Muhammad Haris Kaka Khel

Bio:
Muhammad Haris Kaka Khel was born on 13th January 1997 in Charsadda, Khyber Pakhtunkhwa, Pakistan. He earned his B.S. degree in Electrical Engineering, with a major in Communication, from the University of Engineering and Technology (UET) Peshawar, Pakistan, in 2019. Haris pursued his research master’s degree at the University of Kuala Lumpur, Malaysia, and spent one semester at Politecnico di Torino, Italy, through the Erasmus exchange program. His master’s research focused on Real-Time Crowd Monitoring: Estimating Count, Speed, and Direction of Pedestrians, during which he published several papers in IEEE and other leading journals and conferences.

Currently, Haris is pursuing his Ph.D. at Atlantic Technological University (ATU), with a research focus on Pedestrian Adaptive Trajectory Hypothesis System (PATHS), which aims to predict pedestrian trajectories using behavioral analysis, social interactions, and scene dynamics. He is expected to complete his Ph.D. by 2026. His main supervisor is Dr. Kevin Meehan (ATU Donegal), and his co-supervisors are Dr. Paul Greaney (ATU Donegal), Dr. Marion McAfee (ATU Sligo), and Dr. Sandra Moffet (Ulster University, Derry).

Research Topic:
Pedestrian Adaptive Trajectory Hypothesis System (PATHS)

Research Summary:

In this pedestrian trajectory prediction research, we aim to forecast the future paths of individuals by analyzing multiple factors such as human behavior, past movements, social interactions, and interactions with the surrounding environment. Unlike traditional models that predict a single path, our approach focuses on predicting multiple possible paths, accounting for the inherent uncertainties and variations in human behavior. This makes our prediction framework more robust and adaptable, especially in dynamic environments like busy urban streets or public spaces.

This research is essential in several real-world applications. In autonomous driving, accurate pedestrian path prediction helps vehicles anticipate movements, reducing the risk of accidents and enhancing road safety. Urban planners can benefit by designing pedestrian-friendly spaces that improve walkability and minimize congestion. Additionally, crowd management during large events can be significantly improved by predicting multiple possible paths, which helps prevent bottlenecks and ensures smoother pedestrian flow. Smart surveillance systems can also leverage these predictions to detect unusual or dangerous behavior in crowded areas, enabling preventive interventions. The pedestrian trajectory prediction plays an important role in assisting the visually impaired by improving navigation in public spaces, making cities safer and more accessible. In public transportation, it can optimize schedules by predicting pedestrian flow, reducing wait times, and enhancing overall efficiency. The research holds great promise for increasing safety, convenience, and efficiency across various sectors, contributing to the development of smarter, safer, and more responsive urban environments.

Khel, Muhammad Haris Kaka, Paul Greaney, Marion McAfee, Sandra Moffett, and Kevin Meehan. "GSTGM: Graph, spatial–temporal attention and generative based model for pedestrian multi-path prediction." Image and Vision Computing (2024): 105245.

Khel, Muhammad Haris Kaka, Paul Greaney, Marion McAfee, Sandra Moffett, and Kevin Meehan. "Pedestrian Trajectory Prediction using BiLSTM with Spatial-Temporal Attention and Sparse Motion Fields." In 2023 34th Irish Signals and Systems Conference (ISSC), pp. 1-7. IEEE, 2023.

Albattah, Waleed, Muhammad Haris Kaka Khel, Shabana Habib, Muhammad Islam, Sheroz Khan, and Kushsairy Abdul Kadir. "Hajj crowd management using CNN-based approach." (2020).

Khel, Muhammad Haris Kaka, Kushsairy Abdul Kadir, Sheroz Khan, Mnmm Noor, Haidawati Nasir, Nawaf Waqas, and Akbar Khan. "Realtime Crowd Monitoring—Estimating Count, Speed and Direction of People Using Hybridized YOLOv4." IEEE Access 11 (2023): 56368-56379.

Khel, Muhammad Haris Kaka, Kushsairy Kadir, Waleed Albattah, Sheroz Khan, M. N. M. M. Noor, Haidawati Nasir, Shabana Habib, Muhammad Islam, and Akbar Khan. "Real-time monitoring of COVID-19 SOP in public gathering using deep learning technique." Emerging Science Journal 5, no. Special issue (2021): 182-196.

Khel, Muhammad Haris Kaka, Kushsairy Kadir, Sheroz Khan, Waleed Albattah, Haidawati Nasir, M. N. M. M. Noor, Akbar Khan, and Nawaf Waqas. "Hybridized YOLOv4 for detecting and counting people in congested crowds." In 2022 IEEE International Instrumentation and Measurement Technology Conference (I2MTC), pp. 1-6. IEEE, 2022.

Waqas, Nawaf, Sairul Izwan Safie, Kushsairy Abdul Kadir, Sheroz Khan, and Muhammad Haris Kaka Khel. "DEEPFAKE image synthesis for data augmentation." IEEE Access 10 (2022): 80847-80857.

Khan, Akbar, Kushsairy Kadir, Haidawati Nasir, Jawad Ali Shah, Waleed Albattah, Sheroz Khan, and Muhammad Haris Kakakhel. "Crowd counting and localization beyond density map." IEEE Access 10 (2022): 133142-133151.

Khan, Akbar, Kushsairy Abdul Kadir, Jawad Ali Shah, Waleed Albattah, Muhammad Saeed, Haidawati Nasir, Megat Norulazmi Megat Mohamed Noor, and Muhammad Haris Kaka Khel. "A Deep Learning Approach for Crowd Counting in Highly Congested Scene." Computers, Materials & Continua 73, no. 3 (2022).

Intelligent Real-time Crowd Monitoring System Using Unmanned Aerial Vehicle (UAV) Video and Global Positioning Systems (GPS) Data. Sponsered by Deputyship for Research and Innovation, Ministry of Education in Saudi Arabia

• Lecturer(part-time) at Atlantic Technological University Donegal, Ireland
• Graduate Research Assistant at University Kuala Lumpur -British Malaysian Institute (2022)