Implementing SLAM on a car-like robot using the navigation stack in ROS.
Hi everyone, I have been assigned to work on Ackermann type robot and my task is to make it autonomous and implement SLAM on it. I did consider ros-navigation stack and I came across this post http://answers.ros.org/question/9059/how-can-i-use-the-navigation-stack-on-a-carlike/.
Before I shoot out my questions,I would like to give a picture of whats going on. here is video for my robot:http://www.youtube.com/watch?v=8fZYF8Zn3zA
Components of the robot: 1.sdc2130 motor driver 2.encoders from US digital 3.SICK LIDAR(LMS-200) 4.2 lead acid(12 V ) batteries. 5.2 DC motors
Goals: 1.To achieve SLAM in unkown environment and also to achieve autonomous navigation and obstacle avoidance.
Tasks accomplished: 1.Developed a ros node for the motor control as shown in the video. 2.Developed a ros node for reading the encoder values and pumping out the data to the odometry node.
Things yet to be done: 1.identifying the ideal motion primitive and implementing it in the ros-navigation(which I have no clue how to start). 2. Rewriting the local planner.
Questions: 1. I would like to know whether the navigation stack could be implemented successfully in this type of robot. 2.Is there any other approach in ROS to achieve SLAM in this type of robot?.
I am aware that the files trajectory_planner.cpp ,trajectory_planner.h(Local planner) in navigation/base_local_planner are the ones which need editing. I also need to know how to choose the right motion primitives in SBPL and how to implement it in ros navigation stack.
Please share your views on this. Thanks for your time.
regards, Karthikeyan Yuvaraj