Good morning everyone. I am ti and today I will be discussing our research on resilient health monitoring system, specifically engineered for disaster zone. The subtitle of my presentation captures our core mission, transforming proof of concept health monitoring. Into battle tested system capable of operating when infrastructure fails. This is crucial work because traditional health monitoring system are designed for control environment with reliable power and stable network connectivity conditions that simply doesn't exist in disaster zones. When earthquake strikes, hurricane hits, or other catastrophe occurs. Conventional healthcare infrastructure collapses precisely when it's necessary. Most, our research addresses this challenge by re-imagining health monitoring technology from the ground up, focusing on resilience, power, efficiency, and reliable communication. Even in the most challenging environment, the ultimate challenge. When designing the health monitoring for Disaster Zone, we face three critical challenge challenges that form what I call the ultimate challenge. First, we need to provide life critical monitoring. This means tracking vital signs, including heart rate, blood, oxygen level, SPO two body temperature, and respiratory functions parameters. That can mean the difference between life and death. Unlike consumer fitness tracking trackers that can tolerate occasional inaccuracy, our system must maintain clinical accuracy even under extreme conditions. Second, we operate in hostile environments when traditional infrastructure has collapsed. Think about post earthquake scenario. When no power grid, no internet connectivity, extreme temperature, dust, and debris, our system must continue functioning reliably despite these conditions. Finally, we are targeting 99.99% of time. This means our system can only afford about five minutes of downtime per year when lives depend on continuous monitoring. This level of reliability isn't a luxury, it's an absolute necessary. The BLE connection, resilience. To meet these challenges we have developed several innovations in BLE Connection. Resilience first are dynamic device discovery enables automated reconnection protocols that continuously adapt to challenging environment. This means that as healthcare providers move patients or as condition changes, the system automatically rebuilds connections without requiring manual reconfiguration. We have implemented multi-part communication that ensures signal persistence even when primary route fails. Our field testing shows that approaches has increased connection. Reliability from 92.4% to 99.7% critical improvement when lives are at stake. The system includes sophisticated signal degradation handling that maintains core functionality even at minimal signal strength, even at signal levels as low as minus 95 DBM. The system continues transmitting essential vital signs. Perhaps most importantly, our device from mesh networks where they can rely on data when direct connection fails. This creates a self-healing network architecture that can route around damaged or disconnected nodes, maintaining continuous monitoring capabilities despite localized failure, power optimization, breakthrough. Yes. Power management represents one of our most significant innovations as it directly impacts how long these systems can operate in the field. Our dynamic sensor sampling intelligently varies collection frequency based on the patient's criticality and remain power reserve and remaining power reserves for stable patients. Sampling rates can decrease to conserve power while critical patients are. Monitoring more frequently. This approach has demonstrated as a remarkable 62.4% reduction in power consumption for stable patients while maintaining clinical standards. We have implemented advanced edge computing algorithm that analyzes data locally drastically reducing energy intensive transmissions. This low power processing approach has reduced ideal. Current consumption, idle current consumption from 4.2 milliamps to just 0.8 milliamps and 81% improvement that directly translates to longer operational life. Our transmission optimization uses strategic compression and scheduled data delivery protocols to reduce radioactivity cycle by coordinating transmission windows. Across multiple devices, we have achieved a 28.3% reduction in overall power consumption as compared to standard BLE implementation. Finally, we have integrated innovative energy harvesting technology that captures kinetic thermal and ambient RF energy. These micro generators create self-sustained power systems that can extend operational. Duration by 18 to 26% under favorable conditions, potentially providing indefinite operation in some field settings. OTA update in constrained networks. Maintaining software currently in a currently is vital for security and functionality, but traditionally update mechanism, traditional update mechanism fails in disaster environment. With limited connectivity, our Delta update approach transmits only modified components rather than complete firmware images. This reduces bandwidth demand by over 80% with typical update payload size, dropping from 48 to 96 KB to just eight to 12 kb. This efficiency is critical when network resources are severely constrained. We have developed proprietary extreme compression algorithm that can shrink update payloads by nearly 83%. This enables critical patches, even on severely limited network, the traditional updates would be impossible. Our partial update recovery use sophisticated checkpointing to allow interruptions, interrupted updates to resume from break points. This eliminates redundancy. Data transfer during network fluctuation, ensuring updates are complete even with intermittent connectivity. Finally, our intelligent rollback safety mechanism automatically reverts to the last stable version if deployment integrity check fields. This ensures continuous device operation even when updates, encounter problems with field testing demonstrated in 92.2%. First attempt update success rate. Next, let's talk about distributed observability. Maintaining visibility into the system performance is crucial in disaster scenarios, but traditional monitoring approach fail without reliable infrastructure. Our system provides comprehensive real time virtualization visualization across all deployed monitoring devices in the disaster zone. This gives emergency responders and medical personal immediate insights to, into both system health and patient status. We have implemented dis distributed tracing capabilities that provide crucial visibility into patient interaction during network destruction. This allows us to maintain accountability and data integrity despite challenging conditions. Our advanced machine learning algorithms can provide identify potential system failures and psychological emergency before critical incident records. Field testing shows our system can detect deterioration of two 13 point five minutes before conventional indicators, potential life saving, early warning. Finally, our intelligent bandwidth optimization prioritization, transmission of life. Critical metrics when communication infrastructure is severely compromised. This ensures that most important data gets through even when bandwidth is extremely limited error Budgeting for critical care, we have adopted the site reliability engineering concept of error budgeting to ensure our system maintaining, maintain. Approximately reliability, appropriate reliability for different functions. For overall system reliability, we have targeted 99.9%, which means a maximum downtime of just five minutes per year. This ensures that healthcare providers can depend on the system being available when needed for critical alert delivery notification that can be lifesaving. We aim even higher with 99 point. Nine, nine, 9% reliability target. This translates to less than one minute of annual downtime For these critical functions for non-critical function, we allow slightly more flexible flexibility. With a 98.5% target, the strategic approach allows more innovation in secondary features while maintaining the ionic lag reliability for the most important capability. This error budgeting approach ensures that we focus our reliable efforts where they matter most on the functions that directly impact patient safety and outcomes. Healthcare specific SLIs and SLOs. We have developed healthcare specific service level indicators and service level objectives that directly relate to clinical outcomes. For vital sick sign latency, we target less than one second with a critical threshold of three seconds. This ensures that healthcare providers are working with current patient data, not historical information that might no longer reflect patient's status. Alert delivery type is even more stringent with a target of less than 200 milliseconds and a critical threshold of two seconds. When a patient's condition deteriorates every second counts and rapid alerts enable fast inter intervention. Data accuracy targets 99.5% with a critical threshold of 98.5%. This high standard ensure that critical decisions are based on trustworthy information, reducing risk, reducing the risk of treatment error. Finally, battery life prediction accuracy aims for a plus minus 5% error with a critical threshold of plus minus 10% Accurate battery prediction prevents unexpected device failure that could leave patient unmonitored during the critical periods. These healthcare specific metrics ensures that our system meets the unique requirement of medical monitoring in disaster in the. Minutes, graceful degradation patterns. When resource become constrained, our system implements sophisticated graceful degradation patterns to maintain essential functions. Our priority based functions shedding disabled non-essential features first, as resource es. Preserving critical monitoring functions until absolute failure. This means. Capabilities, like high resolution wave or historical data access might be reduced before vital signs. Monitor is affected. Data resolution scaling dynamically adjust sampling rates and precise precision based on patient status. This mainten maintains higher resolution for abnormal readings, ensures critical accuracy when it matters most while conserving resources. On stable patients, our first, our local first processing approach shifts to autonomous operations when disconnected from central infrastructure. Local alerting continues without central system ensuring patient monitoring continues even when the network fails completely. Finally across, finally, across cross device redundancy allows nearby devices to assure monitoring reliability. For failing units, patient data is seamlessly transferred between devices. Maintaining continuous monitoring, even when individual component fails, real world impact. Our system has already demonstrated significant real world impact across multiple scenarios. During the Napal earthquake, we successfully monitored over 5,000 patients across 15 market makeshift field. Hospitals with 99.3% of time during critical disaster response operation. This provides continuous continuity of care despite complete collapsed infrastructure, enabling more efficient triage and resource allocation following Hurricane Maria. Our system was rapidly deployed when hospital infrastructure collapsed, providing 17,000 patient hours of uninterrupted vital. Monitoring in an extreme condition. This allows healthcare providers to focus on treatment rather than manually monitoring, significantly improving the efficiency of limited medical personnel beyond immediate disaster response. Our technology has transformed healthcare delivery by extending critical monitoring capabilities to facilitate 200 facilities in. Underserved regions without reliable power infrastructure to demonstrate how innovation driven by extreme requirements can broader impact on global healthcare, across global healthcare access. Key takeaways to conclude, I want to emphasize that SRE principles can transform healthcare technology, reliability, and extreme conditions when human lives. Lives depends on uptime. Traditionally reliable standards and approach simply aren't enough. Our research demonstrate that through focus, innovation, and in connection, reliance in connection, reliance, power optimization, update mechanism, and g graceful degradation, we can create health monitoring system that functions reliably even in the most challenging environment. The impact extends beyond immediate disaster response, potentially transforming healthcare delivery in resource constraint environments worldwide, and establishing new standards for medical device reliability. Thank you for your attendance. I'm happy to take any questions about our research or the specific technical approaches that we have developed. Thank you again. Bye bye.