Implementace AI kamer pro zvýšení bezpečnosti provozu vysokozdvižných vozíků

Abstract

This master's thesis addresses the issue of increasing the safety of forklift operations through the implementation of modern artificial intelligence systems. The theoretical part analyzes the current state of safety in intralogistics, the legislative framework, and occupational accident statistics, which confirm that handling equipment is the cause of up to 35% of fatal accidents in the Czech Republic. The analytical part employs the Failure Mode and Effects Analysis (FMEA) method to quantify risks during the interaction between machinery and pedestrians within a manufacturing plant environment. The core of the thesis is the experimental verification of the LUIS Artificial Intelligence (AI) system on a Linde H30CNG-02 forklift. The measurement results demonstrated that the deployment of Edge AI technology allows for a reduction in the Risk Priority Number (RPN) by more than 80%, from 168 to 32 points. However, the experiments also revealed technical limitations of the system, particularly a drop in detection success at speeds above 13 km/h and failure in zero-light conditions. The conclusion of the thesis proposes an implementation methodology including electronic speed limitation, workplace lighting standardization, and operator training.