A low-cost vision based hand gesture interface for real time industrial motor control in resource constrained environments.

Vision based hand gesture interfaces offer an intuitive means of human machine interaction, but their adoption in industrial control systems remains limited, particularly in small and medium scale industries where cost and hardware con- straints are critical factors. This study presents a low cost framework for real time industrial motor control using a simplified vision-based hand gesture approach. A lightweight convolutional neural network based on MobileNetV2 is adapted through transfer learning to recognize a binary set of hand gestures selected to ensure reliable operation under constrained computational conditions. Ges- ture recognition is executed on an external edge device, while control commands are transmitted wirelessly to an Arduino-based controller interfaced with a vari- able frequency drive for three phase motor actuation. Experimental evaluation shows stable gesture classification performance with low end-to-end latency and memory usage compatible with low cost hardware. System-level testing with an operational industrial motor confirms consistent real time response and safe con- trol behavior. The findings indicate that practical gesture based industrial motor control can be achieved without reliance on expensive automation platforms, offering a feasible pathway for incremental adoption of intelligent interfaces in resource limited manufacturing environments.