Ricardo Leite

🚘 Decoding the CAN Frame: What’s Inside a Standard Message❓️🤔 ✅️The Controller Area Network (CAN) is the backbone of modern automotive communication. But what actually goes into a CAN frame? ✅️From Start of Frame (SOF) to Intermission (ITM), each field has a critical role in ensuring reliable data exchange between vehicle ECUs.

352

4 Kommentare

Why is the CAN Protocol Essential in Automotive Engineering? In modern vehicles, seamless communication between hundreds of Electronic Control Units (ECUs) is critical. That’s where the Controller Area Network (CAN) protocol comes in. Originally developed by Bosch, the CAN protocol allows microcontrollers and devices to communicate with each other without a host computer, making it ideal for real-time, robust, and efficient communication across vehicle systems. Why is CAN so vital? Reliability: It ensures fault-tolerant data transmission, even in harsh environments. Efficiency: Reduces wiring complexity, minimizing vehicle weight and cost. Real-Time Performance: Crucial for time-sensitive operations like engine control, braking, and airbag deployment. Scalability: Easily integrates new ECUs and features without overhauling the entire system. As vehicles become smarter and more connected, the CAN protocol remains a cornerstone of in-vehicle communication, paving the way for future mobility innovations. #AutomotiveEngineering #CANProtocol #VehicleNetworking #MobilityTech #EngineeringInnovation

168

1 Kommentar

🚘 Decoding the CAN Frame: What’s Inside a Standard Message❓️🤔 ✅️The Controller Area Network (CAN) is the backbone of modern automotive communication. But what actually goes into a CAN frame? ✅️From Start of Frame (SOF) to Intermission (ITM), each field has a critical role in ensuring reliable data exchange between vehicle ECUs.

296

Last time, I compared std::expected<T, E> in C++ with Result<T, E> in Rust. The takeaway was that modern C++ no longer needs 𝚎𝚡𝚌𝚎𝚙𝚝𝚒𝚘𝚗𝚜 for structured error handling. This post completes that idea. When using std::expected, instead of immediately branching on the returned value, we can treat it as a pipeline and use its 𝗺𝗼𝗻𝗮𝗱𝗶𝗰 operators to make the flow clear. This simply means we can chain computations that can fail and propagate errors back. In the example below, the flow is linear and fluent: ⭢ Load data from a file ⭢ Transform the value (lower-case it) if the file exist ⭢ Process the data if everything succeeded ⭢ Log an error if anything failed All without try/catch, flags, or nested if statements. ━━━━━━━━━━ 🎆 𝐌𝐲 𝟐𝟎𝟐𝟔 𝐅𝐨𝐜𝐮𝐬 This year, I’ll be sharing content on: - Modern C++ & Rust - Software architecture & design - Automotive software development - Safety-critical systems - Book reviews on software engineering and tech topics ♻️ If you’re interested in any of these areas, follow me to stay updated! ━━━━━━━━━━ #CppToRustJourney #ModernCpp #RustLang #SoftwareEngineering #ErrorHandling #Cpp23 #SystemsProgramming #EmbeddedSoftware #SafetyCritical #CleanCode #ProgrammingBestPractices

121

62 Kommentare

💻 To my software engineers, developers, testers, and anyone looking to get into or transition into tech/software. Listen very carefully and mindfully, 🧏🏾‍♀️ You cannot know **every single programming language** AND be great at building with every single one of them. Read that again! 👆🏾👆🏾👆🏾 __________ Take these practical steps instead: 🎯 Get clear on what domain you want to get into or get better at —> whether that be mobile, web, cross platform, cloud, security, infra, data, devops, ai/ml…. (there’s a ton!). 🎯 Pick 2 - 3 (max) languages and get really good at them —> ideally, ones that are versatile and have a wide range of use cases and applications in the industry. Invest your time and effort wisely. 🎯 Understand why that specific language is used and considered industry standard for its relevant domain —> like its pros, cons, and trade offs. 🎯 Some questions to ask yourself —> Does it benefit me in the long run to learn this language? What if I want to switch domains in the future, would this knowledge be easily applicable/transferrable? Does learning this language keep my skills relevant in the industry I’m in? 🎯 Lastly, break stuff and build for fun! It’ll stick and you’ll learn better that way. __________ Knowing every programming language to exist doesn’t make you better developer/engineer. and …. …. stop spreading yourself tooo thin. You still have that issue to debug! 😉🤭 __________ P.s. share this with an engineer who might be feeling overwhelmed with the sheer amount of new language(s) they *think* they (still) need to learn! 😭 ✨Follow Chikodili for more content on career growth and impact ✨

50

5 Kommentare

🤩 Driving automation and autonomous mobility is one of the most interesting topics in automotive, here is an update. 👉 Please add your comments below and reshare with your network. It’s just my personal opinion. ADAS and AD in China 🇨🇳 at crossroads: 👉 GAC MOTOR and DiDi shared an update on their joint AD project. The mass production of the L4 vehicle, a GAC vehicle with the DiDi tech stack is targeted for end of 2025: https://lnkd.in/dZgeZczp 👉 The Ministry of Industry and Information Technology (MIIT) in China tightened ADAS regulations, public beta testing programs are prohibited or marketing terminology for ADAS is restricted. Bad news for Tesla and some of the domestic OEMs: https://lnkd.in/dYDpqjvx 👉 Volkswagen Group China shares an update on the in China 🇨🇳 for China developed ADAS system. Developed with Horizon Robotics and CARIAD in 18 months, capable of L2+ features on highways and in urban areas, launch in 2025: https://lnkd.in/dmKTtnjD 👉 GEELY launched the G-Pilot ADAS portfolio to respond to the overall demand and the God‘s View program of BYD. ZEEKR极氪 will be the first brand deploying the L3-ready G-Pilot H9 performance package with the NVIDIA Thor chip (probably the first OEM with the next SOC generation 🤷‍♂️): https://lnkd.in/dQcfgFMW 👉 Follow the Shanghai Auto closely next week, ADAS solutions might take the center stage. Roads to Japan 🇯🇵: 👉 Waymo shared more details about the data collection project in Japan. 25 vehicles are used to gather driving and mapping data. Nihon Kotsu is responsible for a safe operations based on manual drivers: https://lnkd.in/d4sxKp74 👉 Wayve closed a deal with Nissan Motor Corporation. The next ProPILOT systems based on Wayve software should launch in 2027: https://lnkd.in/dX6E8C_R 👉 Nuro, which raised a $106 million funding round recently is heading to Japan 🇯🇵 as well. It’s not sure if Nuro has a deal with a local OEM or just want to get on the radar of Japanese OEMs like Toyota Motor Corporation, Honda or Nissan. Seems that investors buy the pivot from L4 delivery 🚚 to ADAS: https://lnkd.in/dxpY9UtH What else is happening 🤷‍♂️? 👉 Kodiak, an AD trucking company is going public 🤑 via a SPAC merger. I leave it to you to decide if the business is robust enough and if ghe $2.5 billion valuation is fair. Read through the deck here: https://lnkd.in/dt9mzqY4 👉 Valeo and Mobileye will deliver a L2+ system to Volkswagen Group. The popular MQB platform models should be equipped with the system: https://lnkd.in/dTPegFej What is your favorite story of the last weeks 🤷‍♂️? Please add below. #ADAS #AD #Mobility

376

24 Kommentare

🚗 DTCs in Automotive Diagnostics — What, Why, and How If you work with automotive embedded software, you’ve definitely met DTCs (Diagnostic Trouble Codes). But in real projects, DTCs are much more than error numbers — they are a contract between software, hardware, testers, workshops, and even legislation. Let’s break it down clearly 👇 🔍 What is a DTC? A DTC (Diagnostic Trouble Code) represents a detected fault related to: - Sensors or actuators - Communication (CAN, LIN, Ethernet) - Internal ECU software or hardware - Safety, emissions, or drivability functions Each DTC has: - A unique identifier - A status byte (bit-coded life cycle) - Optional extended data (snapshots, counters, timestamps) Standards typically come from AUTOSAR and ISO 14229 (UDS). 🎯 Why DTCs Are Critical in Real Projects In production vehicles, DTCs are used to: ✔ Enable on-board diagnostics (OBD) ✔ Support workshop troubleshooting ✔ Fulfill legal and emission requirements ✔ Trigger limp-home or degradation modes ✔ Provide data for field analysis & recalls 👉 A wrongly implemented DTC can cause: - False warning lamps - Missed safety issues - Rejected vehicle homologation - Costly customer complaints ⚙️ How DTCs Are Used in Practice (Real ECU Flow) 1) Monitoring Software continuously checks conditions (timeouts, range checks, CRCs, plausibility). 2) Detection If a fault condition is fulfilled → diagnostic event fails. 3) Debouncing / Qualification Counters, time filters, or operation cycles avoid false positives. 4) DTC Storage DTC is stored with a status byte update and optional snapshot data. 5) Reporting Tester reads it via UDS services (e.g. 0x19 ReadDTCInformation). 6) Healing / Aging If the fault disappears for defined cycles → DTC may heal or age. 🧠 The Heart of a DTC: Status Byte The status byte tells the story of the fault, not just that it exists. Bit 0 TestFailed : Fault is currently present Bit 1 TestFailedThisOperationCycle: Fault occurred in the current driving cycle Bit 2 PendingDTC: Fault detected but not yet confirmed Bit 3 ConfirmedDTC: Fault confirmed after qualification Bit 4 TestNotCompletedSinceLastClear: Test hasn’t run since last clear Bit 5 TestFailedSinceLastClear: Fault happened at least once since clear Bit 6 TestNotCompletedThisOperationCycle: Test didn’t run in current cycle Bit 7 WarningIndicatorRequested: MIL / warning lamp requested 📌 Key takeaway: A DTC is not binary (present / not present). Its status byte represents its life cycle. 🏁 Final Thought A good diagnostic implementation doesn’t just detect faults — it tells a reliable story about them. If your DTCs are clean, consistent, and well-qualified: - Testing becomes easier - Workshops trust the ECU - Customers trust the brand 💬 How do you usually implement DTC qualification? Counters or time-based? Let’s exchange experiences 👇

234

6 Kommentare

Why GitHub is Crucial in the Automotive Industry 1. Open Collaboration & Code Sharing Modern vehicle systems are no longer built in silos. OEMs (like Ford, Volvo) and Tier-1 suppliers (like Bosch, Continental) increasingly use open platforms (AOSP, Linux, QNX) to shorten time-to-market. GitHub becomes the bridge: For sharing reusable modules (like Vehicle HALs) For collaborating on standards (e.g., GENIVI, COVESA) For co-developing middleware and tools Example: Google publishes AOSP source and Car Service code on GitHub — many OEMs fork, extend, and contribute back. 2. Transparent Version Control for Complex Projects A modern infotainment stack has: Kernel drivers HALs and middleware System services App layer UI frameworks GitHub helps manage all this complexity with: Branching for features and platform versions (e.g., Android 11 vs. 13) Pull requests for safe code reviews Tags/releases for OTA update management Issue tracking for bugs in production or test fleets 3. Rapid Prototyping with Open Hardware (e.g., Raspberry Pi) When engineers like you build infotainment systems on Raspberry Pi using AOSP, GitHub becomes a place to: Share boot configs, init scripts, SELinux policies Upload HAL samples or AIDL interface code Document build & debug steps for the community This helps others replicate your work, collaborate, or even improve it — especially valuable for small startups or university R&D labs. 4. Reputation & Visibility for Engineers In automotive, roles like IVI System Engineer, Android Automotive Developer, Platform Integrator are in high demand — and GitHub shows: You can write production-level code You understand build systems, CI/CD, board bring-up You contribute to a community, not just work in private Recruiters and tech leads often scan GitHub profiles to verify your skills, not just your resume. 5. Compliance, Licensing, and Security Audits Auto industry has strict safety, security, and legal requirements: GitHub helps document code origin, licenses (GPL, Apache) Track security vulnerabilities (CVEs) in dependencies Collaborate securely on OTA patch releases Some companies even run GitHub Actions or CI pipelines to auto-test builds on embedded boards or in QEMU emulators. You’re not just coding — you’re building a portfolio that speaks for itself. #AndroidAutomotive #AOSPSystemEngineering #TechPortfolio #SELinux #VehicleUI #AutomotiveInnovation #CarInfotainment #EmbeddedSystems #AutomotiveSoftware #RaspberryPi #HALDevelopment #IVI #OpenSource #GitHubProjects #LearningInPublic #DevCommunity #ShowYourWork #MadeWithAndroid

9

𝗢𝗽𝗲𝗻 𝗦𝗼𝘂𝗿𝗰𝗲 & 𝗟𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆: 𝗧𝗵𝗲 𝗨𝗻𝗰𝗼𝗺𝗳𝗼𝗿𝘁𝗮𝗯𝗹𝗲 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻 𝗶𝗻 𝗔𝘂𝘁𝗼𝗺𝗼𝘁𝗶𝘃𝗲 (𝘗𝘦𝘳𝘴𝘰𝘯𝘢𝘭 𝘷𝘪𝘦𝘸; 𝘰𝘱𝘪𝘯𝘪𝘰𝘯𝘴 𝘢𝘳𝘦 𝘮𝘺 𝘰𝘸𝘯) Open-source software (OSS) already represents 𝘂𝗽 𝘁𝗼 𝟳𝟬% 𝗼𝗳 𝘁𝗵𝗲 𝗰𝗼𝗱𝗲 𝗶𝗻 𝗺𝗼𝗱𝗲𝗿𝗻 𝘃𝗲𝗵𝗶𝗰𝗹𝗲𝘀. The trend is irreversible - driven by the need for faster innovation, transparency, and reduced vendor lock-in. But as OSS moves from the toolchain to the in-vehicle stack, it raises an uncomfortable question: When a component with no warranty contributes to a failure, who is liable? This tension will define the next chapter of the Software-Defined Vehicle. 𝟭. 𝗧𝗿𝗮𝗻𝘀𝗽𝗮𝗿𝗲𝗻𝗰𝘆 𝗶𝘀 𝘁𝗵𝗲 𝗡𝗲𝘄 𝗦𝘁𝗮𝗻𝗱𝗮𝗿𝗱 Safety doesn’t come from secrecy - it comes from process, evidence, and traceability. When code is open, quality can be verified, security audited, and interoperability improved by design. Initiatives like Automotive Grade Linux (AGL) and the Eclipse SDV Working Group already prove that openness and safety are not mutually exclusive. 𝟮. 𝗧𝗵𝗲 𝗜𝗻𝗲𝘀𝗰𝗮𝗽𝗮𝗯𝗹𝗲 𝗟𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻 Most OSS licenses explicitly disclaim warranties - a stark contrast to a regulated industry where accountability is non-negotiable. The real challenge isn’t if we should use OSS, but how we build a framework of responsibility around it. 𝗪𝗵𝗼 𝗶𝘀 𝗮𝗰𝗰𝗼𝘂𝗻𝘁𝗮𝗯𝗹𝗲? The OEM? The Tier-1 integrator? The community? 𝟯. 𝗧𝘄𝗼 𝗠𝗼𝗱𝗲𝗹𝘀 𝗪𝗶𝗹𝗹 𝗘𝗺𝗲𝗿𝗴𝗲 We’re heading toward a hybrid future: 𝗢𝗘𝗠-𝗠𝗮𝗶𝗻𝘁𝗮𝗶𝗻𝗲𝗱 𝗢𝗦𝗦: Automakers will own, qualify, and maintain critical parts of the open-source stack internally. 𝗖𝗼𝗺𝗺𝗲𝗿𝗰𝗶𝗮𝗹 𝗗𝗶𝘀𝘁𝗿𝗶𝗯𝘂𝘁𝗶𝗼𝗻𝘀: Partners - much like Red Hat for enterprise Linux—will provide validated OSS stacks with long-term support, security patches, and, crucially, contractual liability. 𝟰. 𝗜𝗻𝗶𝘁𝗶𝗮𝘁𝗶𝘃𝗲𝘀 𝗣𝗮𝘃𝗶𝗻𝗴 𝘁𝗵𝗲 𝗪𝗮𝘆 The Eclipse SDV ecosystem is already shaping this accountable future with projects like:  • 𝗘𝗰𝗹𝗶𝗽𝘀𝗲 𝗦-𝗖𝗼𝗿𝗲 – a reference architecture for certifiable, modular vehicle software.  • 𝗘𝗰𝗹𝗶𝗽𝘀𝗲 𝗢𝗽𝗲𝗻𝗦𝗢𝗩𝗗 – an open standard for vehicle diagnostics and off-board interaction.  • 𝗘𝗰𝗹𝗶𝗽𝘀𝗲 𝗢𝗽𝗲𝗻𝗕𝗦𝗪 – open implementations for essential Base Software modules, defining the next frontier of in-vehicle OSS. 𝗧𝗵𝗲 𝗙𝘂𝘁𝘂𝗿𝗲 𝗶𝘀 𝗢𝗽𝗲𝗻 - 𝗕𝘂𝘁 𝗥𝗲𝘀𝗽𝗼𝗻𝘀𝗶𝗯𝗹𝗲 The shift to open source is unstoppable. The challenge now is to make it 𝘀𝗮𝗳𝗲, 𝗺𝗮𝗶𝗻𝘁𝗮𝗶𝗻𝗮𝗯𝗹𝗲, 𝗮𝗻𝗱 𝗹𝗲𝗴𝗮𝗹𝗹𝘆 𝘀𝘂𝘀𝘁𝗮𝗶𝗻𝗮𝗯𝗹𝗲. 𝗧𝗵𝗲 𝗳𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗮𝘂𝘁𝗼𝗺𝗼𝘁𝗶𝘃𝗲 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝘄𝗶𝗹𝗹 𝗯𝗲 𝗼𝗽𝗲𝗻 -  𝘄𝗶𝘁𝗵 𝗹𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗱𝗲𝘀𝗶𝗴𝗻𝗲𝗱 𝗶𝗻, 𝗻𝗼𝘁 𝗯𝗼𝗹𝘁𝗲𝗱 𝗼𝗻. What’s your view? How is your organization preparing for this shift? #Automotive #SoftwareDefinedVehicle #OpenSource #AutomotiveIndustry #Liability #TechLeadership #EmbeddedSystems #EclipseSDV

59

26 Kommentare

📚 Prelaunch Review: C++ Memory Management by Patrice Roy 🚀 I had the privilege of reviewing this book before its official launch, and I must say—it’s a must-read for C++ developers looking to master memory management! 🔹 Memory management is one of the most critical aspects of C++ programming, and this book provides a well-structured, in-depth guide that balances theory and practical applications perfectly. ✅ Why You Should Read This Book: ✔️ Covers object lifetime, alignment, smart pointers, and custom memory allocators. ✔️ Focuses on modern C++ (C++20, C++23) best practices. ✔️ Discusses arena-based memory management, custom STL-style containers, and memory pools. ✔️ Practical examples make it highly applicable to real-world systems like embedded, game dev, and high-performance computing. 📌 Who Should Read It? 🔹 Experienced C++ Developers – Looking to refine their understanding of memory optimization. 🔹 Embedded & Systems Programmers – Working with real-time, constrained-memory environments. 🔹 Game & HPC Developers – Handling large-scale, low-latency memory operations. 🔹 Software Architects – Preparing for architect-level roles requiring deep system knowledge. ⚠️ Areas for Improvement: Some introductory chapters may be too basic for expert-level readers. Would benefit from more real-world case studies on system-level memory optimization. ⭐ Final Verdict: Highly Recommended! If you are serious about writing safer, faster, and more efficient C++ code, this book is a must-read! 🎯 📖 More details & purchase link: https://lnkd.in/didw54ab 🔖 #C++ #MemoryManagement #SoftwareEngineering #EmbeddedSystems #HighPerformanceComputing #BookReview #CppDevelopers #Programming

64

1 Kommentar

How to apply The 4 Key Metrics in a highly regulated, hardware-centric environment like automotive software development? This article - mainly written by my dear colleague Folker Bernitt with some contributions from myself - sheds some light on how you can introduce the Four Key Metrics for automotive / embedded software and reduce feedback cycle time. In a gist: Apply the 4 Key Metrics to the current thing that comes closest to your production environment, ie. 1️⃣ Change Lead Time becomes the time required to validate changes in available QA environments (e.g., Software in the Loop (SiL), Hardware in the Loop (HiL)). 2️⃣ Deployment Frequency measures how often teams verify and validate their software in the best available environment that approximates production. 3️⃣ Change Failure Rate highlights the percentage of failed deployments (incl. failed validation and verification in preceding environments) 4️⃣ Mean Time to Recovery tracks how quickly teams address issues identified in downstream validation. Although each of these metrics lead to a "local optimization" in the first place - if you apply it consequently and optimize each stage - you eventually improve the org-wide delivery performance. This article took us quite some time to get written and a big Thank You! goes out to all our proof readers Michael Fait, Sven de Ridder, Sudhanshu Garg, Manu Eluvungal Viswambharan, Deng Shufang, Nikolas Buckstegen https://lnkd.in/eNP9hJJG

30

1 Kommentar

🚀 𝗪𝗵𝗮𝘁 𝗶𝘀 𝗠𝗼𝗱𝗲𝗹-𝗕𝗮𝘀𝗲𝗱 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 (𝗠𝗕𝗗)? And why is it a game-changer in Automotive Software Engineering? 🚗 As vehicles become more intelligent—with ADAS, Powertrain, and autonomous driving capabilities—the complexity of embedded software has increased drastically. Traditional development methods just can’t keep up. That’s where Model-Based Development (MBD) steps in. 💡MBD is a design methodology that uses graphical models to simulate, test, and auto-generate code for embedded systems—before writing a single line of manual code.   ✅ 𝗪𝗵𝘆 𝗠𝗕𝗗 𝗠𝗮𝘁𝘁𝗲𝗿𝘀 𝗶𝗻 𝗔𝘂𝘁𝗼𝗺𝗼𝘁𝗶𝘃𝗲 ? 👉 𝗙𝗮𝘀𝘁𝗲𝗿 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 – Iterate rapidly in early stages of the V-cycle 👉 𝗘𝗮𝗿𝗹𝘆 𝗩𝗮𝗹𝗶𝗱𝗮𝘁𝗶𝗼𝗻 – Simulate scenarios before hardware is even available 👉 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗰 𝗖𝗼𝗱𝗲 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 – Reduce manual coding errors using tools like Embedded coder and TargetLink 👉 𝗧𝗿𝗮𝗰𝗲𝗮𝗯𝗶𝗹𝗶𝘁𝘆 – Align with ISO 26262 and ASPICE processes 👉 𝗥𝗲𝘂𝘀𝗲 – Build modular components for easy reuse across vehicle platforms 🚘 𝗥𝗲𝗮𝗹-𝗪𝗼𝗿𝗹𝗱 𝗘𝘅𝗮𝗺𝗽𝗹𝗲: 𝗣𝗼𝘄𝗲𝗿𝘁𝗿𝗮𝗶𝗻 & 𝗕𝗼𝗱𝘆 𝗖𝗼𝗻𝘁𝗿𝗼𝗹 𝗠𝗼𝗱𝘂𝗹𝗲𝘀 In one of my recent projects, we applied MBD for both Powertrain and Body Control Modules (BCM)—handling features like engine start logic, torque management, and lighting, wiper, and door functionalities. Using MATLAB/Simulink and Stateflow, we were able to: 👉 Design and simulate multiple engine states (e.g., ON/OFF, limp-home mode) 👉 Validate user scenarios like wiper operation during ignition or door unlock timing 👉 Generate production-ready embedded C code via Target Link, with traceability to system requirements   💡 𝗪𝗵𝗮𝘁 𝘄𝗲 𝗮𝗰𝗵𝗶𝗲𝘃𝗲𝗱: ✔️ Cut integration testing time by 40% ✔️ Identified logic flaws pre-HIL ✔️ Ensured compliance with ISO 26262 and OEM coding guidelines—well before touching the hardware   👨‍💻 𝗠𝘆 𝗧𝗮𝗸𝗲 Model-Based Development isn’t just another tool—it’s a strategic shift in how we approach embedded system design. It brings clarity, speed, and safety to a domain where these qualities are non-negotiable. Are you or your team using MBD yet? Let’s connect and exchange ideas! 🚀   🔖 #ModelBasedDevelopment #AutomotiveSoftware #Simulink #EmbeddedSystems #ISO26262 #Powertrain #BCM #MATLAB #Stateflow #TargetLink #AUTOSAR #FunctionalSafety #ControlSystems #LinkedInEngineering #SoftwareDevelopment

96

3 Kommentare

What SDV really means for Tier-1s – and how Bosch, ZF, Continental are positioning? Software-defined vehicles (SDV) fundamentally change what OEMs expect from Tier-1s. The shift is not incremental. It is structural. OEM expectation trajectory is becoming clear: Short term: Tier-1s absorb risk, deliver SOP-ready systems, and provide safety credibility (especially by-wire). Mid term: OEMs reclaim vehicle behavior ownership. Tier-1s are expected to provide fail-operational execution layers, safety kernels, and integration support, not vehicle “DNA.” Long term: Chassis control becomes a software platform problem. OEMs multi-source hardware and monetize behavior via OTA. Tier-1s win only where ASIL-D execution, redundancy, and tooling remain hard to internalize. Against this backdrop, the strategies of Bosch, ZF, and Continental start to separate clearly. Bosch appears most aligned with the OEM endgame. Its positioning increasingly emphasizes fail-operational by-wire, safety-certified execution, and vehicle-level motion middleware that enables—but does not define—OEM behavior. Co-development models (e.g., OEM software alliances) and strong investment in tooling and validation suggest Bosch accepts that behavior belongs to OEMs, while Tier-1 value shifts downward into safety and execution. ZF sits in a strong but more ambiguous position. With integrated chassis and centralized VMC platforms, ZF is technically well prepared for SDV. The open question is strategic clarity: will ZF fully embrace an OEM-owned behavior / Tier-1 execution split, or continue to hedge with “integrated chassis intelligence” narratives? The former aligns long-term; the latter risks friction as OEM software maturity grows. Continental shows convincing short- and mid-term strength through cross-domain compute, holistic motion control, and cloud-based SDV workflows. The long-term challenge is scale and standardization: success depends on whether these platforms become repeatable, OEM-embedded building blocks, rather than isolated showcases. Otherwise, chassis control risks sliding toward commoditized hardware plus basic safety software. Bottom line: SDV does not eliminate Tier-1 value—but it redefines it. Short term: system delivery wins. Mid term: execution + integration wins. Long term: fail-operational safety, redundancy, and tooling win. Tier-1s that align early with this reality will remain indispensable. Those that defend behavioral ownership will struggle—no matter how good the hardware is. Based on publicly available signals, this is my interpretation—happy to discuss alternative views.

75

🚗 Understanding “Vanilla AOSP” vs “Qualcomm-Tailored AOSP” and why RB3 Gen2 isn’t supported by default If you’ve ever tried to build AOSP for Snapdragon boards (like RB3 Gen2, RB5, or SA8155) and found it doesn’t boot — here’s why 👇 ⚙️ 1️⃣ Vanilla AOSP (Google’s Pure Version) “Vanilla” AOSP is the clean, open-source Android code released by Google on android.googlesource.com . It contains: 1. Android Framework + System Core 2. Generic HAL interfaces (no vendor code) 3. Build targets for emulators only: aosp_arm64, aosp_x86_64, etc. 4. No BSP (Board Support Package), no device trees, no proprietary binaries ✅ Great for education, research, and emulator builds ❌ Not ready for real hardware (Snapdragon, NXP, Renesas, etc.) ⚙️ 2️⃣ Qualcomm-Tailored AOSP (BSP Enhanced) Qualcomm takes Google’s AOSP base and extends it with: 1. Device-specific folders → device/qcom/<platform> 2. Vendor drivers, HALs, and proprietary binaries → vendor/qcom/ 3. Custom kernel sources (e.g., kernel/msm-5.4) 4. Hardware bring-up scripts, init configs, and SELinux rules This “Qualcomm-flavored AOSP” is distributed through: 1. CodeAurora Forum (CAF) or Qualcomm Innovation Center (QuIC) repos 2. OEM/NDA partnerships for Snapdragon automotive or IoT platforms ✅ Runs on real Snapdragon boards (RB3, RB5, SA8155, etc.) ❌ Not publicly available in Google’s mainline AOSP tree 🧩 3️⃣ Why RB3 Gen2 Isn’t in Vanilla AOSP Because RB3 Gen2’s BSP and vendor blobs are Qualcomm proprietary, Google’s AOSP doesn’t include them. That’s why you’ll find no “rb3gen2” target under device/ or vendor/ in a vanilla checkout. You need the Qualcomm BSP to add: 1. Device tree 2. Kernel source 3. Vendor folder 4. Proprietary binaries Only then can you build Android (AAOS, AOSP) for RB3 Gen2 successfully. 💡 Summary 🔹 Vanilla AOSP = pure open-source Android (no hardware support) 🔹 Qualcomm AOSP = AOSP + BSP + vendor folder for Snapdragon SoCs 🔹 To boot on RB3 Gen2 → use Qualcomm’s tailored AOSP, not vanilla AOSP 💬 Have you ever worked on porting AOSP or AAOS to real hardware (Snapdragon, Renesas, NXP, etc.)? What was your biggest bring-up challenge? I’d love to hear your experience 👇 #AndroidAutomotive #AOSP #Qualcomm #EmbeddedLinux #AAOS #AndroidDevelopment #Infotainment #AutomotiveSoftware #AndroidInternals #QNX #VolvoCars

134

8 Kommentare