Naveen Kumar K V

#Integration and #Testing the complex system in Software Defined Vehicles (#SDV) context is the biggest challenge to be solved by OEMs and Software suppliers!

1

#ChatGPT is now 2 years old. When comparing with Manufacturing. It's AI 1.0 to AI 4.0. I see so many things changed in the past 2 years and more will come in next 2. What happened in last 2 years - ChatGPT launch: The meaning of AI is different. It’s how a product in this area changed how people see this area.How people process language are totally different. - More multi-modal models: AI is able to process video, audio, paint, generate images and video, #suno is able to create songs. - GPT-4 release: AI was improved so much in a short time. Everybody talks about scaling law all the time. The developer tools like #Cursor #Windsurf became something we use everyday. What happened in last 100 years in manufacturing - 1920s: Handcraft Era. - 1960s: Mass Production - Assembly lines and standardization - 1990s: Lean Manufacturing - Just-in-time and quality focus - 2020s: Smart Manufacturing - AI-driven optimization and predictive capabilities One of the biggest reason for not able to applying AI into manufacturing is the diversity of industry use cases and knowledge. But with model GenAI tech, it will be different. No technology is able to learn at that large scale and speed. So what will happen in next 2 years? I see huge opportunities for every part of a manufacturing company from #design #supplychain #factory #sales #service Design: Generative AI revolutionizing product development Supply Chain: Intelligent forecasting and risk mitigation Factory Operations: Autonomous decision-making systems Customer Experience: Personalized service at scale Will it change manufacturing like #ExpertSystem wanted to?

11

1 Comment

You’re driving down the highway - a smooth, seamless ride. Thousands of new features were added to the vehicle over the last few years, increasing safety and functionality. Obviously, your next thoughts are: “I wonder how they squeezed the vehicle through Jenkins”. “Is Kubernetes involved in testing vehicle functionality?” At GM, our ADAS software is fully tested with Jenkins pipelines using a shift-left approach. Every single PR (code submission) triggers a software build and testing on both SIL (software simulation) and HIL (hardware simulation). We have a continuous testing methodology, through which the ADAS components flow through a multi-tier approach so that each level of testing is handled where it needs to be. The components are packaged, integrated and released for calibration and vehicle testing. Where does Kubernetes come in? We originally were managing our HIL boards using from within Jenkins (Lockable Resources plugin). This became challenging as we have multiple Jenkins servers and wanted to have the board pool available to all of them. To handle this, we attached all our boards to Kubernetes as nodes and defined them so that only a single pod could run on each node concurrently.  We can then use the Kubernetes plugin to dynamically allocate the boards to whichever Jenkins server needs them. This also gives us the ability to use other Kubernetes features, such as standard deployment with helm charts, monitoring and health checks. We can also easily allocate boards to uses other than our CI/automation environment, when needed. This is actually a much-simplified description of our Kubernetes deployment, but I wanted to give a taste of how we’re using industry standard DevOps tooling as part of our vehicle software testing process. #iworkforgm #GMrs_lead

23

1 Comment

Layoffs are never easy, and losing great colleagues is always difficult. If you're going through your first layoff, here are some words to live by: - Do things well, no matter the job. You represent yourself and your family each day. - Be a great teammate. Nothing of significance is accomplished in a silo. - Maintain the relationships you build along the way. - Strive for hands-on experience and opportunity. Walk through the door that opens when the opportunity shows up. - Stop wasting the free time you have outside of work. Learn a new skill. Build your resume. - Always have a plan B. Companies will ALWAYS make money decisions. That includes headcount reduction. - Your employer doesn’t own you, they simply pay rent for your time. Remember, this is your life to own, not theirs.

138

5 Comments

Black box, white box, Software Fault Insertion, HSIT, System Integration, PIL, MIL, SIL & Back-to-back testing

3

Must see (or hear) episode of The Garage with Senior Vice President Vidya Rajagopalan from Rivian! 📣 If this is the first time you are tuning in - The Garage covers all topics around SDV with the most influential people in the industry. Listen in! 🔊

4

I have completed the "ChatGPT Prompt Engineering for Developers" course by Deeplearning.ai! This experience has enhanced my skills in crafting effective prompts for AI language models. Looking forward to applying this knowledge and exploring new opportunities in AI! #ChatGPT #PromptEngineering #AI #DeepLearning #ContinuousLearning

21

2 Comments

Here is a statement, which maybe a bit controversial. But let's just say it:- "If you throw enough money into a product, the product is bound to be successful, even if it is not the best product in the market."   After being in the automotive industry for almost 8 years, and primarily working on EV's for Indian OEM's to Indian Startups I have realized that success of a product has very little to do with its actual engineering. Its more to do with the story that you build around the product.   Recent example:- There have been so many incidences of people complaining about their recent scooter purchase. Varying from a spectrum of people complaining about the scooter on Social media to burning their showroom itself.   So this is not a new incident, the product itself has not been very great from the beginning, it had a lot of issues, but with enough money in their pockets, they were able to push it throughout India, without any dealership, which usually takes a lot of time. This inorganic growth and capturing of the market shows the real power of money. They tried making the most futuristic product possible, with all the premium features, and sold it at a price that just throws basic economics out of the picture, which make sense as a lot of startups who get VC funded and need to show growth, do the same. But this company was not funded like a normal company, it had unimaginable amount of investments, for which they needed to just decimate their competitors.   So their engineering timelines were really skewed, bought a company from Europe, reengineered its internal parts to get the famed FAME subsidy. Did the internal testing and announced its entry as the next big thing in the market. They hyped up the product with all cool features(which again not the most important thing in a scooter), the longest range, the best features and asks just for 500Rs to book the product. People booked it in huge numbers, and this was the first testament of their pseudo success.   But with delayed delivery of the product and delivering only 1 variant out of their 2 variants announced, people at the end were just happy to get their hands on it. Forget about the claimed range which was so hyped up, that it threw basic science out of the picture. I had so much difficulty in explaining this pseudo range claims when I myself was working on a similar product for a startup.   So ultimately people were still happy that for this price point this is the best thing in the market, even if they were not able to deliver half of the features they promised.   Now forget about the features, it was not able to deliver its basic functionality as a scooter, that’s when the whole drama started. I would still appreciate the resilience of Indian people that somehow they were still justifying their purchase that at least we are buying something Made in India.

37

3 Comments

With the recent layoffs in various Autonomous driving companies (latest was Motional, wish all friends there all the best in securing new opportunities), here’s how I read the market: It was the trend 3~4 years ago for OEMs & Tier suppliers to work and invest money on ADAS as they saw it the future for selling their cars/products, and they thought that creating a fully autonomous vehicle (L3+ or even Level 4) was just a small step of iterative improvement to the current Level 2 ADAS systems they had and maybe with burning more money and dragging in huge resources they would fastly achieve it. Here’s what they discovered: - The laws/standards for developing and launching such cars (at least in Europe is not ready) - Current technologies (Cameras, sensors) are good but putting them together and processing these huge data to act like a ‚human driver‘ is very expensive and complex to do. - Engineering automotive resources were also a little behind in Tech areas to quickly adapt to the fast paces of other Tech. silicon valley companies. - Robotaxis as a business model is very hard to achieve profitability taking into account that they need a huge fleet of cars (every car is equipped with 50+ K system + car price) running almost 24 hours with the addition of the need of Technical support/Remote guidance and regular maintenance (done by a human, like charging or even checking the tire pressure), with all these costs on top to compete with human taxi drivers or Uber like apps will not be easy. Accordingly the companies were split to 3 streams: 1. Huge Cash companies (such as Waymo supported by Google) who still burn money in this direction hoping it will dominate the market when other players are out of money. 2. companies who decided to slow the pace and investment to achieve reasonable targets (maybe not as fast as it was planned) 3. companies who gave up and shutdown (like my beloved Argo) Again this is a personal opinion and how I read the situation and hope the market of Automotive is going to heal soon and recover. #ADAS #AutonomousDriving

355

15 Comments

Automotive Bootloader Course on Udemy

28

3 Comments

Useful read on details of JTAG https://lnkd.in/giDhXyh4 Thanks Aliaksandr Kavalchuk and Memfault for this #JTAG

54

1 Comment

🚀 Dive Deep into FlexRay: The Future of Automotive Communication In the rapidly evolving automotive landscape, FlexRay stands out as a pivotal communication protocol designed to meet the demands of advanced vehicle systems. Here’s an in-depth look at what makes FlexRay a game-changer in automotive technology: 🔍 What is FlexRay? FlexRay is a high-speed, deterministic communication protocol developed to address the needs of modern automotive systems, especially those requiring high reliability and real-time performance. It’s designed to complement the Controller Area Network (CAN) and other protocols, providing a robust solution for complex applications. 📈 Key Features of FlexRay: - High-Speed Communication: FlexRay supports data transfer rates of up to 10 Mbps, significantly higher than traditional protocols like CAN. This high speed is essential for handling the increasing volume of data in modern vehicles. - Deterministic Performance: The protocol ensures deterministic behavior, which means it provides predictable and guaranteed response times. This is critical for safety systems and real-time applications where timing is crucial. - Redundancy and Fault Tolerance: FlexRay includes built-in redundancy with dual channels, enhancing fault tolerance and ensuring communication reliability even in the case of a channel failure. This feature is vital for safety-critical systems in automotive environments. - Time-Triggered and Event-Triggered Communication: FlexRay supports both time-triggered and event-triggered communication. Time-triggered communication provides precise timing for periodic messages, while event-triggered communication handles spontaneous data transmissions efficiently. - Scalability: The protocol is scalable, making it suitable for both high-end luxury vehicles and more basic models, adapting to various communication needs and system complexities. - Synchronization: FlexRay offers precise synchronization capabilities between nodes, which is essential for coordinating complex operations across multiple electronic control units (ECUs). 📊 Applications in Modern Vehicles: - Advanced Driver Assistance Systems (ADAS): FlexRay’s high speed and reliability are perfect for ADAS applications, including adaptive cruise control, lane-keeping assist, and automated parking systems. - Electronic Stability Control (ESC): The protocol’s deterministic performance ensures timely and accurate control in stability and safety systems. - Infotainment Systems:With its high bandwidth and low latency, FlexRay enhances the performance of sophisticated infotainment systems, providing seamless multimedia experiences. 🔧 Implementation Considerations: - Complexity: Implementing FlexRay requires specialized knowledge and hardware, which can increase system complexity and cost.

47

2 Comments

For Mid-Career Tech Professionals If you are a tech professional and have 15+ years under your belt, with a comfortable job and a nice 6-figure package; you may be at risk of layoff or cost cutting measures for a variety of reasons. 1. True cost cutting going on in the company because of poor financial performance 2. Re-alignment of company focusing on newer products or technologies making you redundant in an otherwise profitable company 3. Cheaper resources may be available with 2-3 years of experience or near shoring/off shoring could be a cause. With remote work options, this has really opened up. In all these scenarios, senior and experienced professionals may find themselves out of work. Once that happens, landing another job may become quite hard given the mid-life income needs. If it has happened, Good Luck to you as the job market is a little tight. However, this is for people who have not yet been laid-off to take notice and ensure they are ready by following some simple guidelines 1. Understand the focus and strategic areas of your company's growth. Make sure you understand what the leadership teams are working on. This information can be gathered from monthly management meetings, internal newsletters/website or corporate communication, CEO messages going to all employees, new executive hire announcements, etc. 2. Do not be complacent and comfortable in your position, make sure you are upskilling yourself regularly. If you're doing the exact same task for 3+ years, you're at risk - subscribe to industry newsletters and newsfeeds so you are fully aware of what is going on in similar companies 3. Do not price yourself too far out from the median income for your position. Gather information from various websites, forums and your own network about median income. That ensures you are not on chopping block for budgetary reasons 4. Continuously building and increasing your network so in case you do get laid-off, other potential hiring managers know about you The key is to make yourself indispensable for your employer by not being protective of your tasks rather demonstrating initiative, upskilling and superior communication.

56

2 Comments

Masterclass Done !!!! A satisfying revision of the Skillset!! #KPIT

36

1 Comment

A behavioral shift is required in the application of the heuristic RASIC framework. - I believe it is the responsibility of the entire development supply chain to understand each technology and their interface impacts involved in the solution. - Ownership would remain with the respective domain experts.

20

🚀 Unlocking the Future of Automotive Communication with MICROSAR QuickCom! 🚗🔧 As the automotive industry continues to evolve, seamless communication between components is crucial for modern vehicle systems. That's where MICROSAR QuickCom comes in, providing a fast, efficient way to design and implement communication models. With MICROSAR QuickCom, developers can easily: Define communication models using a flexible JSON configuration Generate Proxy/Skeleton APIs with a robust SDK Implement application logic in C++ Quickly build and run prototypes for service-oriented communication (SOC) This tool is perfect for rapid prototyping in automotive applications, ensuring faster deployment and real-world validation of communication strategies. MICROSAR QuickCom applications are designed to integrate seamlessly with the MICROSAR Adaptive Stack MICROSAR QuickCom is free for non-commercial use, including research and development. Embrace the future of automotive technology—where speed, reliability, and innovation converge! 🌐 #Vector #SoftwareDevelopment #Prototyping #MICROSARQuickCom #SDV #AdaptiveAUTOSAR

19

Rivian's alternative to CarPlay 1. Rivian is one of the few OEMs (Tesla, Lucid, GM) that are not offering CarPlay in it's vehicles 2. Rivian has introduced a new service called Connect+ which adds many streaming capabilities to the infotainment. Apple Music, Dolby Spatial Audio, Audible 3. The main addition is Apple Music. Other OEMs like Porsche and Mercedes Benz also have this. But the kicker is that Rivian has developed this together with Apple. This to me suggests that Apple has probably faced some backlash of it earlier strategy to take over the infotainment from OEMs. As expected OEMs are not that keen on this strategy. 4. In the new strategy, if Apple works together with OEMs to better integrate their services into infotainments, it is win-win for everyone. Better experience for users, better product for the OEM and Apple can still maintain it's presence in vehicle with or without CarPlay. 5. Trend: The modern car infotainment thrives on streaming capabilities. To read the complete briefing: https://lnkd.in/dYFv4CbW

381

62 Comments