Amir Gholami

Dr. Amir Gholami

orcid

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
12 March 2019 in Proceedings of the Institution of Mechanical Engineers, Part K: Journal of Multi-Body Dynamics
Verified
Journal Article
Closed access
PDF file
AGAmir Gholami
Amir Gholami
MMMajid Majidi
Majid Majidi
In this paper, a neuromuscular driver model for sensing torque feedback or haptic interaction between the vehicle equipped with steer-by-wire (SBW) system and the driver has been developed. The proposed driver model consists of a preview model and a neuromuscular model. The preview driver model calculates the desired angle of the steering-wheel to follow the path, and the neuromuscular driver model, with the ability of perceiving real-time torque feedback, determines the real angle of the steering-wheel angle according to muscular system transfer functions to follow the desired steering-wheel angle. In order to calculate torques on the steering-wheel, the lateral tyre-road forces are estimated by Kalman filter designed using a linear 2-DOF vehicle model. So, the design of the neuromuscular driver model combined with torque feedback estimation is the main contribution of this paper. The simulation results from TruckSim and Simulink software indicate that the novel designed driver model with torque feedback estimation has an important role in the controlling and steering vehicle to follow the desired paths.

13 January 2025 in Sensors
Verified
Journal Article
Open access
PDF file
AGAmir Gholami
Amir Gholami
ARAlex Ramirez-Serrano
Alex Ramirez-Serrano
This paper presents a comprehensive approach to evaluating the ability of multi-legged robots to traverse confined and geometrically complex unstructured environments. The proposed approach utilizes advanced point cloud processing techniques integrating voxel-filtered cloud, boundary and mesh generation, and dynamic traversability analysis to enhance the robot’s terrain perception and navigation. The proposed framework was validated through rigorous simulation and experimental testing with humanoid robots, showcasing the potential of the proposed approach for use in applications/environments characterized by complex environmental features (navigation inside collapsed buildings). The results demonstrate that the proposed framework provides the robot with an enhanced capability to perceive and interpret its environment and adapt to dynamic environment changes. This paper contributes to the advancement of robotic navigation and path-planning systems by providing a scalable and efficient framework for environment analysis. The integration of various point cloud processing techniques into a single architecture not only improves computational efficiency but also enhances the robot’s interaction with its environment, making it more capable of operating in complex, hazardous, unstructured settings.

1 December 2019
Verified
Book Chapter
Closed access
PDF file
AGAmir Gholami
Amir Gholami
MMMilad Moradi
Milad Moradi
MMMajid Majidi
Majid Majidi

1 December 2019
Verified
Book Chapter
Closed access
PDF file
HMHamed Mahmudi
Hamed Mahmudi
AGAmir Gholami
Amir Gholami
MDMohammad Hossein Delavaran
Mohammad Hossein Delavaran
SKSoheil Khatibi
Soheil Khatibi
SBSaeid Bazargan
Saeid Bazargan
MTMeisam Teimouri
Meisam Teimouri

29 February 2024
Verified
Conference Paper
Closed access
AGAmir Gholami
Amir Gholami
FRFatemeh Rashnozadeh
Fatemeh Rashnozadeh
ARArash Rahmani
Arash Rahmani
ANAhmadreza Nazari
Ahmadreza Nazari
PBPegah Behvarmanesh
Pegah Behvarmanesh
ARAlejandro Ramirez-Serrano
Alejandro Ramirez-Serrano
Deep learning methods like semantic segmentation have gained popularity in computer vision, but challenges remain, particularly in the lack of relevant datasets, such as for humanoid soccer robots. Manually annotating these datasets for segmentation is time-consuming and error-prone. To overcome this, we utilized realistic simulation to quickly generate large datasets and corresponding masks. This paper focuses on applying the u-net architecture on low-end hardware, using a synthetic dataset for training. To evaluate the model, a real annotated dataset was created. Although our Sim-to-Real approach produced a dataset close to reality, the results were unsatisfactory. To address this, transfer learning was employed to fine-tune the network and achieve better accuracy in a real environment.