| Title | : | Environment-interaction aided Exploration Approach for an Autonomous Ground Robot in Indoor Environments |
| Speaker | : | Nitin Kumar Dhiman (IITM) |
| Details | : | Mon, 17 Mar, 2025 2:00 PM @ SSB 233 |
| Abstract: | : | In many situations, robots may be deployed for tasks like search and rescue. These robots are equipped with one or more types of sensors. One of the key capabilities demanded of such autonomous systems is the ability to develop spatial awareness of the surroundings. A robot may need to select and identify goal locations to explore. These goal locations are called viewpoints. The class of algorithms that permits the robot to move efficiently in an unknown area is known as exploration algorithms. In this talk, I present an exploration approach that considers the possibility of moving an object during selection of viewpoints. The algorithm evaluates whether moving an object could lead to exploration of some obstructed area. The output of the algorithm is a sequence of viewpoints to visit, where each viewpoint is associated with an action sequence. For example, the action sequence may employ only perception with a coverage sensor, or manipulation of some object which would precede perception with a coverage sensor, etc. The volume of the explored area of the proposed algorithm has been compared against the TARE algorithm in simulation. A robot with a payload robotic arm is used in simulation experiments, which can identify and manipulate objects. Since the proposed solution can propose the movement of objects to create a passage for traversal into obstructed areas, the proposed algorithm results in more explored volume as compared to TARE. |
