Hi, welcome to the conference. Today I will be talking about self-healing BLE networks powering resilient iot at scale. Introduction to self-healing BLE networks. Today we are exploring a transformative approach to internet of things, specifically self feeling Bluetooth networks. In an increasing connected world, we need networks. That can automatically detect, diagnose, and recover from fault without human intervention. The core concept is simple, yet powerful, create intelligent network that can maintain themselves. BLE has emerged as an ideal protocol for iot due to its low power consumption and scalability, but traditional approaches have. Struggle to make these networks truly resilient. So let's talk about the growth and importance of network resilience. Let's talk. Talk about the market dynamics, the self-healing iot market is. Predicted to expand dynamically through 2030, driven primarily by industrial sector adoption. This isn't just technological trends, it's a business imperative. Critical industries like healthcare and infrastructures are targeting 99% uptime. That means less than an hour of downtime per year. A standard that demands extraordinary network reliability as an as enter, enterprises develop IOT systems at mass scale. We are talking thousands of interconnected sensors and devices that meets for advanced self-healing protocol becoming crucial. Current challenges in IOT networks, the obstacles are significant. Significant. BLE devices are in currently resource constrained with limited processing capabilities and memory. This restricts the complexity of algorithms we implement, creating a fundamental challenge in building in intelligent network as networks grows. Reliability become reliability becomes exponential and more complex. More nodes means more potential points of failure leading to unpredictable propagation patterns. Traditional fault recovery methods suffer from unacceptable latency between detecting a fault and recovering the network power consumption. Add another layer of complexity in field development. Continuous network monitoring can rapidly deplete battery services, making long-term autonomous operation, challenging our open source framework approach. We. We developed a comprehensive solution addressing these challenges through four key innovation. First intelligent recovery. We have implemented machine learning based anomaly detection with predictive healing capabilities. This means not just detecting failure, but anticipating and preventing them. Second, a distributed architecture that enables coordination. Between local and network level healing mechanism devices can make intelligent autonomous decisions while contributing to overall network resilience. Third memory efficient design are algorithms are purpose built for resource constrained iot devices, ensuring sophisticated fault recovery mechanisms with minimal memory footprint. Finally, we have leveraged modern programming patterns that enables advanced concurrent processing with minimal computational overhead. Next, let's talk about power efficiency across operational states. Our power optimized strategies cover every network, operational state due. During active scanning, we have reduced power draw by 60 to 75, 70 5% through dynamic channel hopping and adaptive duty cycle in connected mode, optimized packet framing and connection intervals tuning E 40%. Energy saving. Energy consumption. Our fault detection mechanism uses intelligent, threshold based monitoring, reduces power spikes by 83% compared to. Periodic pooling. In standard mode, we have achieved sub 10 point, sub 10 micro amp powers consumption with a vehicle latency under 3.5 milliseconds for critical events. Let's talk about the life weight machine learning implementation. What we have done here's our approach. Becoming truly innovative, our machine learning model models consume just four kb of memory while achieving 95% detection accuracy for network anomalies. Instead of relying on external service, each device conduct local real-time analysis. We have implemented a distributed intelligence model where insights are shared across the network through federated learning, ensuring continuous improvement. Without compromising data privacy, adaptive thresholds, and automatically calibrate based on the environment, conditions, and network traffic dynamically reducing the false positives while maintaining the subsecond detection speed. Let's talk about the performance benchmark. Now. Our frameworks dynamically outperform traditional approaches. We have achieved 90% faster recovery times triple battery life and improved packet delivery reliability by 14%, all while using only 25% of, memory of conventional by using conventional methods. Next, let's talk about the architecture for scale. Our architecture is designed to manage complexity at unpredictable scale. We can orchestrate thousands of BLE device with zero touch provisioning and automated firmware distribution. Advanced scheduling algorithms allow 10 to 20 concurrent BLE connections per gateway with minimal latency. Arm meh coordination enables transparent roaming with sub 50 millisecond handoff. Between coverage zones, ensuring uninterrupted operations. Now let's talk about code implementation and examples. We have created a comprehensive open source code base that delivers production ready implementation of BLE, connection management, anomaly detection, power optimization, and memory efficient package handling. Our goal. Is to solve the most challenging aspect of resilient iot development with minimal integration efforts, empowering developers to build more robust networks. So now let's talk about the real world application. Using these methods, let me illustrate our framework impact across diverse scenarios. In industry automation, we have maintained. We are maintaining 99.9% uptime. Our mission critical equipments monitoring is high interference and in high interference environment. Some cities, can now manage vast sensor networks with self diagnosed capabilities that reduces maintenance visits by 78%. Healthcare monitoring benefits from uninterrupted patient data streams with redundant communication pathway, smart building scan, integrate climate control, security and energy management through resilient mesh networks that maintain operations during connect connectivity challenges. Next, let's start off, talk about the key, takeaways and the next steps and the future directions. Modern programming has fundamentally transformed network resilience. We have proven that resource constraint devices can implement, can implement sophisticated self-healing mechanisms without compromising performance or battery life. Machine learning is now practical on low power devices. Adaptive power management techniques allows extended device functionality without sacrificing. Recovery times our reliability. I invite you to join the open source community. Visit our A link to access our implementation, contribute to the project, and help us shape the future of resilient iot networks. Together we can build that doesn't just connect, but heal, adapts and times. Thank you.