Intermittency is a fundamental property of space plasma dynamics, characterizing turbulent dynamical variables as well as passive scalars. Its qualitative and quantitative description from in-situ data requires an accurate estimation of the probability density functions (PDFs) of fluctuations and their moments, particularly the flatness, a normalized fourth order moment of the PDF. Such a statistical description needs a sufficiently large number of samples for the computation to be meaningful. Due to inherent technological limitations (e.g., limited telemetry bandwidth) not all samples collected on-board can be sent to the ground for further analysis. Therefore, a technology designed to process on-board the data and to compute the flatness is useful to fully exploit the capabilities of scientific instruments installed on robotic platforms, including nanosatellites. We designed, built, and tested in laboratory such a technology based on Field Programable Gate Arrays (FPGA). The building principle is the classical estimation of PDFs and their moments, based on normalized histograms of a measure. The technical design uses the FloPoCo (Floating-Point Cores) framework with customized arithmetic operators; the computation block is a pipelined architecture, which computes a new value of the flatness in each clock cycle. The design and implementation achieve optimization directives of the FPGA resources relevant for operation in space, like area, energy efficiency, and precision. The technology was tested in laboratory using Xilinx SRL16 or SRLC32 macros and provides correct results validated with test time series provided by magnetic field data collected in the solar wind by ULYSSES spacecraft. The characteristics and performance of the laboratory prototype pave the way for a space qualified version of the laboratory design.