Hello and welcome. I'm Sankar Subramanian and I'm glad we can come together online to explore today's topic. I have over 20 years of experience in control and automation systems, bringing together hardware and software expertise in area such as semiconductor lasers, fiber optic transfers, and electric vehicle RMD and manufacturing. My work has often been about integrating precise engineering with intelligent, reliable automation. Over past few years, I've been deeply involved in exploring how modern AI can be combined with robust programming tools to solve real world industrial challenges. Rust, with its focus on safety, speed and concurrency, has quickly become a powerful choice of building high performance and reliable systems. In the system, we will be looking look at how AI powered Mission, mission in Rust can deliver not only accuracy and efficiency, but also the scalability and suitability making it strong option for next generation automation. For next few minutes, I will share key concepts, real world examples, key takeaways you can apply to your own work. Let's get started. AI power mission interest, transforming ev, manufacturing quality and automation. The imperative of advanced division system in EV manufacturing, electrical. Electric vehicle represents a fundamental shift in automotive design on manufacturing complexity due to its higher component density requiring multi micron level precision critical battery safety concerns and necessary a hundred percentage inspection. Manufacturing at scale demands foster throughput, zero defect expectations for safety critical systems. So all this, because of all this high demands in the manufacturing traditional inspection methods can't meet these demands. Creating perfect application for rust power automation mission systems that combines performance with the memory safety guarantees that why rest the characters that are mentioned above. Makes arrest uniquely suited for EV manufacturing where reliability, speed and safety cannot be compromised. The first characteristics you know, that we can consider is memory safety, zero memory related crash in a production system through compile time guarantees critical for 20 4K seven manufacturing environments where downtime cost thousands per minute. Fearless currencies, safety process model. Safely processing multiple camera feeds simultaneously without, data race a race conditioning without any race conditioning, enabling real time inspection across multiple assembly stages and rest performing performance is very close to c. Speeds with the zero cost abstraction. Allow processing of higher resolution images at the line speed exceeding 10 ports per second throughput FFI capabilities. Seamless integration with cc plus vision libraries and emission learning framework like 10 flow and by chart while maintaining safety boundaries. The real time inline quality inspection systems, the higher resolution component inspection that is quite often in the manufacturing environment. So Rust Vision System currently deployed in production, environmental use, rust resolution down to five micron for critical cell inspections. Processing speed of 12 to 15 frames per second at a 4K resolution, arrow camera, image resolution multi spectrum imaging, like visible ir UV for comprehensive defect detections, consistent performance under variable lighting conditions. These are all becoming, very essential. And our multi-camera array leverage rust. That's concurrent con, concurrency model to process multiple inspection angle simultaneously without frame drops. The next presentation is next slide, which is AI enhanced to defect detection which is part of the inspection systems, mission, vision, inspection systems, and in the EV manufacturing. So image acquisition, which is a primary part. That has high speed camera capture capturing multiple views of components at line speed, like approximately like a 10 plus parts per second, which is significantly higher rate in the modern manufacturing. Pre-processing rus, SAMD optimization enables a realtime noise reduction in our contrast enhancements and then feature extractions. AI based classification, custom neural network detects subtle defect missed by traditional threshold based systems. And then automated responses, part rejection rework, flow flagging, or process adjustments based on a defect classifications. Our system have demonstrated a 37 percentage increase in defect detection rate compared to conventional machine vision approaches. While rest, safety guarantees guarantees prevent the system crash common in a CC plus N environments. And the next thing is battery cell on a pack inspection. On this electrode alignment verification is one of the critical process in, in, in the battery assembly. Battery pack assembly, basically. So it means assembling a cell into a pack. Submillimeter precision measurements or demand in this process. For aligning for ano catho alignment using structured light patterns. And rust high performance image processing is vital here. Surface defect detection, all a models identify microscopic cracks, dents, and contaminations. That could lead to a thermal runaway events detecting such, such a defect or critical in the battery cell and pack inspections electrolyte filling verification, and I are imaging combining combined with the custom d, detection algorithm, ensuring proper electrolyte levels and distributions. Thermal anomaly monitoring. On the back level. The IR cameras like FLIR, which is a teledyne. Now the IR camera integrates for early detection of hotspots during formation cycling with rust, handling a real time thermal mapping, advanced inspection methodologies. Battery module level inspection is one of the critical process in the manufacturing. Our rust based system implementation implements several cutting edge techniques like wire bond quality assessments using structured elimination laser weld quality verification through structural analysis, thermal imaging. During module pulsing to identify resistance anomalies on the welding defects 3D profile mapping to verify dimensional accuracy post assembly. So Russ performance enable a realtime processing of these complex inspections routines without dropping a frame. Lesser build inspection requires a precise lighting under multiple specialized camera to identify microscopic defects. 3D mission systems for assembly automation. When it comes for assembly automation. The challenges are quite significant, especially how precisely we are able to detect the depth of multiple components in a 3D perspective, which plays a vital role. So depth sensing, time of flight cameras provides realtime spatial data for robot guidance with two millimeter accuracy. And then the other type could be a dual camera setup with rust optimized disparity mapping for high precision and component handling. We do use laser triangulation. Basically a laser based sensors that, that could measure up to sub micron precision for critical alignment tasks like, motor, straighter assembly, those kind of processes. And then point cloud processing rust performance enables a realtime processing of dense point clouds for complex assembly operations. So once the assembly and then is done, the AI power division analytics, that plays a vital role in in any. High advanced manufacturing environment. Deep planning, integration, I know with our rust system it leverages multiple AI approaches for neural networks for defect classification with a very high accuracy, close to a hundred percent unsupervised anomaly detection, identifying novel defect types, transfer learning techniques to reduce training data requirements by 80% age. Russ FFI capabilities seamlessly connecting to TensorFlow and then OnX runtime interference. Optimizing through Russ enables the deployment of H devices directly on the protection line, eliminating network latencies. Then it comes to manufacturing execution systems integration. In MES in the manufacturing environment, component traceability is one of the, prime role of having ME. So OCR and the barcode reading are the primary on the critical process of traceability implementation. So OCR and barcode reading capability tracks every component from raw material to finished vehicles with 99 percentage or even close to a hundred percentage. Read. Process verification. So vision system confirms the correct assembly sequences and component placements, and on a day and integrate with the digital work instructions. So every critical process was implemented with this vision based assembly sequence, verifications process step. So quality data aggregation. All this data, there's a tons of data that are being collected. All these data are getting aggregated into a one digital form and then that will be tagged to every component that is being manufactured in this line. Regulatory compliance compliances. It also helps to automate documentation generation for safety and certification and recall readiness. Rust. Reliability ensures these mission critical systems maintain uptime exceeding close to a hundred percentage essential for maintaining production flow and traceability requirements. So once the vehicle is assembled, when it comes for advanced driver assistance system calibration under digital twin integration. You know here also it plays a critical role. So adults calibration involves our advanced distribution system validates precise calibration for forward facing camera for accurate land reduction, LIDAR sensors for comprehensive obstacle recognition, rather units for reliable distance measurement. Surround view camera systems for enhanced parking assistance. Rus unparalleled. Performance and choose complex calibration procedures for, that can be completed within, under 90 seconds per vehicle in a high volume manufacturing line. So digital twin integration. So real time vision data seamlessly integrates into factory digital twin, facilitating productive maintenance through insightful trend analysis. That tracks each of the process equipments on the critical process line. Dynamic real time process optimization robust closed loop control systems, efficient virtual commissioning of a new product line future directions and implementation roadmap. So current deployment, so bacteria and powertrain inspection systems with the. Basic AI capabilities achieving 30 percentage reduction in s escape defects. Our near term target is to integrate integration of a multimodal sensing like a vision, ultrasound and thermal sensors with advanced machine learning models capable of product to quality assessments. In the midterm, fully autonomous inspection system with the self optimized algorithm and automated root cause analysis on, on, in cars for long term, the cognitive manufacturing system where vision and robotics. And the process controls form a unified AI production ecosystem which could be achieved in a long term approach. Rest will remind our foundation with its ecosystem maturing alongside these advance months to provide performance, safety, and reliability requires for next generation EV manufacturing. That concludes my presentation. I thank you all for your time and attention today. I truly appreciate your participation and hope the session gave you valuable insights into AI power, the mission in rest language. Thank you.