Mohammad Salehizadeh

Mohammad Salehizadeh

Auto-generated profile, based on publicly available information

Mohammad Salehizadeh

Department of Mechanical and Industrial Engineering, University of Toronto, Toronto, ON, Canada

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 April 2019 in Science Robotics
Verified
Journal Article
Closed access
PDF file
TXTianqi Xu
Tianqi Xu
JZJiachen Zhang
Jiachen Zhang
MSMohammad Salehizadeh
Mohammad Salehizadeh
OOOnaizah Onaizah
Onaizah Onaizah
EDEric Diller
Eric Diller
New types of flexible microrobots are enabled by patterning discrete 3D magnetization.

5 August 2020 in The International Journal of Robotics Research
Verified
Journal Article
Closed access
PDF file
MSMohammad Salehizadeh
Mohammad Salehizadeh
EDEric Diller
Eric Diller
This article presents a method to independently control the position of multiple microscale magnetic robots in three dimensions, operating in close proximity to each other. Having multiple magnetic microrobots work together in close proximity is difficult due to magnetic interactions between the robots, and here we aim to control those interactions for the creation of desired multi-agent formations in three dimensions. Based on the fact that all magnetic agents orient to the global input magnetic field, the local attraction–repulsion forces between nearby agents can be regulated. For the first time, 3D manipulation of two microgripping magnetic microrobots is demonstrated. We also mathematically and experimentally prove that the center-of-mass external magnetic pulling of the multi-agent system is possible in three dimensions with an underactuated magnetic field generator. Here we utilize the controlled interaction magnetic forces between two spherical agents to steer them along two prescribed paths. We apply our method to independently control the motion of a pair of magnetic microgrippers as functional microrobot candidates each equipped with a five-degree-of-freedom motion mechanism and a grasp–release mechanism for targeted cargo delivery. A proportional controller and an optimization-based controller are introduced and compared, with potential to control more than two magnetic agents in three dimensions. Average tracking errors of less than 141 and 165 micrometers are accomplished for the regulation of agents’ positions using optimization-based and proportional controllers, respectively, for spherical agents with approximate nominal radius of 500 micrometers operating within several body-lengths of each other.

25 February 2021 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
MSMohammad Salehizadeh
Mohammad Salehizadeh
EDEric D. Diller
Eric D. Diller
We propose a robotics method to independently control the position of multiple magnetic microrobots (agents) through their local magnetic interactions in close proximity. Utilizing a rapidly-exploring random tree (RRT) planner, we generate a trajectory which is then followed by an optimization-based controller (OBC), while avoiding obstacles. The method generates a path that obeys the governing unique nonholonomic actuation constraints in the agents' motions, and has potential to scale to more agents. This is the first work of its kind to introduce a method for autonomous navigation of multiple magnetic microrobots in close proximity with obstacle avoidance. The ability to independently control the motion of multiple magnetic microrobots in this manner makes the method useful for potential applications such as targeted drug delivery and medical diagnostics. We prove that the two-microrobot system studied here is small-time locally controllable (STLC). We conduct two practical demonstrations to verify the ability of the RRT-based navigation applied to a pair of magnetic microrobots moving in a microchannel: 1) multi-agent micro-factory, and 2) team-targeted cargo delivery. An average tracking error of around one-fifth of the agents' body-length is achieved for the control of agents' positions.

4 January 2024 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
BBBrevin Banks
Brevin Banks
MSMohammad Salehizadeh
Mohammad Salehizadeh
AMAdnan Munawar
Adnan Munawar
RTRussell H. Taylor
Russell H. Taylor
UTUgur Tumerdem
Ugur Tumerdem
The Galen Surgical Robot is an admittance controlled collaborative robot that can enhance the performance of surgeons in microsurgery by sharing the load of instruments, preventing hand tremor and improving precision. In mastoidectomy, contacting the temporal bone of the skull and drilling with the robot can be challenging due to potential instability of the coupled surgeon-robot-patient system. In this letter, we investigate how tuning mass and damping gain properties of an admittance controller affect the stability and transparency of this interaction. We generate and make use of stability-transparency maps to aid in the optimal design of the control system. To address the stability/transparency trade-off, we propose gain switching between free motion and contact using an external force sensor or a force sensing tool to detect contact. We simulate critical phases of mastoidectomy surgery on temporal bone phantoms, including first contact with bone and drilling. In agreement with theoretical results, experiments with the proposed controller show significant improvements in transparency and contact stability.

13 November 2020 in Optics Letters
Verified
Journal Article
Closed access
EAEhsan Alimohammadian
Ehsan Alimohammadian
SLSifan Liu
Sifan Liu
MSMohammad Salehizadeh
Mohammad Salehizadeh
Jianzhao LiJianzhao Li
Jianzhao Li
PHPeter Herman
Peter Herman
Various beam shaping approaches were examined to counter the negative influence of surface aberration arising when inscribing optical waveguides deeply inside of glass with a femtosecond laser. Aberration correction was found unable to completely recover the low-loss waveguide properties, prompting a comprehensive examination of waveguides formed with focused Gaussian–Bessel beams. Diverging conical phase fronts are presented as a hybrid means of partial aberration correction to improve insertion loss and a new, to the best of our knowledge, means of asymmetric beam shaping. In this way, low-loss waveguides are presented over shallow to deep writing depth (2.8 mm) where morphological and modal properties could be further tuned with conical phase front.

1 January 2025 in Healthcare Technology Letters
Verified
Journal Article
Open access
RSRoger D. Soberanis‐Mukul
Roger D. Soberanis‐Mukul
CCChin Hang Ryan Chan
Chin Hang Ryan Chan
RCRyan Chou
Ryan Chou
JMJan Emily Mangulabnan
Jan Emily Mangulabnan
Lalithkumar SeenivasanLalithkumar Seenivasan
Lalithkumar Seenivasan
MUMathias Unberath
Mathias Unberath
Surgical navigation is critical in sinus surgery to enhance the surgeon's spatial awareness and improve precision, particularly around occluded critical structures. While external tracker‐based navigation systems exist, vision‐based solutions are preferred for being less intrusive and for enabling endoscopic image analysis to assist surgeons. However, monocular endoscopy navigation faces challenges associated with monocular reconstruction and camera pose estimation. This paper presents a proof of concept for monocular vision‐based sinus navigation that utilizes only preoperative CT data and the endoscope video stream to navigate the sinus anatomy. We developed a vision‐based navigation system that incorporates a SLAM algorithm to estimate the camera pose and reconstruct the 3D surface of the anatomy. Given an initial semi‐automated registration, the algorithm maps the SLAM‐based trajectories to the CT space while employing the reconstructed point cloud to solve for the scale interactively. The system displays the updates in the CT triplane visualization as SLAM reconstructs the scene and recovers pose information. We tested our system by performing an off‐site navigation in ten recorded endoscopic video streaming generated from sequences obtained from eight cadaveric subjects, comparing the vision‐based navigation to reference optical tracker pose data and obtaining translation and rotation errors of 3.2 mm and 4.9 degrees, respectively. Additionally, we performed three on‐site tests of our system on two different cadaver experiments. Our work evaluates a fully integrated system that closes the loop between image‐based reconstruction and CT visualization, and discusses the challenges to address to achieve clinical level surgical navigation.