
【虹科方案】 嵌入式 AI 崛起: CAN 總線如何重塑實時控制與邊緣智能?
嵌入式 AI 正加速下沉到邊緣設備,CAN 總線憑藉可靠、實時與分散式特性,成為連接 AI 感測器與智能執行器的關鍵通訊架構。虹科解析 CiA 見解、CANopen 應用(含 CiA 462)與未來 PCAN XL 工具鏈,揭示 CAN + AI 的落地路徑與產業機會。
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.
With the rapid development of Software-Defined Vehicles (SDV) and centralized E/E Architecture, traditional CAN buses can no longer support the increasing volume of data and real-time control requirements.
Leveraging our deep experience in automotive network communication, embedded systems and FPGA technology, we have proposed a new technology based on CAN XL × TSN (Time-Sensitive Networking) It is a total solution that enables vehicle manufacturers and Tier-1 suppliers to build efficient, reliable and scalable in-vehicle communication networks of the future.
CAN XL (Controller Area Network eXtended Layer) is the third-generation protocol for CAN networks, combining the three core advantages of "low cost," "high stability," and "strong compatibility," and realizing a cross-generation performance leap through the new design of the physical and protocol layers.
Its main technical features include:
Ultra-high-speed transmission capability: Data rate up to 1-20 Mbit/s (arbitration domain ≤1 Mbit/s, data domain 1-20 Mbit/s)
Data length extension: Supports 1-2048 Byte, far exceeding CAN FD's 64 Byte limit.
Virtual CAN Network (VCID): Supports up to 256 logical virtual networks for greater isolation and flexibility
Multi-Protocol Compatibility: New SDT (Service Data Unit Type) field for interoperability with protocols such as Ethernet.
Hao Chen, Head of Technology at Rainbow Technology, pointed out that CAN XL is a new technology in the ADAS and in-vehicle radar systems This is a significant advantage in the current 2M CAN FD architecture. In the current 2M CAN FD architecture, for example, radar data needs to be split into multiple frames and re-synchronized by the ADAS controller, whereas CAN XL can utilize Single frame 2K payload Transmits all object data at once, dramatically improving synchronization and real-time performance.
In order to help the industry to experience this new standard quickly, HONGKE launched the PCAN XL KitThe content:
USB-to-CAN XL Interface Module
Professional CAN Analysis Software PCAN-Explorer 7
PCAN-Basic SDK programming interface
The kit fully supports CAN CC / FD / XL communications and implements the core functionality of the CAN XL protocol in an FPGA, providing a complete environment for development, test and verification.

As the electrification and intelligence of automobiles increase, the E/E architecture is moving from the traditional decentralized ECU to the "domain controller" and "central computing" forms. In this process, vehicle communication faces five major challenges:
Explosive growth in arithmetic demand
Slow hardware iteration
Fragmentation of communication interfaces and protocols
Increased real-time and functional safety requirements
Cost and Flexibility Contradictions Highlighted
Mr. Law Hin-chi, Senior TSN Engineer of HTC, pointed out thatFPGAs are the "software" of hardware.The three characteristics of "parallel processing", "low latency" and "reconfigurability" can precisely meet the above challenges.
Unlike CPU/GPU time-sharing, FPGAs use space-sharing, where each task is executed simultaneously by independent circuits, realizing true hardware-level certainty.
The FPGA-based TSN (Time Sensitive Networking) solution brings three core values to the automotive E/E architecture:
Breaking the performance bottleneck: Hardware latency in microseconds for ultra-low jitter communication
Unified Communications Architecture: Single chip replaces multiple discrete controllers, supports any-to-any protocol conversion
Upgrade for the future: Adaptation of new standards through OTA reorganization to perfectly match software-defined vehicle requirements
The complete range of FPGA TSN solutions from SOC-E / RelyUm Series TSN IP Cores and Boards,TSN Switches until (a time) Test and Verification Kits The product supports 10M / 100M / 1G / 2.5G multi-rate. The product supports 10M / 100M / 1G / 2.5G multi-rate, and built-in IEEE 802.1AS time synchronization, 802.1Qbv time-aware shaping, and other key TSN standards, which makes it easy to build a highly reliable and accurate in-vehicle network.

CAN XL and TSN are not in competition, but complement each other and work together. The former provides high bandwidth and low cost for low and medium-speed control layers, while the latter provides certainty and high reliability for high-speed backbone networks. In the Central + Zonal Architecture, the combination of the two is driving the automotive industry towards a more efficient and reliable network. Software-Defined Mobility A New Era.
As the automotive E/E architecture evolves from distributed ECUs to domain control, and then to a "centralized computing + domain control" system, the deep convergence of CAN XL and TSN will become a key force in the evolution of in-vehicle networks, laying the foundation for in-vehicle communication, safety and sustainable innovation.

嵌入式 AI 正加速下沉到邊緣設備,CAN 總線憑藉可靠、實時與分散式特性,成為連接 AI 感測器與智能執行器的關鍵通訊架構。虹科解析 CiA 見解、CANopen 應用(含 CiA 462)與未來 PCAN XL 工具鏈,揭示 CAN + AI 的落地路徑與產業機會。

ArangoDB 在金融反詐欺領域扮演關鍵角色,協助機構從傳統規則導向轉向深度關係洞察。面對日益複雜且隱蔽的詐欺網路,傳統風控系統因缺乏全局視角、規則更新滯後及跨系統整合挑戰而顯得力不從心。ArangoDB 透過其獨特的原生多模型(圖、文件、向量)能力,將碎片化的客戶資訊、交易行為與裝置指紋等數據整合,實現「關係+行為+模式識別」的一體化分析。這不僅能大幅提升詐欺識別速度和準確性,降低決策延遲與維運成本,更賦予金融機構主動防禦與全面理解風險的能力,開啟金融反詐欺的新篇章。

虹科 GNSS 模擬器透過軟體定義架構,在實驗室中精準重現城市峽谷、山區、農田等複雜場景,完整驗證巡檢無人機導航精度、多星座兼容性、抗干擾能力及 RTK 高精度定位與多感測器融合性能。告別高成本、低可重複性的外場測試,打造高效、可控、可追溯的 UAV 整機測試流程。