of slots required for T separate executions of the above estimation protocol for homogeneous networks.
{"title":"Node Cardinality Estimation Using a Mobile Base Station in a Heterogeneous Wireless Network Deployed Over a Large Region","authors":"Sachin Kadam, G. Kasbekar","doi":"10.1109/SPCOM50965.2020.9179541","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179541","url":null,"abstract":"We consider the problem of estimation of the node cardinality of each node type in a heterogeneous wireless network with T types of nodes deployed over a large region, where $Tgeq 2$ is an integer. A mobile base station (MBS), such as that mounted on an unmanned aerial vehicle (UAV), is used in such cases since a single static base station is not sufficient to cover such a large region. The MBS moves around in the region, and makes multiple stops, and at the last stop, it is able to estimate the node cardinalities for the entire region. In this paper, we propose two schemes, viz., HSRC-MI and HSRC-M2, to rapidly estimate the number of nodes of each type. Both the schemes have two phases and they are performed at each stop. We prove that the node cardinality estimates computed using our proposed schemes equal and hence are as accurate as the estimates that would have been obtained if a well known estimation protocol designed for homogeneous networks in prior work is separately executed T times. Using simulations, we show that the numbers of slots required by the proposed schemes, viz., HSRC-MI and HSRCM2, for computing the node cardinality estimates are significantly less than the number of slots required for T separate executions of the above estimation protocol for homogeneous networks.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"238 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123327227","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179499
Geethu Joseph, C. Murthy, V. J. Mathews
In this paper, we discuss the problem of placing the minimum number of sensors in a given area such that all points in that area are covered by the sensors. We consider a pairwise sensing model where each pair of sensors has an associated sensing area. This model arises in applications like structural health monitoring where a transducer network monitors a structure by exciting one transducer at a time while other transducers collect the data. We first formulate the placement as an optimization problem using a grid-based approach. We present a greedy algorithm to solve the optimization problem. We also present a faster version of the greedy algorithm based on column elimination and recursive updates. To further improve the greedy method, we use a group-greedy strategy. This strategy is a compromise between the conventional greedy method and an exhaustive search. Finally, we apply our algorithm to the structural health monitoring application. Using numerical results, we demonstrate the efficacy of our algorithm and show that the solution is close to the optimal solution.
{"title":"Sensor Placement for A Pairwise Sensing Model: Framework and Algorithms","authors":"Geethu Joseph, C. Murthy, V. J. Mathews","doi":"10.1109/SPCOM50965.2020.9179499","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179499","url":null,"abstract":"In this paper, we discuss the problem of placing the minimum number of sensors in a given area such that all points in that area are covered by the sensors. We consider a pairwise sensing model where each pair of sensors has an associated sensing area. This model arises in applications like structural health monitoring where a transducer network monitors a structure by exciting one transducer at a time while other transducers collect the data. We first formulate the placement as an optimization problem using a grid-based approach. We present a greedy algorithm to solve the optimization problem. We also present a faster version of the greedy algorithm based on column elimination and recursive updates. To further improve the greedy method, we use a group-greedy strategy. This strategy is a compromise between the conventional greedy method and an exhaustive search. Finally, we apply our algorithm to the structural health monitoring application. Using numerical results, we demonstrate the efficacy of our algorithm and show that the solution is close to the optimal solution.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"754 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132651754","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179559
Suraj Srivastava, Mahendrada Sarath Kumar, Amrita Mishra, A. Jagannatham
This work develops a novel scheme for doubly-selective sparse channel estimation in a space-time trellis coded (STTC) multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) wireless systems. Toward this end, a pilot based doubly-selective sparse channel estimation model is developed, wherein the time-evolution of the wireless channel is modeled using a first-order autoregressive (AR1) process. This is followed by the proposed online pilot-based sparse Kalman filter (P-SKF) for channel estimation. The proposed P-SKF scheme combines the advantages of the Kalman filter (KF) and sparse Bayesian learning (SBL) techniques to exploit the temporal correlation as well as the sparse multipath delay profile of the wireless channel. In addition, the proposed PSKF technique also exploits the simultaneous-sparsity across all the transmit-receive antenna pairs for improved channel estimation. The recursive Bayesian Cramér-Rao lower bound (BCRLB) is also derived to benchmark the mean square error (MSE) performance of the proposed technique. Simulation results are presented to evaluate the MSE and frame error rate (FER) performance of the proposed technique and compare with the existing schemes.
{"title":"Sparse Kalman Filtering (SKF) Aided Doubly-Selective Channel Estimation in STTC MIMO-OFDM Systems","authors":"Suraj Srivastava, Mahendrada Sarath Kumar, Amrita Mishra, A. Jagannatham","doi":"10.1109/SPCOM50965.2020.9179559","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179559","url":null,"abstract":"This work develops a novel scheme for doubly-selective sparse channel estimation in a space-time trellis coded (STTC) multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) wireless systems. Toward this end, a pilot based doubly-selective sparse channel estimation model is developed, wherein the time-evolution of the wireless channel is modeled using a first-order autoregressive (AR1) process. This is followed by the proposed online pilot-based sparse Kalman filter (P-SKF) for channel estimation. The proposed P-SKF scheme combines the advantages of the Kalman filter (KF) and sparse Bayesian learning (SBL) techniques to exploit the temporal correlation as well as the sparse multipath delay profile of the wireless channel. In addition, the proposed PSKF technique also exploits the simultaneous-sparsity across all the transmit-receive antenna pairs for improved channel estimation. The recursive Bayesian Cramér-Rao lower bound (BCRLB) is also derived to benchmark the mean square error (MSE) performance of the proposed technique. Simulation results are presented to evaluate the MSE and frame error rate (FER) performance of the proposed technique and compare with the existing schemes.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131106065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179492
Shashank Vatedka, P. Vontobel
Currently the best deterministic polynomial-time algorithm for approximating the permanent of a non-negative matrix is based on minimizing the Bethe free energy function of a certain normal factor graph (NFG). In order to improve the approximation guarantee, we propose a modified NFG with fewer cycles, but still manageable function-node complexity; we call the approximation obtained by minimizing the function of the modified normal factor graph the modified Bethe permanent. For nonnegative matrices of size $3times 3$, we give a tight characterization of the modified Bethe permanent. For non-negative matrices of size $ntimes n$ with $ngeq 3$, we present a partial characterization, along with promising numerical results. The analysis of the modified NFG is also interesting because of its tight connection to an NFG that is used for approximating a permanent-like quantity in quantum information processing.
{"title":"Modified Bethe Permanent of a Nonnegative Matrix","authors":"Shashank Vatedka, P. Vontobel","doi":"10.1109/SPCOM50965.2020.9179492","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179492","url":null,"abstract":"Currently the best deterministic polynomial-time algorithm for approximating the permanent of a non-negative matrix is based on minimizing the Bethe free energy function of a certain normal factor graph (NFG). In order to improve the approximation guarantee, we propose a modified NFG with fewer cycles, but still manageable function-node complexity; we call the approximation obtained by minimizing the function of the modified normal factor graph the modified Bethe permanent. For nonnegative matrices of size $3times 3$, we give a tight characterization of the modified Bethe permanent. For non-negative matrices of size $ntimes n$ with $ngeq 3$, we present a partial characterization, along with promising numerical results. The analysis of the modified NFG is also interesting because of its tight connection to an NFG that is used for approximating a permanent-like quantity in quantum information processing.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121475340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179634
G. Senthil, K. Nandhakumar, G. R. S. Subrahmanyam
Handwritten Text Recognition (HTR) of Hindi Documents is a challenging research problem of interest which could enable digitization of millions of official documents. Due to challenges in character segmentation, Segmentation-free Word Recognition is the preferred approach. Lack of a large, diverse Hindi Handwritten Word dataset for pre-training deep learning architectures is a pressing issue. In this paper, we propose a novel way of generating diverse Handwritten Hindi Word images using only Handwritten Hindi Characters and further analyze its effectiveness in enabling Few Instance Learning of Handwritten Hindi Documents.
{"title":"Handwritten Hindi Word Generation to enable Few Instance Learning of Hindi Documents","authors":"G. Senthil, K. Nandhakumar, G. R. S. Subrahmanyam","doi":"10.1109/SPCOM50965.2020.9179634","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179634","url":null,"abstract":"Handwritten Text Recognition (HTR) of Hindi Documents is a challenging research problem of interest which could enable digitization of millions of official documents. Due to challenges in character segmentation, Segmentation-free Word Recognition is the preferred approach. Lack of a large, diverse Hindi Handwritten Word dataset for pre-training deep learning architectures is a pressing issue. In this paper, we propose a novel way of generating diverse Handwritten Hindi Word images using only Handwritten Hindi Characters and further analyze its effectiveness in enabling Few Instance Learning of Handwritten Hindi Documents.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115152088","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-07-01DOI: 10.1109/SPCOM50965.2020.9179578
Mridul Mahajan, Tryambak Bhattacharjee, Arya Krishnan, Priya Shukla, G. Nandi
For a robot to perform complex manipulation tasks, it is necessary for it to have a good grasping ability. However, vision based robotic grasp detection is hindered by the unavailability of sufficient labelled data. Furthermore, the application of semi-supervised learning techniques to grasp detection is underexplored. In this paper, a semi-supervised learning based grasp detection approach has been presented, which models a discrete latent space using a Vector Quantized Variational AutoEncoder (VQ-VAE). To the best of our knowledge, this is the first time a Variational AutoEncoder (VAE) has been applied in the domain of robotic grasp detection. The VAE helps the model in generalizing beyond the Cornell Grasping Dataset (CGD) despite having a limited amount of labelled data by also utilizing the unlabelled data. This claim has been validated by testing the model on images, which are not available in the CGD. Along with this, we augment the Generative Grasping Convolutional Neural Network (GGCNN) architecture with the decoder structure used in the VQ-VAE model with the intuition that it should help to regress in the vector-quantized latent space. Subsequently, the model performs significantly better than the existing approaches which do not make use of unlabelled images to improve the grasp.
{"title":"Robotic Grasp Detection By Learning Representation in a Vector Quantized Manifold","authors":"Mridul Mahajan, Tryambak Bhattacharjee, Arya Krishnan, Priya Shukla, G. Nandi","doi":"10.1109/SPCOM50965.2020.9179578","DOIUrl":"https://doi.org/10.1109/SPCOM50965.2020.9179578","url":null,"abstract":"For a robot to perform complex manipulation tasks, it is necessary for it to have a good grasping ability. However, vision based robotic grasp detection is hindered by the unavailability of sufficient labelled data. Furthermore, the application of semi-supervised learning techniques to grasp detection is underexplored. In this paper, a semi-supervised learning based grasp detection approach has been presented, which models a discrete latent space using a Vector Quantized Variational AutoEncoder (VQ-VAE). To the best of our knowledge, this is the first time a Variational AutoEncoder (VAE) has been applied in the domain of robotic grasp detection. The VAE helps the model in generalizing beyond the Cornell Grasping Dataset (CGD) despite having a limited amount of labelled data by also utilizing the unlabelled data. This claim has been validated by testing the model on images, which are not available in the CGD. Along with this, we augment the Generative Grasping Convolutional Neural Network (GGCNN) architecture with the decoder structure used in the VQ-VAE model with the intuition that it should help to regress in the vector-quantized latent space. Subsequently, the model performs significantly better than the existing approaches which do not make use of unlabelled images to improve the grasp.","PeriodicalId":208527,"journal":{"name":"2020 International Conference on Signal Processing and Communications (SPCOM)","volume":"1053 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123150634","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：