The version presented here may differ from the published version or from the version of the record. Please see the repository URL above for details on accessing the published version and note that access may require a subscription.

No abstract

One of the long standing challenging aspect in mobile robotics is the ability to navigate autonomously, avoiding obstacles especially in crowded and unknown environment. The path followed by a mobile robot and its behavior plays an important role in the quality of localization and mapping as well. To combat this problem, we introduced a real time and robust recursive line extraction algorithm for object based navigation. For navigation task robot can't neglect any object and laser scanner is not reliable in office environment which crystalline objects are common, so the sensor fusion is quite essential for navigation a mobile robot. In this paper we have extended our navigation approach using fuzzy controller that will take path based on extracted lines and fusing data from sonar modules. Some benefits of using fuzzy logic controller include, higher speed and smoother control command because it assigns different speeds to wheels and generate curved path instead of zigzagged path by r and commands in polar coordinate system. Furthermore, it has good real-time capability and it is implemented on NAJI V mobile search and rescue robot platform and the results show better performance in localization and mapping as well.

Deep Neural Network methods have been used to a variety of challenges in automatic 3D recognition. Although discovered techniques provide many advantages in comparison with conventional methods, they still suffer from different drawbacks, e.g., a large number of pre-processing stages and timeconsuming training. In this paper, an innovative approach has been suggested for recognizing 3D models. It contains encoding 3D point clouds, surface normal, and surface curvature, merge them to provide more effective input data, and train it via a deep convolutional neural network on Shapenetcore dataset. We also proposed a similar method for 3D segmentation using Octree coding method. Finally, comparing the accuracy with some of the state-of-the-art demonstrates the effectiveness of our proposed method. INDEX TERMS Object recognition, recurrent neural networks, multi-layer neural network, octrees.

Robot navigation and obstacle avoidance are from the most important problems in mobile robots, especially in unknown environments. The autonomous navigation of mobile robots has attracted a number of researchers over the years. A wide variety of approaches and algorithms have been proposed to tackle this complex problem. There are many successful localization methods that can determine the robot's position relative to a map using sonar, laser and camera data; however, most localization methods fail under common environmental conditions. In this paper we present a novel method for navigation of autonomous mobile robots in several indoor environments that optimize our old navigation approach, using fuzzy Kalman filter based on extracted lines and fusing data from laser scanner sensors. In this approach, the sonar data and laser measurement are combined using the fuzzy Kalman filter. Finally, the simulation results show that the proposed algorithm offers advantages over previous methods and has improved performance.