Designing a CAN-Based Communication Hub with the S32N55 using Model Based Design Toolbox
1
Table of Contents
• Introduction
• Overview
• Context
• References
• Conclusion
2
Introduction
This article presents an automotive system built around a central computer that processes high volumes of data to manage interactions and decisions across the vehicle. Implemented on an NXP S32N55 board, a main node orchestrates peripheral nodes — Lighting, Motor Control, Steering, Radar, and Parking Sensors — over CAN, demonstrated through real-time interactions and Driver-in-the-Loop (DiL) simulations. The same architecture also enables stimuli and scenarios to be injected directly from Simulink/MATLAB via the Model-Based Design Toolbox (MBDT), turning the setup into both a functional prototype and a flexible test bench that shortens the loop between design, validation, and refinement.
3
Overview
The communication hub acts as a comprehensive aggregator and decision-maker, serving as the central intelligence of the entire automotive control network. This architectural choice follows industry's best practices by consolidating critical decision-making processes into a single, robust processing unit capable of efficiently managing multiple concurrent data streams and executing time-sensitive commands. Centralizing this logic also simplifies maintenance and traceability, since the rules governing vehicle behavior live in one well-defined place rather than being scattered across multiple ECUs.
For a project of this nature, the NXP Model-Based Design Toolbox (MBDT) offers a practical development path: control logic and application behavior can be designed in Simulink/MATLAB and deployed directly onto the S32N55, without a separate hand-coding step. The graphical, model-based workflow makes the system's structure easier to follow and adjust, while built-in support for CAN communication and integration with tools like FreeMASTER for live telemetry simplify both stimulus injection and runtime observation. The result is a smoother path from initial concept to a working prototype that can be iterated on and validated in a controlled, repeatable way.
In this specific implementation, the main node hosts an application that fulfills two complementary roles: data aggregator and decision-maker. As an aggregator, it collects, synchronizes, and interprets incoming signals from the sensing nodes; as a decision-maker, it translates that fused view of the environment into concrete commands for the actuators. Practically, our system receives data over CAN from the peripheral sensing nodes (Radar, Parking Sensors) and dispatches commands to the actuator nodes (Motor Control, Lights, Steering).
The main node is also designed to make safety-critical decisions based on the incoming inputs — for example, triggering Automated Emergency Braking (AEB) when the Parking Node or the Radar Node detects a hazardous situation. Because these decisions are made centrally, the response logic can take the full context into account (vehicle speed, proximity of obstacles, current steering input) rather than reacting to a single sensor in isolation.
4
Context
At its core, the main node receives a continuous stream of data over the CAN bus from peripheral nodes distributed throughout the vehicle. These peripheral nodes include:
Radar sensors — provide long-range object detection and relative velocity measurements, making them ideal for highway-speed scenarios and forward collision awareness.
Parking sensors — monitor the immediate vicinity of the vehicle for obstacles and potential collision risks, typically at very short range and at low speeds.
Fault sensors — for actuator nodes, like the motor control, steering and lighting systems.
The CAN bus protocol guarantees the reliable, deterministic communication required to meet the stringent timing demands of automotive safety systems. Its built-in arbitration, error detection, and message prioritization make it a natural fit for a distributed architecture in which safety-relevant signals must always reach the main node within a bounded time window.
To streamline communication across components, a CAN Database ( DBC ) file has been created that contains all the signals and messages used throughout the system. The DBC file acts as a single source of truth for the entire network: every node — whether sensing or actuating — references the same definitions for message IDs, signal layouts, scaling factors, and value ranges. This drastically reduces the risk of integration mismatches when multiple boards are developed in parallel.
Beyond its data aggregation role, the main node also serves as the command center for the vehicle's actuator systems. After receiving data from the simulation, it is being processed and then it transmits precisely timed control signals to critical subsystems, including the motor control unit, lighting system, and steering mechanism. This bidirectional architecture enables closed-loop control strategies, in which sensor feedback continuously informs actuator commands to achieve the desired vehicle behavior. Each actuator node remains responsible for the low-level handling of its hardware, while the main node provides the high-level command to the actuators.
Since the main node is responsible for receiving, analyzing, processing and sending data, it also becomes the one responsible for sharing the telemetry information upstream, either to the cloud, or to real time monitoring tools like FreeMASTER.
A particularly valuable aspect of this system is its seamless integration with the Simulink/MATLAB environment, which unlocks extensive possibilities for system validation and scenario testing. Engineers can inject stimuli into the simulation and analyze a wide range of driving conditions and edge cases without requiring a full-scale prototype. This is especially useful for reproducing rare or dangerous situations — such as sudden obstacles or sensor faults — in a fully controlled and repeatable environment.
To achieve two-way communication between the main node and the simulation, the CAN bus itself is used to communicate with the Simulink model. This way, the physical prototype can feed stimuli into the simulation — and vice versa — on the same CAN bus that devices are using to communicate, significantly expanding the boundaries of the testing environment. The same DBC file that defines the on-vehicle communication is reused on the simulation side, ensuring that the messages exchanged between the real and virtual worlds remain perfectly consistent.
Note: Perhaps one of the most noteworthy features of the main node's active functions is its ability to make safety-critical decisions in real time based on aggregated sensor inputs.
The system continuously monitors data from both the parking sensors and the radar node, detecting potentially dangerous situations that require immediate intervention:
At low speeds — hazard detection is typically driven by the parking sensors mounted on the front and/or rear of the vehicle, where short-range, high-resolution distance measurements are most relevant.
At driving speeds — the radar module takes over, collecting and analyzing data that is then forwarded to the main node for higher-level interpretation.
In both scenarios, the main node remains the ultimate decision-maker, fusing all available data to determine the appropriate response. This clear separation between sensing, decision-making, and actuation keeps each component focused on a single responsibility and makes the overall system easier to reason about, extend, and validate.
5
References
NXP Model-Based Design Toolbox (MBDT) Community
Interacting with Digital Inputs/Outputs on MR-CANHUBK344
Communicating over the CAN Bus
S32N Vehicle Super-Integration Processors
6
Conclusion
This article has provided an overview of the communication hub's core functionality, offering a high-level perspective on how key systems interact within the overall architecture. The main node was presented both as a data aggregator and as a decision-maker, with a particular emphasis on its role in safety-critical scenarios and its integration with the Simulink/MATLAB environment.
Future installments in this series will take a deeper dive into the communication hub — covering the specific board in use, detailed hardware and software requirements, and other technical considerations and implementation nuances. Subsequent articles will also explore individual peripheral nodes in more detail, building up a complete picture of the system one subsystem at a time.
記事全体を表示
