Ph.D. Mechanical Engineering, University of Utah (2021)
M.S. Mechanical Engineering, AmirKabir University of Technology
B.S. Mechanical Engineering, K.N. Toosi University of Technology
Lab: MEB 2215
1. Improvement of Human Safety and In Human-and-Robot Interaction
Due to recent developments in collaborative robots, industries are more intended to use them in collaboration with human workers. In this project, we tried to improve human safety while interacting with collaborative robots by implementing safety measures into planning and control algorithms of these robots. The goals of this research project are:
- Study human motion in a shared autonomy workplace interacting with a collaborative robot.
- To have better understanding on human body motion in collaboration tasks, a motion capture system connected to ROS is used which provides online human postures and also records those motions for later studies
- Develop a reinforcement learning algorithm to find the cost function of human motion and use it to predict the human motion
- Develop an algorithm for real-time optimal motion planning and failure-tolerant control of robots which maintains safety measures and improves productivity at the same time.
- The novel algorithm is based on convex optimization and MPC and ensures the collision avoidance and safety measures between human and robot and simultaneously improves productivity by completing the task without stopping or lowering the speed and follow the global optimal path. In addition, the algorithm reduces harmful velocity and force jumps caused by the actuator failure in the robot.
- Use the prediction of human motion to improve the performance of motion planner algorithm
2. Estimation and Ergonomic Analysis of Human Posture in Tele-Manipulation Tasks
In this project, we focused on improvement of ergonomics in tele-manipulation task in which human has physical interaction with tele-operation robots and it is where evidences show high rate of musculoskeletal injuries. The goals of this projects include:
- Posture estimation of human doing a tele-manipulation task
- Posture of human while having physical interaction with a tele-operation robot is estimated by only using the trajectory of the robot without any vision system. The problem modeled as a partially-observed Hidden Markovian Model (HMM) and solved via a filtering approach.
- Online ergonomic analysis of human posture in tele-manipulation
- Develop an online ergonomics analysis algorithm that is based on current ergonomics study methods (such as RULA–Rapid Upper-Limb Assessment) .
- Use OpenSim software with improved cost functions to score a posture
- Smart re-mastering of the tele-manipulation robot to maintain human posture in an ergonomic manifold while doing different tele-manipulation tasks
- Model the problem as a POMDP problem
Research Interests: Human-Centered Robotics, Artificial Intelligence, Probabilistic Robotics, Robot Learning and Perception, Telemanipulation, HRI/HRC, Safety, and Ergonomics
Personal Interests: Fly fishing, Hiking, and Mountain Biking