Path planning for mobile robots

Project Title

Path planning for mobile robots

Project Area of Specialization RoboticsProject Summary

Path planning is one of the most fundamental problems in the navigation of mobile robots. Determination of a collision-free path for a robot between start and goal positions in an obstacles cluttered workspace is central to the navigation of an autonomous robot. This project proposes a novel technique that incorporates both formal and neural network techniques to plan an optimal path for the navigation of snake robots in an environment. A robust path planning algorithm does not only save operation time but also reduce the wear and capital of mobile robot. In this study, we propose an optimal path planning algorithm for a bio-inspired snake robot. Due to the modular structure and rapid movement, it is a challenging task.
 For navigation, the proposed algorithm will read the map of the environment or workspace and subsequently attempts to avoid obstacles in the environment and find a free path for the snake robot to traverse in the workspace without collision. The proposed algorithms will be able to find an optimal path for the robot to traverse, even if there are a large number of obstacles cluttered in a complex environment. To achieve this, the main aim is to replace environment sampling with a Target bias strategy to reduce the complexity and convergence problems and use a neural network for curve post-processing.
     
  A lidar (Light Detection and Ranging) sensor will be used to make the robot able to perceive the environment and take decisions accordingly. If an obstacle happens in its way, it will change its direction. The algorithms help the robot decide which predefined path it should select for a particular scenario. For optimization, neural network will be utilized. 

Project Objectives

The project focuses on the hardware and software implementation of a snake robot. Which starts from simulation to the final hardware product.

1. Deep understanding of robotics, ROS, and different real-time simulators.
2. Simulation of the robot in CoppeliaSim (a physics engine for robotics simulation).
3. Understanding the path planning algorithms with respect to optimization.
4. Designing and implementation of sampling and Neural Network-based path-planning algorithm for the navigation of a mobile robot.
5. Designing and implementation of a path planning algorithm for the navigation of a snake robot.
6. Implementation of the designed path planning algorithm on a hardware snake robot.

Project Implementation Method

This project will mainly be focusing on three sections. Which are as follows:

Path Planning Algorithm:

Path planning is the basis of the navigation of a mobile robot from its initial state to its goal state. Path planning is one of the main sections of navigation in a mobile robot. Because it helps the robot choose the best available path for its navigation. The best available path should be free from collision, low computation cost, and safe. There are many path algorithms including A*, RRT, RRT*, D*, etc. which are different using their optimality and computational criteria. So, in this project, RRT* is being used because it is suitable and optimal in almost all environments. Which uses samples of the environment to get the best available path for its navigation. Whereas other algorithms do not use samples. 
The following figure is showing the methodology of path planning.

                                                                          Figure 1 Methodology of Path Planning

Robotic Simulator:

We are using CoppeliaSim for the simulation of the robot. which provides all types of tools for real-time simulation. So first we will test our proposed algorithm in a simulation environment. This environment will be developed in CoppeliaSim this environment will have all types of known obstacles to replicate the real environment for critical testing. For coding, Python language will be used to make logic (detect an obstacle, its height, distance) using external APIs that CoppeliaSim provides.

Hardware Requirements:

After evaluation of the algorithm in a simulation environment, it will be implemented on the snake robot. The available snake robot has 10 actuated joints which give them superior ability to flex, reach and approach a huge volume in their workspace. We will mount all the other required components such as LIDAR for environment perception and Raspberry pi 4 for the processing.  

Benefits of the Project

This project focuses on a novel technique that finds the optimal path for a snake robot that can be used for rescue operations in a collapsed building, spying, and inspection of pipes in any industry. Our path planning technique will help the snake robot to navigate easily in any environment. So, the proposed algorithm will help the application of mobile robots to decide which path will be best for navigation. The most popular mobile robot application is as follows

1. Navigation in autonomous vehicles
2. Unmanned Aerial Vehicles (UAVs)
3. Planetary and space missions
4. Surveillance Planning

Technical Details of Final Deliverable

The snake robot will be simulated first in CoppeliaSim. The snake robot that we are using consists of 10 joints. Which helps the robot to navigate between points where each joint is build with a motor. So, in order to move the snake first, we will define any suitable environment with all types of obstacles so that the robot can easily navigate between obstacles in the environment. Then the path planning technique will be used which helps the robot to decide the best path. For coding, Python language will be used to make logic (detect an obstacle, its height, distance) using external APIs that CoppeliaSim provides, and then we will switch from simulation to real-world implementation and interface with the hardware with the help of ROS (Robot Operating System).

The robot will get samples from the environment with the help of the lidar sensor, which allows the creation of 3D images of the detected objects and mapping of the surroundings. The snake robot will be able to take self-decisions and reduce human interaction and effort. which helps snake robots to find an optimal path in lesser time. The snake robot can be used in surveillance and rescue operations. Moreover, the lidar sensor can be configured to create a full 360-degree map around the vehicle rather than relying on a narrow field of view and image segmentation technique, the robot will estimate the distance of the obstacle and determine its path. The depth images from the lidar sensor will be received by a raspberry pi. The controller will process the images and the result will be sent to the robot joints and it will then act accordingly.

The end product when presented will be a snake robot that will be intelligent enough to take self-decisions and perform its respective tasks.

Final Deliverable of the Project HW/SW integrated systemCore Industry OthersOther Industries Transportation , Security Core Technology OthersOther Technologies Artificial Intelligence(AI), RoboticsSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	79974
Raspberry Pi 4	Equipment	1	32500	32500
PLA 3D Printing Filament	Equipment	2	3200	6400
Lidar Sensor	Equipment	1	31100	31100
3D Printing	Miscellaneous	2	412	824
Power Cables	Miscellaneous	4	1100	4400
Jumpers	Miscellaneous	50	19	950
Thesis Printing and Binding	Miscellaneous	4	950	3800