Habib Ullah Manzoor
@glasgow.ac.uk
Communication Sensing and Imaging Lab
University of Glasgow, UK
EDUCATION
PhD scholar
MS in Electrical Engineering
BSc Electrical Engineering
RESEARCH INTERESTS
Machine learning, communication system, Optical Engineering, Renewable Energy, Solar cells, Fiber optics, free space optics
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Attia Shabbir, Habib Ullah Manzoor, Ridha Alaa Ahmed, and Zahid Halim
IEEE
Federated learning (FL) has established itself as a communication-efficient, privacy-aware, and cost-effective technique for training machine learning models in energy forecasting. This approach enables simultaneous model training across multiple smart grids while keeping data decentralized at edge nodes. However, FL is not immune to backdoor adversarial attacks, such as data and model poisoning. In this paper, we scrutinize the impact of two data poisoning techniques: scaling and random noise effects. The attack was initiated on one client among ten. As the attack percentage increases, the Mean Absolute Percentage Error (MAPE) of the local model also rises. Our simulation results reveal that the scaling effect elevated MAPE from 0.193% to 32.72%, while random noise increased MAPE from 0.183% to 129.75% as the attacked percentage rose from 10% to 100%. It is concluded that data poisoning solely affects the local model and does not significantly impact the global model; hence, it can provide more resilience than centralized machine learning models.
Habib Ullah Manzoor, Sanaullah Manzoor, Muhammad Ali Jamshed, and Tareq Manzoor
Institution of Engineering and Technology (IET)
AbstractSolar cells are increasingly being utilised for both energy harvesting and reception in free‐space optical (FSO) communication networks. The authors focus on the implementation of a mid‐band p‐In0.01Ga0.99 N/p‐In0.5Ga0.5 N/n‐In0.5Ga0.5 N (PPN) solar cell, boasting an impressive 26.36% conversion efficiency (under 1.5AM conditions) as a receiver within an indoor FSO communication network. Employing a solar cell with dimensions of 1 mm in length and width, the FSO system underwent simulation using Optisystm software, while the solar cell's behaviour was simulated using SCAPS‐1D. The received power from the solar cell was then compared to that of four commercially available avalanche photodiode (APD) receivers. Exploring incident wavelengths spanning 400–700 nm within the visible spectrum, across transmission distances of 5, 10, 15, and 20 m, the study presented current‐voltage (IV) and power‐voltage curves. Notably, the InGaN solar cell exhibited superior electrical power output compared to all commercial APDs. In conclusion, the findings underscore that augmenting received power has the potential to enhance FSO network quality and support extended transmission distances.
Ahsan Raza Khan, Habib Ullah Manzoor, Fahad Ayaz, Muhammad Ali Imran, and Ahmed Zoha
MDPI AG
Human activity recognition (HAR) using wearable sensors enables continuous monitoring for healthcare applications. However, the conventional centralised training of deep learning models on sensor data poses challenges related to privacy, communication costs, and on-device efficiency. This paper proposes a federated learning framework integrating spiking neural networks (SNNs) with long short-term memory (LSTM) networks for energy-efficient and privacy-preserving HAR. The hybrid spiking-LSTM (S-LSTM) model synergistically combines the event-driven efficiency of SNNs and the sequence modelling capability of LSTMs. The model is trained using surrogate gradient learning and backpropagation through time, enabling fully supervised end-to-end learning. Extensive evaluations of two public datasets demonstrate that the proposed approach outperforms LSTM, CNN, and S-CNN models in accuracy and energy efficiency. For instance, the proposed S-LSTM achieved an accuracy of 97.36% and 89.69% for indoor and outdoor scenarios, respectively. Furthermore, the results also showed a significant improvement in energy efficiency of 32.30%, compared to simple LSTM. Additionally, we highlight the significance of personalisation in HAR, where fine-tuning with local data enhances model accuracy by up to 9% for individual users.
Satyam Bhatti, Habib Ullah Manzoor, Bruno Michel, Ruy Sebastian Bonilla, Richard Abrams, Ahmed Zoha, Sajjad Hussain, and Rami Ghannam
Wiley
Machine learning (ML) and artificial intelligence (AI) methods are emerging as promising technologies for enhancing the performance of low‐cost photovoltaic (PV) cells in miniaturized electronic devices. Indeed, ML is set to significantly contribute to the development of more efficient and cost‐effective solar cells. This systematic review offers an extensive analysis of recent ML techniques in designing novel solar cell materials and structures, highlighting their potential to transform the low‐cost solar cell research and development landscape. The review encompasses a variety of ML approaches, such as Gaussian process regression (GPR), Bayesian optimization (BO), and deep neural networks (DNNs), which have proven effective in boosting the efficiency, stability, and affordability of solar cells. The findings of this review indicate that GPR combined with BO is the most promising method for developing low‐cost solar cells. These techniques can significantly speed up the discovery of new PV materials and structures while enhancing the efficiency and stability of low‐cost solar cells. The review concludes with insights on the challenges, prospects, and future directions of ML in low‐cost solar cell research and development.
Habib Ullah Manzoor, Ahsan Raza Khan, David Flynn, Muhammad Mahtab Alam, Muhammad Akram, Muhammad Ali Imran, and Ahmed Zoha
MDPI AG
Increased demand for fast edge computation and privacy concerns have shifted researchers’ focus towards a type of distributed learning known as federated learning (FL). Recently, much research has been carried out on FL; however, a major challenge is the need to tackle the high diversity in different clients. Our research shows that using highly diverse data sets in FL can lead to low accuracy of some local models, which can be categorised as anomalous behaviour. In this paper, we present FedBranched, a clustering-based framework that uses probabilistic methods to create branches of clients and assigns their respective global models. Branching is performed using hidden Markov model clustering (HMM), and a round of branching depends on the diversity of the data. Clustering is performed on Euclidean distances of mean absolute percentage errors (MAPE) obtained from each client at the end of pre-defined communication rounds. The proposed framework was implemented on substation-level energy data with nine clients for short-term load forecasting using an artificial neural network (ANN). FedBranched took two clustering rounds and resulted in two different branches having individual global models. The results show a substantial increase in the average MAPE of all clients; the biggest improvement of 11.36% was observed in one client.
Alishah Khalid, Habib Ullah Manzoor, Hafiz Ashiq Hussian, and Moustafa H. Aly
Springer Science and Business Media LLC
Sana Hafeez, Habib Ullah Manzoor, Lina Mohjazi, Ahmed Zoha, Muhammad Ali Imran, and Yao Sun
IEEE
Exploiting unmanned aerial vehicles (UAVs) for delivery services is expected to reduce delivery time and human resource costs. However, the proximity of these UAVs to the ground can make them an ideal target for opportunistic criminals. Consequently, UAVs may be hacked, diverted from their destinations, or used for malicious purposes. Furthermore, as a decentralized (peer-to-peer) technology, the blockchain has immense potential to enable secure, decentralized, and cooperative communication among UAVs. With this goal in mind, we propose the Blockchain-Empowered, Immutable, and Reliable Delivery Service (BIRDS) framework to address data security challenges. BIRDS deploys communication hubs across a scalable network. Following the registration phase of BIRDS, UAV node selection is carried out based on a specific consensus proof-of-competence (PoC), where UAVs are evaluated solely on their credibility. The chosen finalist is awarded a certificate for the BIRDS global order fulfillment system. The simulation results demonstrate that BIRDS requires fewer UAVs compared to conventional solutions, resulting in reduced costs and emissions. The proposed BIRDS framework caters to the requirements of numerous users while necessitating less network traffic and consuming low energy.
Habib Ullah Manzoor, Ahsan Raza Khan, Tahir Sher, Wasim Ahmad, and Ahmed Zoha
IEEE
Federated Learning (FL) is susceptible to backdoor adversarial attacks during the training process, which poses a significant threat to the model's performance. Existing adversarial mitigation solutions mainly rely on the neural network (NN) model statistics and discard an entire client model if attacked. This approach is not feasible as it results in suboptimal performance. Hence, it is crucial to develop lightweight backdoor attack mitigation solutions that efficiently utilize clients' model statistics. To address this issue, we propose (Layer Based Anomaly Aware) LBAA-FedAVG, a modified version of the common aggregation mechanism FedAVG. Our proposed framework employs a clustering-based technique and addresses each NN layer individually. Depending on the type of adversarial attack, this method selectively eliminates one or multiple layers of the NN during the aggregation process. Furthermore, we focused on the model inversion attack and varied the percentage of compromised clients from 10% to 50%. Our experimental findings demonstrate that LBAA-FedAVG outperforms Federated Averaging (FedAVG) in reducing the negative effects of backdoor adversarial attacks. The complexity analysis suggests that the extra training time is the only additional resource limitation in LBAA-FedAVG, which is 19% greater than that of FedAVG. Additionally, we conducted experiments on short-term load forecasting using grid-level datasets to show the effectiveness of LBAA-FedAVG in lightweight backdoor attack mitigation in FL settings, offering a trade-off between time efficiency and enhanced defense.
A.K. Tan, H.U. Manzoor, N.A. Hamzah, M.A. Ahmad, S.S. Ng, and Z. Hassan
Elsevier BV
Habib Ullah Manzoor, Aik Kwan Tan, Sha Shiong Ng, and Zainuriah Hassan
Penerbit Universiti Kebangsaan Malaysia (UKM Press)
In this study, the indium gallium nitride (InxGa1-xN) p-n junction solar cells were optimized to achieve the highest conversion efficiency. The InxGa1-xN p-n junction solar cells with the whole indium mole fraction (0 £ x £ 1) were simulated using SCAPS-1D software. Optimization of the p- and n-InxGa1-xN layer's thickness and carrier density were also carried out. The thickness and carrier density of each layer was varied from 0.01 to 1.50 µm and 1015 to 1020 cm-3. The simulation results showed that the highest conversion efficiency of 23.11% was achieved with x = 0.6. The thickness (carrier density) of the p- and n-layers for this In0.6Ga0.4N p-n junction solar cell are 0.01 (1020) and 1.50 μm (1019 cm-3), respectively. Simulation results also showed that the conversion efficiency is more sensitive to the variations of layer's thickness and carrier density of the top p-InxGa1-xN layer than the bottom n-InxGa1-xN layer. Besides that, the results also demonstrated that thinner p-InxGa1-xN layer with higher carrier density offers better conversion efficiency.
M Hijaaj Tahir, A Arshad, and Habib Ullah Manzoor
IOP Publishing
Abstract Silicone rubber is one of the most used outdoor insulation materials in the last few decades due to its improved performance in contaminated and humid conditions. The improved performance of silicone rubber insulators is due to their hydrophobic nature, however, the organic nature of silicone molecules makes them vulnerable to ageing and degradation. This paper aims at investigating the loss and recovery of hydrophobicity of four different silicone rubber micro/nanocomposites exposed to corona discharge. The samples were exposed to corona discharge generated by pin-plate electrode configuration under AC stress. A series of tests were performed to observe the impact of different electrode-sample gaps and for various periods of corona exposure. The hydrophobicity of samples was measured pre and post corona exposures at various intervals up to 72 h. This time could confirm the hydrophobicity recovery process. Numerical simulations were also performed in COMSOL Multiphysics to investigate the electric fields along the sample surface at different electrode gaps. Experimental results showed that samples recovery time was proportional to the duration of exposure to corona discharge and inversely proportional to the electrode-sample gap. Among all, samples with 2.5% nano-silica as additive showed better hydrophobicity recovery. Simulation results showed that an increase in electrode gap resulted in decreased electric field intensity, hence supporting the experimental outcomes.
Habib Ullah Manzoor, Muhammed Shahzeb Khan, Ahsan Raza Khan, Fahad Ayaz, David Flynn, Muhammad Ali Imran, and Ahmed Zoha
IEEE
With the ever-increasing internet of things (IoT) and the rise of edge computing, federated learning (FL) is considered a promising solution for privacy and latency-aware applications. However, the data is highly distributed among several clients, making it challenging to monitor node anomalies caused by malfunctioning devices or any other unforeseen reasons. In this paper, we propose FedClamp, an anomaly detection algorithm based on the hidden Markov model (HMM) in the FL environment. FedClamp identifies the anomalous node and isolates them before aggregation to improve the performance of the global model. FedClamp was tested in a short-term energy forecasting problem using artificial neural networks when the FL environment had five clients. The algorithm uses mean absolute percentage error (MAPE) generated from local models and clusters them in normal and faulted nodes using HMM. The anomalous nodes identified using this algorithm are isolated before aggregation and achieve global model convergence with few communication rounds.
Habib Ullah Manzoor, Ahsan Raza Khan, Mohammad Al-Quraan, Lina Mohjazi, Ahmad Taha, Hasan Abbas, Sajjad Hussain, Muhammad Ali Imran, and Ahmed Zoha
IEEE
Smart building technologies transform buildings into agile, sustainable, and health-conscious ecosystems by leveraging IoT platforms. In this regard, we have developed a Persuasive Energy Conscious Network (PECN) at the University of Glasgow to understand the user-centric energy consumption patterns in an agile workspace. PECN consists of desk-level energy monitoring sensors that enable us to develop user-centric models that can be exploited to characterize the normal energy usage behavior of an office occupant. In this study, we make use of staked long short-term memory (LSTM) to forecast future energy demands. Moreover, we employed statistical techniques to automate the detection of anomalous power consumption patterns. Our experimental results indicate that post-anomaly resolution leads to 6.37% improvement in the forecasting accuracy.
Muhammad Zafar, Muhammad Saif Ullah, Tareq Manzoor, Muddassir Ali, Kashif Nazar, Shaukat Iqbal, Habib Ullah Manzoor, Rizwan Haider, and Woo Young Kim
Computers, Materials and Continua (Tech Science Press)
H.U. Manzoor, M.A. Md Zawawi, M.Z. Pakhuruddin, S.S. Ng, and Z. Hassan
Elsevier BV
Hafiz Shehbaz Ali, Fakhar ul Hassan, Siddique Latif, Habib Ullah Manzoor, and Junaid Qadir
IEEE
Speech emotion sensing in communication networks has a wide range of applications in real life. In these applications, voice data are transmitted from the user to the central server for storage, processing, and decision making. However, speech data contain vulnerable information that can be used maliciously without the user’s consent by an eavesdropping adversary. In this work, we present a privacy-enhanced emotion communication system for preserving the user personal information in emotion-sensing applications. We propose the use of an adversarial learning framework that can be deployed at the edge to unlearn the users’ private information in the speech representations. These privacy-enhanced representations can be transmitted to the central server for decision making. We evaluate the proposed model on multiple speech emotion datasets and show that the proposed model can hide users’ specific demographic information and improve the robustness of emotion identification without significantly impacting performance. To the best of our knowledge, this is the first work on a privacy-preserving framework for emotion sensing in the communication network.
Tareq Manzoor, K. Nazar, S. Iqbal, and Habib Ullah Manzoor
Elsevier BV
Muhammad Waqas Ashraf, Sheikh Muhammad Aaqib, Habib Ullah Manzoor, Muhammad Waqar Asharaf, Tareq Manzor, and Muzammal Hussain Sethi
IEEE
This paper evaluates the optimum tilt angle and available direct solar radiation on the inclined surface of the PV Cell in the province Punjab, Pakistan. This mathematical model provides the maximization of Solar Radiation impinging on the Solar collectors. Instead of Solar Collector placed at latitude angle with the help of parameters like Slope, orientation, azimuth surface angle and the hour angle of the PV Cell. The variation in the Solar Radiation on the PV Cell from Latitude to Tilt angle is represented graphically. Considered cities of Punjab are Lahore, Mangla, Sialkot, Gujranwala and Multan. This study reveals that as latitude angle varies from 33° to 28°, intensity increases approximately from 0.3KWh/m2 to 0.6KWh/m2 which will result in increased output power. Seasonal optimum angle has also been listed. Optimization of tilt angle has the highest effect on solar irradiance in winter season when day is shorter hence more light is required to fulfill our requirement.
Habib Ullah Manzoor, Sanaullah Manzoor, Tareq Manzoor, Talha Khan, and Ashiq Hussain
Wiley
Tareq Manzoor, Muhammad Zafar, Sanaullah Manzoor, Habib Ullah Manzoor, Muddassir Ali, and Woo Young Kim
IOP Publishing
Abstract In this paper, micro-mechanics of shape memory alloys (SMA) is investigated for design of smart structure and devices, based upon three dimensional constitutive model. In this work, a micro-mechanics based approach is presented. A modification in temperature field of the above model is proposed, where temperature is taken as function of time. Thermal parameters corresponding to transformation temperatures are introduced here. The model solution is presented for the of a set of initial and boundary conditions based upon any type of thermal parameters. Three particular cases for different thermal and structural loading-stresses are presented here. The solution given here expresses the response of smart dampers for various thermal and structural loading conditions. Moreover, the applications of this system were also investigated.
Habib Ullah Manzoor, Muhammad Zafar, Sana Ullah Manzoor, Talha Khan, Songzuo Liu, Tareq Manzoor, Saqib Saleem, Woo Young Kim, and Muddassir Ali
Elsevier BV
Habib Ullah Manzoor, Tareq Manzoor, and Masroor Hussain
Springer Science and Business Media LLC
This paper presents a canonical boundary value problem to understand the behavior of electromagnetic surface waves propagated on the planar interface of rugate filter and hyperbolic material. The electromagnetic surface waves are generated on the planar interface of two different materials. The hyperbolic materials constructed by making a dyadic negative in columnar thin films. Multiple electromagnetic surface waves were investigated by varying phase of rugate filter form 0 to 180° and wavelength of incident light from 400 to 700 nm. Material’s manufacturing fault is also introduced by adding 0.001 i in refractive index of hyperbolic medium.Propagation of multiple electromagnetic waves were observed on the planar interface of rugate filter and hyperbolic material. In order to verify the existence of surface electromagnetic waves at the interface of hyperbolic medium and rugate filter, electric and, magnetic fields are plotted for different material specifications.
Tareq Manzoor, Tariq Javed, Ghulam Mustafa, Habib Ullah Manzoor Ahmed, and Abdul Razzaq
Springer Science and Business Media LLC
Since the advent of material science, nanomaterials have been the most attractive and alluring research domain of nanotechnology with a variety of applications. Considering the significance of nanomaterials specifically in industrial progressions, the present work demonstrates facile synthesis approach of copper–zinc ferrite nanoparticles and their thermo-physical characterization and evaluation. Analytical grade chemicals were used to synthesize the respective nanoparticles employing the co-precipitation method, with the base solution of NaOH to maintain pH of the solution within range of 12–14. A series of nanoparticles were synthesized varying the amount of copper and zinc precursors, and their thermal and physical properties were evaluated using various analytical tools including XRD (X-ray diffraction), SEM (Scanning Electron Microscope), FTIR (Fourier transform infrared spectroscopy) and thermal constant analyzer.
Habib Ullah Manzoor, Tareq Manzoor, Ashiq Hussain, Moustafa H. Aly, and Sanaullah Manzoor
Springer Science and Business Media LLC
Next-generation optical communication networks require high input power and more number of channels at low spacing. This can be achieved using dense wavelength division multiplexing (DWDM). However, increasing the number of channels and decreasing channel spacing can enhance fiber nonlinearities, especially the four-wave mixing (FWM). In this paper, different modulation techniques and optical filters are considered and investigated to reduce the FWM effect in DWDM optical communication systems. System performance is evaluated through its quality factor (Q-factor), optical signal-to-noise ratio, optical received power and FWM efficiency. All used techniques have shown a reduction in FWM efficiency. The highest reduction in FWM efficiency is 25 dB and is reported while using modified Duobinary modulation with an increase of 2 in the Q-factor. A comparative study is carried out for the different techniques at 10–20 Gbps bit rate. All simulations are performed through Optisystem.