Robot localization in a mapped environment using Adaptive Monte Carlo algorithm

Sagarnil Das

2025-01-02

Abstract:Localization is the challenge of determining the robot's pose in a mapped environment. This is done by implementing a probabilistic algorithm to filter noisy sensor measurements and track the robot's position and orientation. This paper focuses on localizing a robot in a known mapped environment using Adaptive Monte Carlo Localization or Particle Filters method and send it to a goal state. ROS, Gazebo and RViz were used as the tools of the trade to simulate the environment and programming two robots for performing localization.

Robotics

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper mainly explores how to use the Adaptive Monte Carlo Localization (AMCL) algorithm to determine the position and pose of a robot in a known environmental map. Specifically, it solves the problem of accurate positioning of the robot in the environment, which is a crucial step for mobile robots to achieve autonomous navigation. The following are the specific types of problems involved in the paper:

1. **Local Localization**:
- This is the simplest positioning problem, also known as position tracking.
- The robot knows its initial position and direction, and the challenge is to estimate the position and pose of the robot as it moves in the environment.
- Although there is motion uncertainty, this uncertainty is limited to the area around the robot.

2. **Global Localization**:
- This is a more complex positioning problem.
- The initial position and direction of the robot are unknown, and its position must be determined according to the real - time map.
- The uncertainty is high because the robot could be located anywhere on the map.

3. **The Kidnapped Robot Problem**:
- This is the most challenging positioning problem.
- Similar to the global localization problem, but the robot may be "kidnapped" at any time and moved to a new position on the map.
- The robot needs to be able to recognize this sudden change and re - position itself.

### Solutions

To address the above problems, the author adopts the Adaptive Monte Carlo Localization (AMCL) algorithm. This algorithm simulates the positioning process of two robots (UdacityBot and SagarBot) in a known environment through the Particle Filters method. Specific steps include:

- **Sensor Data Processing**: Use data collected by sensors such as RGB cameras and laser rangefinders to update the position and pose estimates of the robot.
- **Particle Resampling**: Adjust the weights of particles according to sensor data and resample to improve positioning accuracy.
- **Path Planning**: After successful positioning, guide the robot to the target position.

### Tools and Environment

- **ROS (Robot Operating System)**: A framework for developing and running robot applications.
- **Gazebo**: A tool for simulating robots and their environments.
- **RViz**: A tool for visualizing robot states and sensor data.

Through these tools and algorithms, the paper demonstrates how to achieve accurate positioning and navigation of robots in complex environments.