Localization of a snake robot

Humans perform many tasks efficiently, but some operations negatively impact human health and life, such as working in a hazardous environment, diffusion of bombs, locating survivors in collapsed buildings, and working in nuclear power plants. Therefore, one should perform such tasks without endange

Project Title

Localization of a snake robot

Project Area of Specialization

Robotics

Project Summary

Humans perform many tasks efficiently, but some operations negatively impact human health and life, such as working in a hazardous environment, diffusion of bombs, locating survivors in collapsed buildings, and working in nuclear power plants. Therefore, one should perform such tasks without endangering human life. The researchers and scientists proposed numerous intelligent robotics to perform operations in these hazardous environments. These emerging technologies are advancing at a rapid pace, and the development of robots has now progressed to the development of bio-inspired robots so they can eliminate dangerous jobs for humans because they are capable of working and navigating in a hazardous environment.

Mobile robots require knowledge of their position and orientation in an environment, known as localization. Localization enables the robot to have effective navigation.

In this study, we consider a bio-inspired snake robot is considered for search and rescue operations because human and wheeled robots can’t search for survivors or explore the area under collapsed buildings or in any other disaster-affected areas, whereas snake robots can effectively navigate on rough surfaces, complex terrains, do dangerous jobs and they can also move through and around pipes, giving them applications in surveillance and inspection of nuclear power plants and other industrial plants, which is not suitable for wheeled robots.

Navigation of mobile robots comprises localization and path planning. And navigation of this bio-inspired snake robot is a challenging task. In this study, we focus on localization, which estimates the position of a mobile robot in an environment. Without position information, it is unable to perform any task. Localization demands the map of the environment and finding the location and orientation of the robot on this map. The localization is a state estimation problem due to uncertainties in sensors. For the localization of a mobile robot, we combine multiple sensor data and employ probabilistic estimation techniques such as Kalman and particle filters to estimate its position and orientation. In this study, we focus on the localization of a bio-inspired snake-like robot.

Project Objectives

OBJECTIVES:

1. Comparison and evaluations of the state-of-the-art mobile robot localization techniques.
2. Data fusion of various sensors for robust and accurate localization.
3. Simulation and testing of a snake robot in various unknown terrain.
4. Implementation and testing of the developed localization method on a simulated and real bio-inspired snake-like robot.

Project Implementation Method

Software:

We use Robot Operating System (ROS) to develop a localization algorithm. ROS is an open-source and flexible software framework where we can employ reusable algorithms for robots. Firstly, we test the localization algorithm in a simulation environment and then develop in CoppeliaSim. CoppeliaSim is a simulator for robots. It provides advanced tools which make testing and simulation easy. ROS and CoppeliaSim can be easily interfaced, which allows for parallel computation and real-time testing. All sensor data is published using ROS topic by CoppeliaSim, and ROS nodes subscribe to these topics and utilize this data for algorithm development. ROS has several already developed packages that can be utilized for any robot.

Probabilistic approaches:

We propose a probabilistic approach for the localization of a mobile robot. Many probabilistic techniques exist in the literature, such as the Kalman Filter (KF), Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF). We fuse the data of multiple sensors to perform an accurate localization and evaluate the algorithm critically and assess the trade between computational cost, accuracy, and noise.

Kalman Filter estimates the future decision based on previous information and performs efficiently in a linear environment. But its performance degrades in a nonlinear environment. So, to overcome this challenge, its modified versions (EKF and UKF) are used, which linearize nonlinear systems, reduce uncertainty, and estimate data accurately for a nonlinear environment. We apply EKF to the acquired data by fusing the IMU data, odometry data, and stereo camera to achieve effective results. Stereo cameras use three-dimensional images of the surrounding.

Hardware:

We use multiple Inertial Measurement Units (IMU), which have built-in accelerometers and gyroscopes. An accelerometer measures linear acceleration (position), while a gyroscope estimates the angular velocity (orientation). We localize a robot using IMU, but it has limited performance with uncertainties. Therefore, we compensate for these uncertainties by fusing odometry data and a stereo camera with IMU. Odometry uses data from moving sensors to estimate changes in position over time. Visual Perception of the physical world and identification of survivors is carried out using a stereo camera. It allows to simulate human binocular vision and gives it the ability to perceive depth. We mount it on the head of our snake robot to detect the survivors.

                                                                                                                                              

Benefits of the Project

The hyper-redundant and modular snake-robot can maneuver in complex environments such as collapsed buildings, coal mines, and irregular terrains to identify human survivors. The snake robots rely on their modular structure and physical dimensions to move in very narrow places, on different terrains, on the ground, and in the water, where the wheeled and legged robots fail. Our goal is to build this robot close to a biological snake's capabilities as much as possible to exploit its flexible movements. Our long-term goal is to turn this project from a prototype into an actual robotic system that can be deployed in the earthquake and other disaster-affected areas and save lives by bringing state-of-the-art rescue equipment for disaster relief teams in Pakistan.

Technical Details of Final Deliverable

The deliverable of our work is a bio-inspired snake robot that can localize and traverse various environments for search and rescue operations. This robot can perform numerous tasks such as exploration, spying, and blockage detection in pipelines. One can implement the proposed localization algorithm to other mobile robots and intergrade for SLAM.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Others

Other Industries

Education

Core Technology

Robotics

Other Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com