Skip to content

Latest commit

 

History

History
139 lines (101 loc) · 6.71 KB

README.md

File metadata and controls

139 lines (101 loc) · 6.71 KB
Raw

IITISoC-24-IVR7-Motion-Planning-with-Controls-for-Self-Driving-Vehicles

Goal:

To develop a reliable Motion Planner with Control algorithms for collision-free path planning and efficient navigation in a simulated environment.

People Involved :

Mentors:

Members:

Project Overview & Solution

A three-wheeled vehicle is modeled and is visualized in gazebo simulator and RViz. The world file provided is mapped using Simeltaneous Mapping and Loacalization (SLAM) method. For Path planning RRT* algorithm were used on the map generated and PID algorithm is used for control algorithm.

Vehicle Modelling

Here we explains how to model a three-wheeled vehicle using URDF files and Xacro in ROS2, focusing on splitting the model into components for better organization and understanding:

Prerequisites

1.Create aworkspace

mkdir ~/ros2_ws/src
   cd ~/ros2_ws/src

2.clone the repo

git clone https://github.com/IVDC-Club-IIT-Indore/IITISoC-24-IVR7-Motion-Planning-with-Controls-for-Self-Driving-Vehicles.git

2.Build the workspace with the colcon build

cd ~/ros2_ws
colcon build
source install/setup.bash

File discription

robot.urdf.xacro-main file contains other conmponents of the file such as chassis,lidar,imu chassis.xacro-contains upper and lower part of the chassis robot_core.xacro-contains the right and left wheel, caster wheel inertial.xacro-contains the interial properties of the vechile lidar.xacro-contains details of the lidar and pulgins foe the lidar imu.xacro- has the details of the imu sensors and plugin ros2_control_wheel.xacro-this file for define the diff_controllers for the better control over the vechile

Launching the vehicle in the gazebo

ros2 launch july_9 launch_sim.launch.py

image

Launching the vehicle in the gazebo with world

ros2 launch july_9 new_launch_sim.launch.py

During SLAM the lidar range extended to infinity, due to which queue size got filled up and therefore was not able to send messages to RViz.

Mapping the world file

The orange_igvc.world world file is mapped using SLAM method with the help of Navigation2 stack of ROS2 and turtlebot3 robot.

sudo apt update
sudo apt install ros-humble-navigation2 ros-humble-nav2-bringup ros-humble-turtlebot3*

Go to the bashrc file and set TURTLEBOT3_MODEL=waffle

Run the launch file to open the orange_igvc world file in gazebo and run the turtlebot3_teleop node to control the turtlebot3 robot. To visualize it in RViz the turtlebot3_cartographer is used and the map is saved using nav2_map_server

ros2 run nav2_map_server map_saver_cli -f maps/maps2
cd maps
ls

On opening the .pgm file the map we get

world_map

Path Planning

Path Planning gives the vehicle's trajectory from its starting state to some goal state in the map. There are most popularly two approaches for path planning - Search based methods & Sampling based methods. In Search based algorithms, cost to access each node from start node in graph is calculated, to get the most optimal or shortest path to reach the goal node. Common examples being Dijkstra's and A* algorithms. As the size of map and dimensions increases they become computational expensive but still give us the shortest possible path.

Sampling based algorithms doesn't require fully exploring the configurational space and it also requires lesser computational power, though it doesn't give the most optimal path as output. RRT* Algorithm( modification of RRT ), which gives the nearest to most optimal path for the robot, is an example of sampling based algorithm.

The RRT* algorithm has been used for path planning. The code implementation takes the binary occupancy grid map as .npy file for detecting the obstacles and the start point, goal point and waypoints are given as input co-ordinates. The final path is visualized using the matplotlib library.

Visualisation of waypoints is as follows:

Visualisation of path planned on the binary occupancy map is as follows:

Comparison of Path Planning Algorithms https://docs.google.com/spreadsheets/d/17M89Dh4DRQxYJfTK4u48PiHGHFnKuStqf_kaLztgRA0/edit?gid=0#gid=0

Control Algorithm

Path Following and Control Algorithm

  • The path following (Controller) algorithm controls the bot to follow the waypoints generated by path planning algorithm. It consists of two types of controls, longitudinal and lateral. Longitudinal control regulates the cruise velocity of the vehicle while the lateral control steers the bot to follow the path.
  • Closed-loop controls like black box controls, geometric controls and optimal controls are widely used for these purposes. This project uses PID as a black box control method for both lateral and longitudinal controls.
  • The errors, namely Cross Track Error (CTE) and Orientation Error that can occur during autonomous driving can be minimized and subsequently eliminated using the control algorithms.

Process flow