EMBEDDED IN WHAT MATTERS.
Mission- and safety-critical embedded software from concept to deployed code
Embedded engineering partner for R&D teams
X2CON is a specialized embedded engineering consultancy. We work in Model-Based Design and multi-domain system engineering under industry standards, like ECSS, DO-178C, ISO 26262 and IEC 62304. We are one of fewer than 50 official MathWorks Consulting Partners worldwide, and an Altera Solution Acceleration Partner (ASAP). Within ASAP we are the only European FPGA Model-Based Design consultancy.
We are a Budapest-based SME founded in 2019, with customers in space, nuclear, medical, automotive, wireless and defence across Central and Western Europe and the US.
Official MathWorks Consulting Partner
Everything we build runs on MATLAB and Simulink. The model is the executable requirement, the C/C++ or HDL comes out of that model, and the verification keeps the two in agreement all the way to the target. That is the toolchain your algorithm team already works in, and it lets us move verification forward: model-in-the-loop, SIL and PIL, HDL co-simulation, all of it before the first board comes back.
We publish through the Connections Program as well. FdIdent Toolbox does frequency-domain system identification, MP Toolbox does arbitrary-precision arithmetic, and both came out of the MATLAB Laboratory at BME, where Quantization Noise (Widrow and Kollár, Cambridge University Press) was written. That book is the theory our fixed-point work rests on, and its instructor solution manual is still issued from our team.
Altera Solution Acceleration Partner (ASAP)
MathWorks on the model side, Altera on the silicon side. The expensive problem sits between them.
What the algorithm team hands over: a Simulink model, floating-point reference results and a specification document.
What the FPGA team needs: fixed-point word lengths that hold accuracy, an architecture that closes timing, and a test bench proving the two still match.
Months disappear in between. That step is our work: fixed-point conversion and noise budgeting grounded in the textbook above, HDL out of the model with DSP Builder or HDL Coder, verification by co-simulation and FPGA-in-the-loop, and training on your own design.
Capabilities
- Embedded SW Development. High-level system modelling in Simulink and System Composer, automatic C/C++ code generation with Embedded Coder, hand-written C/C++ and Python where a target calls for it, target integration, and CI pipelines for embedded control platforms and test systems. AUTOSAR-compatible automotive development under ISO 26262.
- Verification & Validation. Requirements-based testing, MIL, SIL, PIL and HIL simulation, static code analysis with Polyspace Bug Finder and Code Prover, formal verification, and full traceability per the applicable safety standard.
- Signal Processing & Data Acquisition. Custom signal conditioning, sensor and detector interfaces, and data acquisition chains for embedded measurement systems.
- Control System Design. Power electronics and motor control, system identification, simulation environments and virtual prototyping with Model-Based Design.
- FPGA & HDL Development. HDL code generation, algorithm optimization, FPGA integration, and hardware-software co-design, from schematic through to a working device.
- Engineering Applications. MATLAB applications, measurement data processing, and trending and reporting tools built to a customer’s own technical specification, including instrument automation, operator interfaces, and work inside regulated condition monitoring and ageing management programmes.
Specialist areas
Four problem areas our engineers have worked in for decades, and the ones customers most often bring us in for.
- Quantization, noise and finite word length. Signal quantization, finite bit length and roundoff, noisy signals, non-ideal system behaviour, external disturbances and sampling. Transform-domain processing with FFT, DCT and PCT. This is the theory behind the textbook above, and it is what a fixed-point conversion stands or falls on.
- Wireless and satellite communications. Modulation and demodulation, error correction coding with LDPC and Turbo codes, nonlinear effects across the transceiver chain, synchronization, and multicarrier modulation, OFDM in particular.
- System identification. Time- and frequency-domain identification of dynamic systems with nonlinearities, parameter estimation, adaptive signal processing and Kalman filters.
- Radar. Signal processing and simulation for pulse and FMCW radar, target detection, tracking and identification.