AI-Driven Safety Systems: Reducing Risk in Complex Workplaces and High-Stakes Task

Complex workspaces that involve workers, machines, and tools often harbor residual risks due to both intended and unintended interactions among these elements. For instance, many incidents arise from operator misuse of machinery, such as working with tampered safeguards or completely removed protections. Additionally, dangerous situations can occur when machines operate without specific auxiliary devices, increasing potential risks within the workspace. The state of the art only provides written warnings in the use and maintenance manual for most of these risks and no more efficient technical solutions have been proposed. In this study, a prototypal machine assembly, comprising a robot and a multimodal lathe, is utilized to assess and mitigate risks in a complex workplace. This workspace is divided into different pre-defined zones depending on the task to be completed and the presence of operators. A comprehensive risk assessment is conducted before and after integrating IoT sensors, such as RFID tags and Computer Vision, integrated with an Artificial Intelligence (AI) module with standard safety systems that comply with the Machine Directive. This approach offers promising solutions to mitigate the consequences of operator errors and potential machine malfunctions. The paper explains how well-known hazards in workspace can be significantly reduced by strategically deploying sensors to monitor specific tasks. Multiple sensors are employed to oversee the case under examination, ensuring a redundant check with sensors based on different technologies to reduce Common Cause Failures of those innovative systems. By combining traditional safety measures with advanced sensor technologies and AI, it is possible to enhance the overall safety of complex workspaces. The proposed system not only addresses common hazardous situations but also provides a proactive approach to managing risks, ultimately contributing to a safer working environment.

Keywords: Artificial intelligence, Machinery safety, Errors, Computer vision, RFID.