Loading...

Energy Optimization of an Arm with Seven Degrees of Freedom Using Imperialistic Competitive Algorithm (ICA) and Designing an Artificial Neural Network

Abedini, Amin | 2014

630 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 45638 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Ahmadian, M.T; Asghari, M
  7. Abstract:
  8. In recent years, great attention has been devoted to the design of artificial arms. The most crucial problem in such a design is the trajectory of movement. In this paper, a seven degree of freedom arm is modeled and simulated. Also, the optimization method named “Imperialistic Competitive Algorithm” has been modified and better performance of the new version is presented. Energy optimization is performed based on trajectory of the arm with angular velocity, angular acceleration and joint angles using modified imperialistic competitive algorithm (ICA). Considering ICA as a fast optimization algorithm, it would be reasonable to use this algorithm for robotic purposes for online answering. The simulated arm has a set of five links joint together where each has its own angular velocity, angular acceleration and torque. A set of six imperialists has been chosen as a case study via ICA. Simulation results reveals that the method is capable of finding the global minimum as well as local minimums with high performance. In the current non-convex problem, the global minimum has been found with 70 iterations.
    At the end, a neural network is designed and trained with 100 sets of learning data and tested with other 24 sets of test data. The neural network shows acceptable error and seems to be suitable for the case study.
  9. Keywords:
  10. Energy Optimization ; Imperialist Competitive Algorithm ; Global Minimum ; Energy Minimization ; Artificial Hand ; Seven Degree of Freedom Arm

 Digital Object List

 Bookmark

No TOC