Analysis and Development of Computational Intelligence based Navigational Controllers for Multiple Mobile Robots

Mohanty, Prases Kumar (2015) Analysis and Development of Computational Intelligence based Navigational Controllers for Multiple Mobile Robots. PhD thesis.



Navigational path planning problems of the mobile robots have received considerable attention over the past few decades. The navigation problem of mobile robots are consisting of following three aspects i.e. locomotion, path planning and map building. Based on these three aspects path planning algorithm for a mobile robot is formulated, which is capable of finding an optimal collision free path from the start point to the target point in a given environment. The main objective of the dissertation is to investigate the advanced methodologies for both single and multiple mobile robots navigation in highly cluttered environments using computational intelligence approach. Firstly, three different standalone computational intelligence approaches based on the Adaptive Neuro-Fuzzy Inference System (ANFIS), Cuckoo Search (CS) algorithm and Invasive Weed Optimization (IWO) are presented to address the problem of path planning in unknown environments. Next two different hybrid approaches are developed using CS-ANFIS and IWO-ANFIS to solve the mobile robot navigation problems. The performance of each intelligent navigational controller is demonstrated through simulation results using MATLAB. Experimental results are conducted in the laboratory, using real mobile robots to validate the versatility and effectiveness of the proposed navigation techniques. Comparison studies show, that there are good agreement between them. During the analysis of results, it is noticed that CS-ANFIS and IWO-ANFIS hybrid navigational controllers perform better compared to other discussed navigational controllers. The results obtained from the proposed navigation techniques are validated by comparison with the results from other intelligent techniques such as Fuzzy logic, Neural Network, Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO) and other hybrid algorithms. By investigating the results, finally it is concluded that the proposed navigational methodologies are efficient and robust in the sense, that they can be effectively implemented to solve the path optimization problems of mobile robot in any complex environment.

Item Type:Thesis (PhD)
Uncontrolled Keywords:Mobile robot, Navigation, ANFIS, Cuckoo Search, Invasive weed Optimization
Subjects:Engineering and Technology > Mechanical Engineering > Robotics
Divisions: Engineering and Technology > Department of Mechanical Engineering
ID Code:7253
Deposited By:Mr. Sanat Kumar Behera
Deposited On:30 Mar 2016 16:06
Last Modified:30 Mar 2016 16:09
Supervisor(s):Parhi, D R

Repository Staff Only: item control page