Momin Uppal’s research while affiliated with Lahore University of Management Sciences and other places

Several Pakistani cities are among the world’s most polluted. In the previous three years, air pollution in Lahore has been considerably over World Health Organization guideline levels, endangering the lives of the city’s more than 11 million citizens. In this paper, we investigate the city’s capability to combat air pollution by analyzing three essential aspects: (1) Data, (2) Capacity, and (3) Public awareness. Several studies have reported the need for expansion of the current air quality monitoring network. In this work, we also provide a context-aware location recommendation algorithm for installing the new air quality stations in Lahore. Data from four publicly available reference-grade continuous air quality monitoring stations and nine low-cost air quality measuring equipment are also analyzed. Our findings show that in order to measure and mitigate the effects of air pollution in Lahore, there is an urgent need for capacity improvement (installation of reference-grade and low-cost air quality sensors) and public availability of reliable air quality data. We further assessed public awareness by conducting a survey. The questionnaire results showed huge gaps in public awareness about the harms of the air quality conditions. Lastly, we provided a few recommendations for designing data-driven policies for dealing with the current apocalyptic air quality situation in Lahore.

Estimating the spatio-temporal profile of a building’s construction using high-resolution satellite images is a critical problem since it can be utilized for a variety of data-driven urban initiatives. One strategy to achieve this is to extract building footprints and track them in multi-temporal data as observed in SpaceNet’s Challenges. Although several unique solutions have been presented for this problem, this task can become extremely difficult for partially obscured buildings with densely overlapping boundaries, such as those found in underdeveloped countries like Pakistan. Consequently, in this paper we propose a framework to address this problem by merging built-up area segmentation with digital maps. In the first step, satellite image is passed to a deep learning model that predicts segmentation masks over the built-up area following which building construction profiles are generated by overlaying digital maps over these predicted masks. We compare the results with ground truth profiles and our results show that the proposed method extracts building counts and construction profiles with an accuracy of 95%.

The e-commerce industry has seen significant growth over the past few years. One significant issue that has sprung up as a result of this growth is unstructured addresses during last mile delivery. These ambiguous addresses are an established issue, particularly in developing countries like Pakistan. They are difficult to read and locate by last mile delivery riders thereby increasing delivery times and cost, negatively impacting the business of the company. Increased delivery times are also detrimental to the environment. In this paper, we aim to quantify the effects of unstructured addresses on last mile logistics. Many attempts have been made to standardise addresses to tackle this problem. Deep learning based approaches using recurrent neural networks (RNN) as well as probabilistic approaches using hidden Markov models (HMM) have been used. However, the main downside to these approaches are the underlying variation in address schemes in housing societies. We present an end to end rule based pipeline using Levenshtein distance (LD) and regular expressions (RegEx) rules which takes those unstructured addresses and outputs their structured forms along with their Geo-coordinates. The pipeline also returns the optimized route to minimize the last mile distance traveled.

p>In this paper, we consider the problem of removing clouds and recovering ground cover information from remote sensing images by proposing novel framework based on a deep unfolded and prior-aided robust principal component analysis (DUPA-RPCA) network. Clouds, together with their shadows, usually occlude ground-cover features in optical remote sensing images. This hinders the utilization of these images for a range of applications such as earth observation, land-cover classification and urban planning. We model these cloud-contaminated images as a sum of low rank and sparse elements and then unfold an iterative RPCA algorithm that has been designed for reweighted l1-minimization. As a result, the activation function in DUPA-RPCA adapts for every input at each layer of the network. Our experimental results on both Landsat and Sentinel images indicate that our method gives better accuracy and efficiency when compared with existing state of the art methods. </p

p>In this paper, we consider the problem of removing clouds and recovering ground cover information from remote sensing images by proposing novel framework based on a deep unfolded and prior-aided robust principal component analysis (DUPA-RPCA) network. Clouds, together with their shadows, usually occlude ground-cover features in optical remote sensing images. This hinders the utilization of these images for a range of applications such as earth observation, land-cover classification and urban planning. We model these cloud-contaminated images as a sum of low rank and sparse elements and then unfold an iterative RPCA algorithm that has been designed for reweighted l1-minimization. As a result, the activation function in DUPA-RPCA adapts for every input at each layer of the network. Our experimental results on both Landsat and Sentinel images indicate that our method gives better accuracy and efficiency when compared with existing state of the art methods. </p

In this paper, we consider bidirectional relaying between two diffusion-based molecular transceivers (bio-nodes). As opposed to existing literature, we incorporate the effect of direct diffusion links between the nodes and leverage it to improve performance. Assuming network coding type operation at the relay, we devise a detection strategy, based on the maximum-likelihood principle, that combines the signal received from the relay and that received from the direct link. At the same time, since a diffusion-based molecular communication channel is characterized by high inter-symbol interference (ISI), we utilize a decision feedback mechanism to mitigate its effect. Simulation results indicate that the proposed setup incorporating the direct link can achieve notable improvement in error performance over conventional detection schemes that do not exploit the direct link and/or do not attempt to mitigate the effect of ISI.

Urban air quality is increasingly becoming a cause for concern for the health of the human population. The poor air quality is already wreaking havoc in major cities of the world, where serious health issues and reduction of average human life by a factor of years are reported. The air quality in developing countries can become worse as they undergo development. The urban air quality varies non-linearly depending upon the various factors such as land use, industrialization, waste disposal, traffic volume, etc. To address this problem, it is necessary to look at the plethora of available literature from multiple perspectives such as types and sources of pollutants, meteorology, urban mobility, urban planning and development, health care, economics, etc. In this paper, we provide a comprehensive survey of the state-of-the-art in urban air quality. We first review the fundamental background on air quality and present the emerging landscape of urban air quality. We then explore the available literature from multiple urban air quality measurement projects and provides the insights uncovered in them. We then take a look at the sources that are significantly contributing to polluting the air quality. Finally, we highlight open issues and research challenges in dealing with urban air pollution.

Considering a common case where measurements are obtained from independent sensors, we present a novel outlier-robust filter for nonlinear dynamical systems in this work. The proposed method is devised by modifying the measurement model and subsequently using the theory of Variational Bayes and general Gaussian filtering. We treat the measurement outliers independently for independent observations leading to selective rejection of the corrupted data during inference. By carrying out simulations for variable number of sensors we verify that an implementation of the proposed filter is computationally more efficient as compared to the proposed modifications of similar baseline methods still yielding similar estimation quality. In addition, experimentation results for various real-time indoor localization scenarios using Ultra-wide Band (UWB) sensors demonstrate the practical utility of the proposed method.

Clouds, together with their shadows, usually occlude ground-cover features in optical remote sensing images. This hinders the utilization of these images for a range of applications such as earth observation, land-cover classification and urban planning. In this work, we propose a deep unfolded and prior-aided robust principal component analysis (DUPA-RPCA) network for removing clouds and recovering ground-cover information in multi-temporal satellite images. We model these cloud-contaminated images as a sum of low rank and sparse elements and then unfold an iterative RPCA algorithm that has been designed for reweighted $\ell _{1}$ minimization. As a result, the activation function in DUPA-RPCA adapts for every input at each layer of the network. Our experimental results on both Landsat and Sentinel images indicate that our method gives better accuracy and efficiency when compared with existing state of the art methods.

State estimation of dynamical systems is crucial for providing new decision-making and system automation information in different applications. However, the assumptions on the standard computational models for sensor measurements can be violated in practice due to different types of data abnormalities such as outliers and biases. In this work, we focus on the occurrence of measurement biases and propose a robust filter for their detection and mitigation during state estimation of nonlinear dynamical systems. We model the presence of bias in each dimension within the generative structure of the state-space models. Subsequently, employing the theory of Variational Bayes and general Gaussian filtering, we devise a recursive filter which we call the Bias Detecting and Mitigating (BDM) filter. As the error detection mechanism is embedded within the filter structure its dependence on any external detector is obviated. Simulations verify the performance gains of the proposed BDM filter compared to similar Kalman filtering-based approaches in terms of robustness to temporary and persistent bias presence.