A brief guide to multi-objective reinforcement learning and planning JAAMAS track
- Authors: Hayes, Conor , Bargiacchi, Eugenio , Källström, Johan , Macfarlane, Matthew , Reymond, Mathieu , Verstraeten, Timothy , Zintgraf, Luisa , Dazeley, Richard , Heintz, Frederik , Howley, Enda , Irissappane, Aathirai , Mannion, Patrick , Nowé, Ann , Ramos, Gabriel , Restelli, Marcello , Vamplew, Peter , Roijers, Diederik
- Date: 2023
- Type: Text , Conference paper
- Relation: 22nd International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2023, London, 29 May to 2 June 2023, Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS Vol. 2023-May, p. 1988-1990
- Full Text:
- Reviewed:
- Description: Real-world sequential decision-making tasks are usually complex, and require trade-offs between multiple - often conflicting - objectives. However, the majority of research in reinforcement learning (RL) and decision-theoretic planning assumes a single objective, or that multiple objectives can be handled via a predefined weighted sum over the objectives. Such approaches may oversimplify the underlying problem, and produce suboptimal results. This extended abstract outlines the limitations of using a semi-blind iterative process to solve multi-objective decision making problems. Our extended paper [4], serves as a guide for the application of explicitly multi-objective methods to difficult problems. © 2023 International Foundation for Autonomous Agents and Multiagent Systems (www.ifaamas.org). All rights reserved.
A case for causal loop diagrams to model electronic health records ecosystems
- Authors: Hashmi, Mustafa , McInnes, Angelique , Sahama, Tony , Stranieri, Andrew
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 Australasian Computer Science Week, ACSW 2023, Melbourne, Australia, 31 January-3 February 2023, ACSW '23: Proceedings of the 2023 Australasian Computer Science Week p. 238-239
- Full Text: false
- Reviewed:
- Description: Causal loop diagrams (CLD) that emerged from systems thinking disciplines have been used to simulate complex inter-dependencies between causal factors in diverse phenomena. This paper highlights a process for generating a casual loop diagrams to represent the quality of electronic health record (EHR) ecosystem in a medical context. The quality inherent in the use of electronic health records for specific clinical purposes is taken to depend on factors including data integrity, reliability, relevance, timeliness and completeness. By improving the electronic health record ecosystem quality, health care providers can enhance their data sharing practices, and personalised patient care, while reducing the probabilities of medical errors. Ultimately the CLD can be used to run multiple simulations for several clinical case scenarios to understand the impact of various case phenomena on the quality of the electronic health record ecosystem. © 2023 ACM.
A hybrid PWM strategy for SMES integrated grid-feeding transformerless PV Inverters
- Authors: Mondal, Sudipto , Biswas, Shuvra , Bin Islam, Md Sabbir , Islam, Md Rabiul , Shah, Rakibuzzaman
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 IEEE International Conference on Applied Superconductivity and Electromagnetic Devices, ASEMD 2023, Tianjin, China, 27-29 October 2023, 2023 IEEE International Conference on Applied Superconductivity and Electromagnetic Devices, ASEMD 2023
- Full Text:
- Reviewed:
- Description: This paper presents a new hybrid pulse width modulation (PWM) strategy for improving the performance of superconducting magnetic energy storage (SMES) integrated solar photovoltaic (PV) based grid-feeding transformerless H6 inverter systems. The proposed PWM strategy offers significantly lower total harmonic distortion (THD) of output voltage compared to existing unipolar PWM (UPPWM), bipolar PWM (BPPWM), and some well-established advanced PWM techniques. A new carrier and reference signals are used to develop the proposed hybrid PWM strategy, which ensures low leakage current compared to existing counterparts. The presented inverter system is simulated in MATLAB/Simulink and PLECS computer simulation environment. Apart from simulated results, some preliminary experimental test results are also shown to validate the superiority of the proposed PWM strategy. © 2023 IEEE
A new tour on the subdifferential of the Supremum function
- Authors: Hantoute, Abderrahim , López-Cerdá, Marco
- Date: 2023
- Type: Text , Conference paper
- Relation: International Meeting on Functional Analysis and Continuous Optimization, IMFACO 2022, Elche, Spain, 16-17 June 2022, Functional Analysis and Continuous Optimization In Honour of Juan Carlos Ferrando's 65th Birthday, Elche, Spain, June 16–17, 2022 Vol. 424, p. 167-194
- Full Text: false
- Reviewed:
- Description: This chapter is a survey presenting various characterizations of the subdifferential of the pointwise supremum of convex functions, as well as some featured applications. We gathered here the main outcomes we obtained in a series of recent papers, dealing with different models, assumptions and scenarios. Starting by the maximum generality framework, we move after to particular contexts in which some continuity and compactness assumptions are either imposed or inforced via processes of compactification of the index set and regularization of the data functions. Some relevant applications of the general results are presented, in particular to derive rules for the subdifferential of the sum, and for convexifying a general (unconstrained) optimization problem. The last section gives some specific constraint qualifications for the convex optimization problem with an arbitrary set of constraints, and also contains different sets of KKT-type optimality conditions appealing to the subdifferential of the supremum function. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
A Novel Common Grounded Type 1-\varphi Five-Level Boost PV Inverter with Reduced Device Count
- Authors: Ardashir, Jaber , Ghadim, Hadi , Heydari, Dorsa , Hu, Jiefeng
- Date: 2023
- Type: Text , Conference paper
- Relation: 8th International Conference on Technology and Energy Management, ICTEM 2023, Babol, Iran, 8-9 February 2023, 8th International Conference on Technology and Energy Management, ICTEM 2023
- Full Text: false
- Reviewed:
- Description: Common-grounded multi-level inverters with the input voltage boost ability are more common for grid-tied PV system applications. In this paper, a new five-level inverter topology is introduced that can generate a five-level voltage at the output stage, statically increase the input voltage, fully transfer the reactive power, eliminate the leakage current due to common-grounded capability, reduce the number of the power electronic components, self-balancing capability of switched capacitors voltage based on operation modes, and equality of switched capacitors voltage. The operation of the proposed inverter along with its comparison with other similar inverter topologies and the simulation results of the proposed inverter under different loads are demonstrates in this paper. The results indicate that the performance of the proposed inverter in the grid-connected PV system is best than other compared topologies. Also, experimental results with a prototype have been presented, which validates the simulation results and proves the applicability of the proposed inverter in grid-tied PV systems. © 2023 IEEE.
A robust ensemble regression model for reconstructing genetic networks
- Authors: Gamage, Hasini , Chetty, Madhu , Lim, Suryani , Hallinan, Jennifer , Nguyen, H.
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 International Joint Conference on Neural Networks, IJCNN 2023 Vol. 2023-June
- Full Text: false
- Reviewed:
- Description: Genetic networks contain important information about biological processes, including regulatory relationships and gene-gene interactions. Numerous methods, using high-dimensional gene expression data have been developed to capture these interactions. These gene expression data, generated using high-throughput technologies, are prone to noise. However, most existing network inference methods are unable to cope with noisy data, making genetic network reconstruction challenging. In this paper, we propose a novel ensemble regression model combining quantile regression and cross-validated Ridge regression, RidgeCV, to infer interactions from noisy gene expression data. The application of quantile regression to GRN inference is novel, and its design makes it appropriate for noisy data. RidgeCV also addresses other important issues, such as data overfitting and multicollinearity. First, each regression method is independently applied to gene expression data and the output of these methods, in the form of ranked gene lists, is aggregated using a novel gene score-based method by considering the gene rank and model importance. The model importance score is evaluated based on an adjusted coefficient of determination. This method implicitly includes majority voting by averaging each gene score value across all models. The proposed model was tested on the DREAM4 datasets and publicly available small-scale real-world network datasets. Experiments with noisy datasets showed that the proposed ensemble model is more accurate and efficient than other state-of-the-art methods. © 2023 IEEE.
A study into the impact of data breaches of electronic health records
- Authors: Pilla, Ravi , Oseni, Taiwo , Stranieri, Andrew
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 Australasian Computer Science Week, ACSW 2023, Melbourne Australia, 31 January-3 February 2023, ACSW '23: Proceedings of the 2023 Australasian Computer Science Week p. 252-254
- Full Text: false
- Reviewed:
- Description: The research study deals with electronic health records (EHRs) data breaches, their impact., Electronic health records play an important role in digital healthcare services. However, confidentiality and integrity of sensitive EHRs are critical to ensure patient privacy. Although the existing traditional cybersecurity practices provide some protection, they cannot prevent EHRs data breaches. Therefore, this research's primary focus will be critically reviewing the impact of data breaches and current cybersecurity practices. Finally, the paper's key findings highlight the type of cyberattacks and options to reduce them. © 2023 ACM.
A survey on image classification of lightweight convolutional neural network
- Authors: Liu, Ying , Xiao, Peng , Fang, Jie , Zhang, Dengsheng
- Date: 2023
- Type: Text , Conference paper
- Relation: 19th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery, ICNC-FSKD 2023, Harbin, China, 29-31 July 2023, 2023 19th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD)
- Full Text: false
- Reviewed:
- Description: In recent years, deep neural networks have achieved tremendous success in image classification in both academic and industrial settings. However, the high hardware requirements imposed by their intensive and complex computations pose a challenge for deployment on low-storage devices. To address this challenge, lightweight networks provide a viable solution. This paper provides a detailed review of recent lightweight image classification algorithms, which can be categorized into low-redundancy network model design and neural network compression algorithms. The former reduces network computations by replacing traditional convolution with efficient lightweight convolution, while the latter reduces redundancy in the network by employing methods such as network pruning, knowledge distillation, and parameter quantization. We summarize the experimental results of some classical models and algorithms on ImageNet2012 and CIFAR-10 datasets, and analyze the characteristics, advantages and disadvantages of these models respectively. Finally, future research directions for lightweight algorithms in the field of image classification are identified. © 2023 IEEE.
Artificial intelligence enabled digital twin for predictive maintenance in industrial automation system : a novel framework and case study
- Authors: Siddiqui, Mustafa , Kahandawa, Gayan , Hewawasam, Hasitha
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 IEEE International Conference on Mechatronics, ICM 2023, Leicestershire UK, 15-17 March 2023, Proceedings - 2023 IEEE International Conference on Mechatronics, ICM 2023
- Full Text: false
- Reviewed:
- Description: Industrial automation systems are excessively used in advanced manufacturing environments. These systems are always prone to failure which not only disturbs smooth manufacturing operations but can also cause injuries to operators. Therefore, in this research, a novel predictive maintenance algorithm is proposed that can be used to detect anomalies in the automation system to avoid asset failure. Artificial Intelligence enabled Digital Twin model was used to detect early anomalies to avoid catastrophic effects of equipment failure. Real-time sensor data were used to validate the proposed novel algorithm. The data were recorded via sensors mounted on the physical system. This paper presents the effectiveness of the proposed algorithm to detect anomalies in industrial automation systems under faulty conditions. © 2023 IEEE.
Bilateral insider threat detection : harnessing standalone and sequential activities with recurrent neural networks
- Authors: Manoharan, Phavithra , Hong, Wei , Yin, Jiao , Zhang, Yanchun , Ye, Wenjie , Ma, Jiangang
- Date: 2023
- Type: Text , Conference paper
- Relation: 24th International Conference on Web Information Systems Engineering, WISE 2023, Melbourne, 25-27 October 2023, Web Information Systems Engineering – WISE 2023, 24th International Conference, Melbourne, VIC, Australia, October 25–27, 2023, Proceedings Vol. 14306 LNCS, p. 179-188
- Full Text: false
- Reviewed:
- Description: Insider threats involving authorised individuals exploiting their access privileges within an organisation can yield substantial damage compared to external threats. Conventional detection approaches analyse user behaviours from logs, using binary classifiers to distinguish between malicious and non-malicious users. However, existing methods focus solely on standalone or sequential activities. To enhance the detection of malicious insiders, we propose a novel approach: bilateral insider threat detection combining RNNs to incorporate standalone and sequential activities. Initially, we extract behavioural traits from log files representing standalone activities. Subsequently, RNN models capture features of sequential activities. Concatenating these features, we employ binary classification to detect insider threats effectively. Experiments on the CERT 4.2 dataset showcase the approach’s superiority, significantly enhancing insider threat detection using features from both standalone and sequential activities. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
Clinically prioritized data visualization in remote patient monitoring
- Authors: Arora, Teena , Balasubramanian, Venki , Stranieri, Andrew , Neupane, Arun
- Date: 2023
- Type: Text , Conference paper
- Relation: 19th IEEE International Conference on Wireless and Mobile Computing, Networking and Communications, WiMob 2023, Montreal, Canada, 21-23 June 2023, International Conference on Wireless and Mobile Computing, Networking and Communications Vol. 2023-June, p. 5-12
- Full Text: false
- Reviewed:
- Description: Understanding and integrating physiological data collected from wearable sensors in remote patient monitoring (RPM) is challenging. Data streams may be interrupted due to the sensor's sensitivity, movement, and electromagnetic interference leading to inconsistent, missing, and inaccurate data. Existing approaches to summarize data flows into a single score such as the traditional Modified early warning score (MEWS) is limited. Data visualization approaches have the potential to address this challenge, but few studies have focused on visualization of RPM streams. The study presents a transformation of observed raw RPM physiological data into parameters identified as trust, frequency, slope, and trend. This facilitated visualization and enabled automated assessments of prioritized alerts. Experimental results have shown that the transformations led to the prioritization of clinically significant conditions, and improved visualization has the potential to better support clinical decisions compared with traditional MEWS. © 2023 IEEE.
Comprehensive analysis of feature extraction techniques and runtime performance evaluation for phishing detection
- Authors: Nath, Subrata , Islam, Mohammad , Chowdhury, Abdullahi , Rashid, Mohammad , Islam, Maheen , Jabid, Taskeed , Naha, Ranesh
- Date: 2023
- Type: Text , Conference paper
- Relation: 6th International Conference on Applied Computational Intelligence in Information Systems, ACIIS 2023, Bandar Seri Bagawan, Brunei, 23-25 October 2023, 2023 6th International Conference on Applied Computational Intelligence in Information Systems: Intelligent and Resilient Digital Innovations for Sustainable Living, ACIIS 2023 - Proceedings
- Full Text: false
- Reviewed:
- Description: The digital landscape is continually evolving, bringing with it numerous cybersecurity challenges, notably the rise of phishing websites targeting unsuspecting users. These deceptive websites jeopardize digital identities, emphasizing the critical need for precise detection mechanisms. This research provides a deep analysis of feature extraction nuances and critically evaluates the runtime performance of detection models. Through intensive refinement of Random Forest classification models, an integrative approach is adopted, which encompasses feature selection, outlier mitigation, and hyperparameter optimization using advanced data mining techniques. Leveraging a pre-established dataset with 87 distinct features from 11,430 URLs, this research narrows down the features to a pivotal set of 56. The outcome is a robust model that achieves an accuracy of 97.069% and a precision rate of 97.326%. A noteworthy aspect of this study is the incorporation of ensemble models, which amplify prediction accuracy by harnessing the capabilities of multiple algorithms. By employing the ensemble approach, the research ensures the model's heightened accuracy and adaptability, making it resilient against ever-changing phishing strategies. The findings underscore the symbiotic relationship between comprehensive feature extraction techniques and the paramount importance of runtime efficiency, laying the groundwork for a fortified digital landscape. © 2023 IEEE.
Data-efficient graph learning meets ethical challenges
- Authors: Tang, Tao
- Date: 2023
- Type: Text , Conference paper
- Relation: 16th ACM International Conference on Web Search and Data Mining, WSDM 2023, Singapore, 27 February to 3 March 2023, WSDM 2023 - Proceedings of the 16th ACM International Conference on Web Search and Data Mining p. 1218-1219
- Full Text: false
- Reviewed:
- Description: Recommender systems have achieved great success in our daily life. In recent years, the ethical concerns of AI systems have gained lots of attention. At the same time, graph learning techniques are powerful in modelling the complex relations among users and items under recommender system applications. These graph learning- based methods are data hungry, which brought a significant data efficiency challenge. In this proposal, I introduce my PhD research from three aspects: 1) Efficient privacy-preserving recommendation for imbalanced data. 2) Efficient recommendation model training for Insufficient samples. 3) Explainability in the social recommendation. Challenges and solutions of the above research problems have been proposed in this proposal. © 2023 Owner/Author.
Distributed formation trajectory planning for multi-vehicle systems
- Authors: Nguyen, Binh , Nghiem, Truong , Nguyen, Linh , Nguyen, Tung , La, Hung , Sookhak, Mehdi , Nguyen, Thang
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 American Control Conference, ACC 2023 Vol. 2023-May, p. 1325-1330
- Full Text: false
- Reviewed:
- Description: This paper addresses the problem of distributed formation trajectory planning for multi-vehicle systems with collision avoidance among vehicles. Unlike some previous distributed formation trajectory planning methods, our proposed approach offers great flexibility in handling computational tasks for each vehicle when the global formation of all the vehicles changes. It affords the system the ability to adapt to the computational capabilities of the vehicles. Furthermore, global formation constraints can be handled at any selected vehicles. Thus, any formation change can be effectively updated without recomputing all local formations at all the vehicles. To guarantee the above features, we first formulate a dynamic consensus-based optimization problem to achieve desired formations while guaranteeing collision avoidance among vehicles. Then, the optimization problem is effectively solved by ADMM-based or alternating projection-based algorithms, which are also presented. Theoretical analysis is provided not only to ensure the convergence of our method but also to show that the proposed algorithm can surely be implemented in a fully distributed manner. The effectiveness of the proposed method is illustrated by a numerical example of a 9-vehicle system. © 2023 American Automatic Control Council.
Effect of wood/binder ratio, slag/binder ratio, and alkaline dosage on the compressive strength of wood-geopolymer composites
- Authors: Gigar, Firensenay , Khennane, Amar , Liow, Jong-leng , Tekle, Biruk
- Date: 2023
- Type: Text , Conference paper
- Relation: International Symposium of the International Federation for Structural Concrete, fib Symposium 2023, Istanbul, 5-7 June 2023, Building for the Future: Durable, Sustainable, Resilient: Vol. 349 LNCE, p. 658-667
- Full Text: false
- Reviewed:
- Description: The impact of building construction on the environment is significant. Occupying large land areas (urban footprint), buildings are one of the most important consumers of resources and raw materials. They are responsible for 38% of greenhouse gas (GHG) emissions in both developed and developing countries. Therefore, incorporating sustainability and resilience into all aspects of urban infrastructure has become necessary. To curb emissions, part of the answer lies in the use of construction and building materials made from recycled materials. Bio-sourced materials, like wood chips, combined with a cementitious matrix, offer an alternative to conventional materials. They are sustainable, lightweight, and have good thermal insulation. However, because of their inferior mechanical strength, they have limited use as load-bearing structural parts. Furthermore, the use of Portland cement as a binder still poses some challenges due to its high carbon footprint. This study investigates the potential of wood-geopolymer composites for better mechanical performance and environmental sustainability. A 6x2x2x2 fractional factorial-based experimental design was used to simultaneously study the effect of slag content, wood binder ratio, and alkaline on the compressive strength of the wood-geopolymer composite. The experiments showed encouraging results for developing ambient cured wood geopolymer composites. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Efficient graph learning for anomaly detection systems
- Authors: Febrinanto, Falih
- Date: 2023
- Type: Text , Conference paper
- Relation: 16th ACM International Conference on Web Search and Data Mining, WSDM 2023, Singapore, 27 February to 3 March 2023, WSDM 2023 - Proceedings of the 16th ACM International Conference on Web Search and Data Mining p. 1222-1223
- Full Text: false
- Reviewed:
- Description: Anomaly detection plays a significant role in preventing from detrimental effects of abnormalities. It brings many benefits in real-world sectors ranging from transportation, finance to cybersecurity. In reality, millions of data do not stand independently, but they might be connected to each other and form graph or network data. A more advanced technique, named graph anomaly detection, is required to model that data type. The current works of graph anomaly detection have achieved state-of-the-art performance compared to regular anomaly detection. However, most models ignore the efficiency aspect, leading to several problems like technical bottlenecks. This project mainly focuses on improving the efficiency aspect of graph anomaly detection while maintaining its performance. © 2023 Owner/Author.
Elastic step DDPG : multi-step reinforcement learning for improved sample efficiency
- Authors: Ly, Adrian , Dazeley, Richard , Vamplew, Peter , Cruz, Francisco , Aryal, Sunil
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 International Joint Conference on Neural Networks, IJCNN 2023 Vol. 2023-June
- Full Text: false
- Reviewed:
- Description: A major challenge in deep reinforcement learning is that it requires more data to converge to an policy for complex problems. One way to improve sample efficiency is to use n-step updates to reduce the number of samples required to converge to a good policy. However n-step updates are known to be brittle and difficult to tune. Elastic Step DQN has shown that it is possible to automate the value of n in DQN to solve problems involving discrete action spaces, however the efficacy of the technique when applied on more complex problems and against problems with continuous action spaces is yet to be shown. In this paper we adapt the innovations proposed by Elastic Step DQN onto the DDPG algorithm and show empirically that Elastic Step DDPG is able to achieve a much stronger final training policy and is more sample efficient than DDPG. © 2023 IEEE.
Embodied carbon footprint analysis of signage industry : insights from two case studies
- Authors: Paresi, Prudvireddy , Javidan, Fatemeh , Sparks, Paul
- Date: 2023
- Type: Text , Conference paper
- Relation: International Conference on Green Building, ICoGB 2023, Malmo, Sweden, 19-21 May 2023, Proceedings of 2023 International Conference on Green Building p. 69-76
- Full Text: false
- Reviewed:
- Description: Embodied carbon has recently become a hot topic among environmentalists and designers, especially after the Paris Agreement on climate change. Embodied carbon refers to the carbon emissions associated with the manufacturing and transportation of building materials and the process of construction. The “Global Status Report for Buildings and Construction” report estimated that the building and construction sector alone contributed nearly 37–39% of global carbon emissions in 2017–2020. To tackle embodied carbon, the World Green Building Council (WorldGBC) has set a bold vision to reduce it by at least 40% by 2030 and achieve net-zero operating carbon in all new buildings. The signage industry plays a significant role in the building industry, as signages are a key component of buildings. Signages serve multiple purposes, such as providing information, enhancing brand identity, and promoting safety. Therefore, it is essential to understand the embodied carbon emissions associated with signage materials used to minimise the overall carbon emissions of construction projects. The present paper aims to study the embodied carbon footprint of the signage industry with the help of two case studies. The embodied carbon factors required while estimating the overall footprint of the signages are taken from Environmental Performance in Construction (EPiC) database. The study identifies the aluminum as the major contributor of the embodied emissions in the signage projects. This study provides insight into the other sources of embodied carbon and makes more informed decisions while selecting signage materials used in designs to create sustainable and economic projects. This information helps to increase sustainability and reduce the carbon footprint of signage projects in the early decision-making stages. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Exploring the relationship between testosterone and diabetes within the UK Biobank data
- Authors: Oatley, Giles
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 Australasian Computer Science Week, ACSW 2023, Melbourne Australia, 31 January-3 February 2023, ACSW '23: Proceedings of the 2023 Australasian Computer Science Week p. 244-247
- Full Text: false
- Reviewed:
- Description: The UK Biobank (UKB) cohort data aims to improve the prevention, diagnosis, and treatment of a wide range of serious diseases, including diabetes. Presented is a population-based retrospective cohort study to explore the relationship between steroid hormones and the prevalence of diabetes. In particular, free testosterone is calculated from available serum biochemical markers in the UKB data, prevalent diabetes is calculated across a range of UKB data fields and ICD10 codes are generalized to their top-level classifications. It is then possible to explore relationships between testosterone levels, diabetes presence, and associated morbidities. © 2023 ACM.
Federated learning based trajectory optimization for UAV enabled MEC
- Authors: Nehra, Anushka , Consul, Prakhar , Budhiraja, Ishan , Kaur, Gagandeep , Nasser, Nidal , Imran, Muhammad
- Date: 2023
- Type: Text , Conference paper
- Relation: 2023 IEEE International Conference on Communications, ICC 2023 Vol. 2023-May, p. 1640-1645
- Full Text: false
- Reviewed:
- Description: We present a moving mobile edge computing architecture in which unmanned aerial vehicles (UAV) serve as an equipment, providing computational power and allowing task offloading from mobile devices (MD). By improving user association, resource allocation, and UAV trajectory, we optimizing the energy consumption of all MDs. Towards that purpose, we provide a Trajectory optimization technique for making real-time choices while considering all the situation of the environment, followed by a DRL-based Trajectory control approach (RLCT). The RLCT approach may be adapted to any UAV takeoff point and can find the solution faster. The FL is introduced to address the Optimization problem in a Semi-distributed DRL technique to deal with UAV trajectory constraints. The proposed FRL approach enables devices to rapidly train the models locally while communicating with a local server to construct a network globally. The simulation results in the result section shows that the proposed technique RLCT and FRL in the paper outperforms the existing methods' while the FRL performs best among all. © 2023 IEEE.