Mohammad Soltanshah

I am an instructor and Ph.D. candidate in Engineering Science at Simon Fraser University (SFU). My research focuses on the intersection of machine learning and visual servoing for robotic tasks, with an emphasis on Deep Reinforcement Learning and Image-space Path Planning. I am part of RAMP Laboratory at SFU. Previously, I completed my M.Sc. at Isfahan University of Technology, where my thesis focused on SLAM-based navigation strategies for mobile robots in dynamic environments.

Publications

Toward Autonomous Aerial Object Retrieval Utilizing RL-based Eye-to-hand/Eye-in-hand Feature Matching and Controllable-space Image Planning for Partitioned Visual Servoing

Mohammad Soltanshah, Abolfazl Eskandarpour, Kamal Gupta, and Mehran Mehrandezh

Submitted to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025
Robust Partitioned Visual Servoing for Aerial Manipulation Utilizing Controllable-space Planning and Adaptive Image Representation

Mohammad Soltanshah, Abolfazl Eskandarpour, Mehran Mehrandezh, and Kamal Gupta

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024
Hybrid Dynamic Modelling Using Feedforward and Temporal Convolutional Networks and Robust Control Scheme for Aerial Manipulators

Abolfazl Eskandarpour, Mohammad Soltanshah, Kamal Gupta, and Mehran Mehrandezh

IEEE International Conference on Automation Science and Engineering (CASE), 2024
A Constrained Robust Switching MPC Structure for tilt-rotor UAV Trajectory Tracking Problem

Abolfazl Eskandarpour, Mehran Mehrandezh, Kamal Gupta, Alejandro Ramirez-Serrano, and Mohammad Soltanshah

Nonlinear Dynamics, 111(18), 17247-17275, 2023

Some Personal Projects

Aerial Manipulation Configuration Estimation Using RGB Images and Keras API

This project focuses on estimating the configuration of an aerial manipulator using RGB images. The system leverages eye-to-hand and eye-in-hand cameras for perception and applies regression-based deep learning models for pose estimation, along with visualization and analysis tools.

[GitHub]
Moving-object Visual Servoing and Grasping using Open Manipulator

A ROS-based system for vision-based grasping of moving objects. The framework integrates pose estimation, visual servoing, and grasp execution, along with utilities for transformations and matrix operations.

[GitHub]
Rotary-tilting-wing Robot Modeling and Execution

Simscape/Simulink models for simulating the dynamics and control of a rotary-tilting-wing aerial system, enabling testing of control strategies and system behaviors in a configurable environment.

[GitHub]
A* Pathplanning Implementation and Visualization

A C++ visualization of the A* algorithm using SFML. The tool demonstrates pathfinding in a 2D grid with obstacles, start/end points, and animated exploration.

[GitHub]

Teaching

Instructor — CMPT 310: Introduction to Artificial Intelligence

Simon Fraser University, Spring 2026 — [course outline]

Artificial intelligence is an area with vast application scenarios. This course serves as a prerequisite for advanced AI courses and covers fundamental concepts and mathematical formulations in machine learning. Topics include supervised learning, linear regression, gradient descent, search algorithms, Markov decision processes, and Bayes nets.
Instructor — ENSC 204: Graphical Communication for Engineering

Simon Fraser University, Fall 2024 — [course outline]

An overview of engineering graphical communication, CAD, and workshop practices. The course covers visualization and development of 3D models, use of CAD tools, and GD&T concepts, along with simulation of mechanical and electrical components.