Planning and Decision Making in Robots

Coursework - Fall 2021

Motion Planning for Precise Scanning of Additively Manufactured Parts

Term Project

• Implemented PRM algorithm to develop a roadmap of the robotic manipulator scanning environment.
• The motion plan construction was guided by A* search using a joint space heuristic function.
• The motion planner thus developed is an important module in a robotic inspection system to traverse an ordered set of scan path waypoints.
• Conducted a parametric study based on the metrics listed below to determine the best parameters to use for the roadmap.
  • No. of vertices
  • No. of connected components
  • Avg. no. of neighbors of each node
  • PRM build time
  • Avg. query time
  • Avg. path cost
  • Success rate
• Currently working on simultaneous motion planning of the DENSO robotic arm and turntable for scanning.

Catch me if you can!

Problem Statement 01

• Developed a planner that allowed a point robot to catch a moving object.

Planning for High-DOF planar arm

Problem Statement 02

• Implemented different sampling-based planners for the arm to move from its start joint configuration to a goal configuration avoid collision with the obstacles.
• The types of planners included Probabilistic Roadmaps, RRT, RRT-Connect, RRT*

Symbolic Planning

Problem Statement 03

• Programmed a generic symbolic planner. The planner takes in a set of symbols, initials conditions, goal conditions as part of the environment description.
• Given a set of valid actions, the planner is able to generate a logical plan to achieve the goal condition.