UAV Track Planning Algorithm Based on Graph Attention Network and Deep Q Network

Abstract

This study investigates the enhancement of the traditional Deep Q-Network (DQN) trader model through the integration of cutting-edge techniques such as Prioritized Experience Replay, Regularized Q-Learning, Noisy Networks, Dueling DQN, and Double DQN. Through rigorous empirical testing on a spectrum of financial instruments including BTC/USD and AAPL, the research delineates clear performance improvements over the original model. The augmented DQN trader showcases remarkable gains, with the DQN-vanilla variant achieving an arithmetic return increase from 261% to 287% and an enhanced Sharpe Ratio, indicative of better risk-adjusted returns. The innovative use of CNN1D and CNN2D architectures further amplifies returns, highlighting the efficacy of convolutional layers in capturing market dynamics. The enhanced model's consistency is evident in its application to AAPL stock, where substantial gains are observed. The DQN-pattern variant maintains a stable performance, while the CNN-based models demonstrate their architectural potency through exceptional returns. These results not only eclipse those of the baseline model but also underscore the potential of utilizing convolutional neural networks within financial trading systems. The study's findings confirm that the application of these sophisticated deep learning techniques within a reinforcement learning framework can significantly improve the performance of automated trading strategies. This performance consistency across various financial instruments underpins the importance of continued innovation in this domain. The research concludes by advocating for future exploration into new reinforcement learning methods and their potential to expand the model's effectiveness across a wider array of financial environments. The abstract encapsulates the core advancements and implications of the study, setting the stage for future developments in the realm of AI-driven financial trading.

Comprehensive Review of Drones Collision Avoidance Schemes: Challenges and Open Issues

“…It employed a partly visible state that focuses primarily on the UAV's immediate surroundings rather than the whole deployment region, which results in a slow UAV and substantial computing costs. UAV track planning approach based on Graph Attention Network and Deep Q Network to solve the problem of mission failure caused by erroneous data acquired from UAV during flight present by [128]. It uses the camera to capture photographs and then runs them through a ResNet that has previously been trained to recognize and classify different items within those pictures.…”

“…Geometric [135], [143], force field [131], [132], optimum trajectory [119], and Markov Decision Process (MDP) [108], [110] approaches are the most used [144]. Numerous suggested algorithms using mathematical approaches [5], [72], [89], [105], [129] and ML [100], [128] need a strong processor and memory space, which cannot accommodate most UAVs operating in the real world. On the other hand, owing to the delay and time necessary for setting up and training these algorithms, their implementation in real-time is not anticipated, and the present latency may result in operation failure or drone collisions.…”

Relation extraction: advancements through deep learning and entity-related features

“…2019 ) and graph convolutional networks (GCNs) (Hu et al. 2022 ; Zhu et al. 2019 ), have been proposed to capture multi-level features and global dependencies.…”

A Spatial-Reduction Attention-Based BiGRU Network for Water Level Prediction

“…The attention mechanism is a method for rapidly selecting high-value information from huge amounts of information. Based on this mechanism, Hu et al [33] developed a runoff forecasting model and predicted the runoff of the basin, and the results demonstrated that this improved method can enhance the accuracy of runoff forecasting.…”