Handwritten Image Detection using DCGAN with SIFT and ORB Optical Features

– Rohan Dubey; Ipshita Das

Abstract:

This paper presents a novel approach for the Handwritten Image Detection task that combines the deep-convolution GAN (DCGAN) model with optical feature-extraction techniques, including Oriented FAST with Rotated BRIEF (ORB) and Scale Invariant Feature Transform (SIFT) using the Bag-of-Features approach. Our model uses the image and its optical features (SIFT or ORB) as the inputs to a network trained to recognise the shape and orientation of characters from each class. This study examines the effect of both regular and dense optical features. We evaluate our algorithm on the MNIST, EMNIST, and MADBase datasets. The results indicate that GANs using optical feature approaches are superior to classical GANs, with an accuracy improvement of 4.88% in the MNSIT dataset, 6.89% in the EMNIST dataset, and 0.62% in the MADBase dataset. In addition, our approach surpasses the state-of-the-art methods on the MADBase dataset.

Date of Conference: 03-04 March 2023 DOI: 10.1109/ISCON57294.2023.10112139
Date Added to IEEE Xplore: 04 May 2022 Publisher: IEEE
Conference: 2023 6th International Conference on Information Systems and Computer Networks (ISCON) Conference Location: Mathura, India

Cite this:
R. Dubey and I. Das, “Handwritten Image Detection using DCGAN with SIFT and ORB Optical Features,” 2023 6th International Conference on Information Systems and Computer Networks (ISCON), Mathura, India, 2023, pp. 1-6, doi: 10.1109/ISCON57294.2023.10112139.

Coordinated load frequency control of a smart hybrid power system using the DEMA-TD3 algorithm

– Renuka Loka; Rohan Dubey; Alivelu Manga Parimi

Abstract:

In this paper, a novel deep reinforcement (DRL) algorithm is proposed for the coordination of controllers in a multi-source hybrid power system (HPS) for load frequency control (LFC). Decentralized controllers regulate the distributed sources with uncertainties in load and source power deviations to maintain the power system frequency. Coordination of decentralized controllers becomes challenging in the presence of dynamic disturbances. A novel LFC framework is introduced via a dynamic environment-based multi-agent twin delayed deep deterministic policy gradient (DEMA-TD3) algorithm to coordinate the decentralized controllers. Moreover, system stability is a concern in systems with renewable integration during islanding operations. To achieve successful islanding, a smart home inverter (SHI) based coordinated control is proposed. Virtual inertia (VI) is added to provide additional frequency support for robust control. Various simulation test scenarios are simulated to analyze the effectiveness of the proposed novel LFC architecture in different modes of operation.

Journal : Control Engineering Practice DOI: 10.1016/j.conengprac.2023.105480
Date Added : 06 March 2023 Publisher: Elsevier (ScienceDirect)

Cite this:
R. Loka, R. Dubey and A. M. Parimi, “Coordinated load frequency control of a smart hybrid power system using the DEMA-TD3 algorithm”, Control Engineering Practice, 2023, volume-134: 10.1016/j.conengprac.2023.105480.

Maintaining the Frequency of AI-based Power System Model using Twin Delayed DDPG(TD3) Implementation

– Rohan Dubey; Renuka Loka; Alivelu Manga Parimi

Abstract:

This paper proposes a multi-agent deep reinforcement learning (MA-DRL) method for load frequency control of a renewable energy single-area power system in a continuous action-space domain. This method can non-linearly adapt the control strategies for cooperative LFC control through off-policy learning. Multi-agent twin delayed deep deterministic policy gradient (TD3) is proposed to adjust and refine the control system parameters considering variational load and source behaviour. Implementation of the model requires only local information for each control area to achieve an optimal control state. Comparison between TD3 and DDPG model proves the edge of TD3 model. Simulations and numerical data comparison on a renewable energy single-area power system demonstrate that the proposed model can successfully reduce control errors and stochastic frequency deviations caused by load and renewable power fluctuations.

Date of Conference: 21-22 January 2022 DOI: 10.1109/PARC52418.2022.9726615
Date Added to IEEE Xplore: 09 March 2022 Publisher: IEEE
Conference: 2022 2nd International Conference on Power Electronics & IoT Applications in Renewable Energy and its Control (PARC) Conference Location: Mathura, India

Cite this:
R. Dubey, R. Loka and A. M. Parimi, “Maintaining the Frequency of AI-based Power System Model using Twin Delayed DDPG(TD3) Implementation”, 2022 2nd International Conference on Power Electronics & IoT Applications in Renewable Energy and its Control (PARC), 2022, pp. 1-4, doi: 10.1109/PARC52418.2022.9726615.