This paper presents a gesture recognition system for automotive applications. The system consists of four self-capacitance sensors, an Analog Front-End (AFE) designed in standard 0.18 um CMOS technology and a digital core for signal processing and gesture identification. All blocks are taken into account during the design process and therefore, the most demanding tasks during the data acquisition and processing chain are implemented in the most appropriate and optimal block, minimizing the resources needed to perform the tasks. This leads to an optimized system design of high performance and reduced cost. A proof-of-concept system is designed for architecture verification reaching promising results.