Sensor-based Collision-Free Motion Planning and Control for Autonomous Cart Pullers at Advanced Intelligent Systems (AIS)

The objective is to develop and implement safe navigation algorithms and frameworks for autonomous cart pullers that are supposed to move planted pots on carts within an unknown and dynamically-changing environment in plant nurseries. The cart puller operation needs to be done safely in an environment that is shared by other vehicles, nursery utilities such as planted pots, and human operators. In this regard, the developed software- and hardware-based framework should be capable to overcome following challenges: (1) perform perception using on board LiDAR and camera vision, (2) construct a safe global path between current state of the cart puller and its desired destination, (3) develop dynamic/local motion planning algorithms to avoid collision with obstacles, (4) generate control commands for trajectory tracking, and (5) integration of all modules. In this project, it is intended to use Robot Operating System (ROS) as the basic software framework for developed motion planning and control algorithms. Also for real-time purposes, Robot Operating System 2 (ROS2) is planned to be used alongside ROS1.

Mohammad Soltanshah
Faculty Supervisor: 
Mehran Mehrandezh
Partner University: