In a recent article published in the Journal of Engineering and Applied Science, researchers introduced a new low-cost system for simultaneous localization and mapping (SLAM) in unknown indoor environments. The aim was to develop a system capable of mapping and localizing itself while identifying larger objects in real indoor settings.
Study: Low-Cost SLAM System for Indoor Mapping. Image Credit: Gorodenkoff/Shutterstock.com
Background
SLAM (Simultaneous Localization and Mapping) is a computational process used by mobile robots to create a map of an unknown environment while simultaneously determining their location within that map. This dual challenge arises because mapping requires accurate localization, and localization relies on the map being created. The robot executes its tasks autonomously without external aids such as GPS or indoor positioning systems (IPS), relying solely on its onboard sensors for environmental perception and navigation.
To tackle the SLAM problem, various approaches and algorithms have been developed, including Kalman filters, particle filters, and smoothing techniques. These methods rely on sensors such as laser scanners, light detection and ranging (LiDAR) systems, or stereo cameras to measure distances and angles to environmental landmarks.
LiDAR systems are known for their high accuracy but come with a higher cost. On the other hand, stereo cameras offer a balance between cost and accuracy, making them a more economical choice while still providing satisfactory performance.
About the Research
In this paper, the authors designed, developed, and implemented a low-cost SLAM system using a stereo camera as the primary distance sensor. The system is composed of four main components: a mobile robot platform, a stereo camera, a Linux-based microcomputer, and a ground station machine.
-
Mobile Robot Platform: The platform is a differential drive robot featuring two wheels for movement and turning. Constructed from affordable plywood and acrylic, it includes optical wheel encoders to estimate the robot's position and heading. An Arduino board is used to control the robot and manage communication with the microcomputer.
-
Stereo Camera: The system utilizes an Xbox 360 Kinect module, which captures environmental images and calculates depth through computer vision and stereo imaging algorithms. The Kinect module is connected to the Raspberry Pi 3 microcomputer via USB.
-
Linux-Based Microcomputer: The Raspberry Pi 3, running a Linux operating system, executes the SLAM algorithm and processes data from the Kinect and the wheel encoders. It also handles wireless communication with the ground station through a Wi-Fi router.
-
Ground Station Machine: A laptop serves as the ground station, where it processes data received from the robot, constructs the environmental map, and estimates the robot's pose. The ground station uses the Robot Operating System (ROS) for map visualization and to facilitate communication with the microcomputer.
The SLAM algorithm employed is the Rao-Blackwellized Particle Filter (FastSLAM). FastSLAM combines particle filtering with the Rao-Blackwellized technique, dividing the SLAM problem into robot localization and landmark estimation. It uses a modified particle filter to track the robot's path and multiple Kalman filters to estimate landmark locations.
Research Findings
The SLAM system was tested both in a virtual simulation platform and in real indoor environments. It successfully identified and mapped objects larger than 30 cm × 30 cm × 30 cm. The system demonstrated effective autonomous operation without the need for external sensors and achieved lower costs compared to similar LiDAR-based systems.
Two experiments were conducted:
- Tall Corridor with Obstacles: This experiment involved navigating a narrow corridor filled with obstacles.
- Large Hall with Many Obstacles: This experiment took place in a spacious hall with numerous obstacles.
During testing, the researchers presented the following:
- Real Environment Images: Red-green and blue (RGB) images captured by the Kinect module.
- Occupancy Grid Maps: Generated by the system, showing obstacles as black lines, free spaces as white, and unknown spaces as grey.
- Robot Trajectory and Landmark Locations: Visual representations of the robot's path and estimated positions of landmarks.
Performance evaluation was based on comparing the estimated robot pose and map with ground truth data obtained from a motion capture system. Metrics reported included:
- Root Mean Square Error (RMSE): For assessing robot pose and map accuracy.
- Computation Time: To evaluate the efficiency of the system.
The system achieved satisfactory accuracy and efficiency for commercial applications, such as cleaning and garbage collection in shopping malls, where extremely high precision is not critical.
Applications
The proposed system has potential applications in various fields requiring autonomous navigation and mapping unknown environments. These include exploration and rescue missions in hazardous or inaccessible areas, such as disaster zones, mines, and caves. It can also be used for surveillance and security tasks in indoor or outdoor environments, such as warehouses, parking lots, and airports.
Additionally, the system is suitable for service and entertainment tasks in places like hotels, museums, and amusement parks. Furthermore, it can be employed for education and research in robotics and computer vision, supporting teaching, learning, testing, and the development of new algorithms and techniques.
Conclusion
Applications
The proposed SLAM system has a range of potential applications across various fields that require autonomous navigation and mapping of unknown environments. Key areas of application include:
-
Exploration and Rescue Missions: The system can be employed in hazardous or inaccessible areas, such as disaster zones, mines, and caves, to navigate and map these environments for rescue and recovery operations.
-
Surveillance and Security: It is well-suited for indoor and outdoor surveillance tasks in environments such as warehouses, parking lots, and airports, enhancing security and monitoring capabilities.
-
Service and Entertainment: The system can be used in service roles within hotels, museums, and amusement parks, providing navigation and interaction capabilities for enhancing visitor experiences.
-
Education and Research: It offers valuable resources for educational and research purposes in robotics and computer vision. It supports teaching, learning, testing, and the development of new algorithms and techniques.
Conclusion
In summary, the novel system was low-cost and effective for SLAM in unknown indoor environments. It could map and localize itself simultaneously using noisy sensors and operate autonomously without external sensors. The system showed feasibility and effectiveness for commercial applications, such as cleaning and garbage collection in malls and similar facilities, where high accuracy is not essential. Moving forward, the researchers suggested improving the system by adding more sensors, optimizing the SLAM algorithm, and integrating additional navigation functions.
Journal Reference
Kassem, A.H., Asem, M. A low-cost robotic system for simultaneous localization and mapping. J. Eng. Appl. Sci. 71, 158 (2024). DOI: 10.1186/s44147-024-00486-8, https://link.springer.com/article/10.1186/s44147-024-00486-8
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.