Toward Fully Automated Inspection of Critical Assets Supported by Autonomous Mobile Robots, Vision Sensors, and Artificial Intelligence

Javier Sanchez-Cubillo,Javier Del Del Ser,José Luis Martin

DOI: https://doi.org/10.3390/s24123721

IF: 3.9

2024-06-08

Sensors

Abstract:Robotic inspection is advancing in performance capabilities and is now being considered for industrial applications beyond laboratory experiments. As industries increasingly rely on complex machinery, pipelines, and structures, the need for precise and reliable inspection methods becomes paramount to ensure operational integrity and mitigate risks. AI-assisted autonomous mobile robots offer the potential to automate inspection processes, reduce human error, and provide real-time insights into asset conditions. A primary concern is the necessity to validate the performance of these systems under real-world conditions. While laboratory tests and simulations can provide valuable insights, the true efficacy of AI algorithms and robotic platforms can only be determined through rigorous field testing and validation. This paper aligns with this need by evaluating the performance of one-stage models for object detection in tasks that support and enhance the perception capabilities of autonomous mobile robots. The evaluation addresses both the execution of assigned tasks and the robot's own navigation. Our benchmark of classification models for robotic inspection considers three real-world transportation and logistics use cases, as well as several generations of the well-known YOLO architecture. The performance results from field tests using real robotic devices equipped with such object detection capabilities are promising, and expose the enormous potential and actionability of autonomous robotic systems for fully automated inspection and maintenance in open-world settings.

engineering, electrical & electronic,instruments & instrumentation,chemistry, analytical

What problem does this paper attempt to address?

The key problem that this paper attempts to solve is how to use autonomous mobile robots (AMRs), visual sensors and artificial intelligence technologies to achieve fully - automatic detection of key assets in an industrial environment. As the industry increasingly depends on complex machinery, pipelines and structures, accurate and reliable detection methods become crucial to ensure the integrity of operations and reduce risks. Traditional detection methods have limitations, especially in object recognition and classification in complex environments. Therefore, this research aims to evaluate and verify the performance of AI - assisted AMRs in performing specific tasks and self - navigation, especially in three real - world transportation and logistics cases, by using different generations of the YOLO architecture for object detection.

Specifically, the paper explores the following points:

1. **Improving Detection Accuracy and Reliability**: Traditional sensors are very efficient in measuring distance, speed, acceleration, temperature and frequency, etc., but have limited ability to recognize objects in complex situations (such as key asset detection). AI - assisted AMRs improve the accuracy and reliability of detection by providing a more advanced cognitive understanding.

2. **Reducing Human Error**: Humans are easily distracted during the detection process and have a slow reaction time. Autonomous robots can recognize and respond to emergencies more quickly, reducing human error.

3. **Real - time Insights**: AI - assisted AMRs can provide real - time information on asset conditions, which helps to discover and solve problems in a timely manner and improve operational efficiency.

4. **Adapting to Unknown and Unstructured Environments**: In unknown and unstructured environments, traditional sensors and signal processing methods are difficult to work effectively. AI - assisted AMRs can reliably perform tasks in these environments through advanced perception and decision - making capabilities.

5. **Multipurpose Applications**: The AMRs and AGVs involved in the research are not only used for simple cargo handling, but also for complex detection and maintenance tasks. For example, weed detection on railway tracks, automatic detection of port containers and automatic loading of goods.

Through these studies, the paper demonstrates the great potential of AI - assisted AMRs in industrial detection and maintenance tasks and provides an important reference for future development.