I'm really excited for this opportunity to share what's possible for robotics with AWS us. Let's get started. The use of robots continue to accelerate as industries and organizations realize the role of automation in improving productivity, gaining business continuity, and ensuring flexibility and adaptability. We see increased adoptions across all major sectors, including industrial services, consumer agricultures and more. The number of robots is estimated to grow to 20 million by 2030. People have been creating devices and machines to perform mundane tasks for hundreds of years. However, the ability to create a device that can sense, compute and act with a degree of autonomy is relatively new. The first generations of robots, such as the robotic arm, could perform simple, systematic tasks. They were largely pre programmed or human directed. Since those early years in robotics, there have been remarkable advancements in computing, sensor technology and machine learning that has enabled a shift towards next generation robots. Robots with a wide range of autonomy, powered by the advanced data analytics and natural human interfaces that can perform advanced tasks in collaboration with humans. We believe that connecting robots to the cloud and cross edges is powering the next evolution in robotics. With AWS, connected robots can capture, store and aggregate limitless amounts of data to train and develop new functionality, optimize and drive efficiency. Connected robots can update themselves to ensure reliability and reduce downtime. Operators and technicians can remotely connect and interact. So why do customers choose AWS to build, deploy and manage their robots? First, AWS helps you build mature solutions from day one. With AWS, you can become agile in your experimentation and innovation. With AWS, you can deliver a differentiated value proposition by collecting and harness the power of data to drive innovation. AWS provides unmatched scalability, continuity and reliability. The elasticity of the cloud means more compute power to provide increasing levels of autonomy and facilitate orchestration. AWS enable you to extract the full value of robots automation. Whether you are looking to reduce downtime or implement a new business case, let's look at how customers are using AWS for robots. Robot builders use AWS to build intelligent connected robots faster. You are able to train robots to execute complex tasks, collect telemetry data securely, and continuously improve your robotics devices and generations software. In addition, you are able to test the new capabilities using simulation across a variety of scenarios as well as accelerate development. By using pretrained machine learnings and artificial intelligence model, you are able to build for scalability and operating robots solution at scale as well as maintaining the solution remotely. With cloudbased software provisioning and over the air update, you are able to connect and monitor homogeneous robot fleets with vendor integrations and the ability to build applications that help optimize traffic flow. We have a portfolio of services that can help support all of these use tasks, whether you want to connect and manage your robotics application or to enable new workflows. At AWS, we always start with connectivity and collecting data from the devices to enable many other capabilities and use cases. You can leverage AWS services to add capabilities such as over the air update, predictive preventive maintenance generations, or machine learning interfaces at the edge if you are operating robots at a scale and are looking for a single view glass, we have services that can help you monitor desperate robots fleets more efficiently. Interfaces of robotic things where intelligent devices can monitor events, fuse sensor data from variety of sources, use local and distributed intelligence to determine the best course of action and then act to control or manipulate objects in the real world. It have four aspects sense connect learn activate the sense is about collect and process data streams from sensors, lighters, cameras and other sensor in the environment and store the data in Amazon SS three or Amazon Kinesis data stream. Connect is about sending the data to the cloud or even connect and interpret with other robotics devices, system and equipments learn is about using the data collected from the robots and run and train machine learning models. Then deploy the model in the robot where robot can execute complex function and workloads and make decisions. Accoutate is about interacting the environment in a safe manner, interact with human with other robots and other automation safely. Let's look at common use cases architecture. In this architecture, the robot is collecting sensor data and log data and transfers the data to the cloud using AWS IoT core where the data could be routed to be stored in varieties of database systems and also can be stored to build analytics data lake which could be consumed by visualization tools in order to provide reports and life dashboard. In the same time, IoT core providing a function which is called device shadow which is a mainly adjacent document representation of the actual and the desired state of the robots. So an application can interact with the robots through a shadow where the application sets the desired state for the robot and when the robot is connected to the Internet, it will connect to the IoT core, fetch the desired state from the shadow and start act to present this desired state. Also, when the robot state itself is a change, it will connect to the IoT core and update the shadow. Say this is my current state. AWS IoT core lets you connect billions of IoT devices and robots and route trillions of messages to AWS services without managing any infrastructure. IoT core is a managed cloud platform that lets connected devices easily and securely interact with a cloud application and with other services. There are different communication protocols including MQTT, HTTPs, MQTT over Websocket and LoRaWAN AWS. IoT core also secure device connections and data with mutual authentication and end to end encryption. IoT core can filter, transform and act upon devices data on the fly based on the business defined rules. Another architecture is about software development and managed application architecture where you have two fleets of robots, fleet a and fleet b, and you want to deploy different version of your application like an a b testing. For example, you can use IoT core to create things group. In this case we create thing group a and things group b and ensures that all robots in fleet a is under things group a and all robots in fleet b is under things group b. Then you can use IoT green grass to deploy software a to group a and software group b to group b, which in turn will need to be deployed to only fleet a and fleet b. So IoT Greengrass is an open source edge runtime and cloud services for building, deploying and managing device software. Iot greengrass make it easy to bring intelligence to edge devices such AWS, anomaly detection and powering autonomous devices. You can deploy new or legacy app across fleets using Java, bison node js or even running a container image. IoT greengrass can collect, aggregate, filter and send data locally, also manage and control what data go to the cloud for optimized analytics and storage. Another architecture is where you want to run machine learning at the edge and this is also using IoT greengrass as it make it easily to perform machine learning inference locally on robots using models that are created and trained and optimized in the cloud. IoT green grass gives you the flexibility to use machine learning trained in Amazon Sagemaker or even to bring your own trained model and save it. In Amazon SSV, you can use machine learning model that are built, trained and optimized in the cloud and run interfaces on robots. For example, you can build a predictive model in sagemaker for sense detection analysis, optimize it to run on any camera and then deploy it to predict suspicious activity and send an alert data gathered from the robots itself running on it. Green grass can be sent back to stagemaker where it can be tagged so it can be used continuously to improve the quality of the machine learning model. In this architecture, the user want to stream videos from robots and also have a playback through a mobile application. Amazon can use things video stream make IoT easy to securely stream videos from connected robots to AWS for analytics, machine learning, playback and other processing. Kinesis video stream automatically provision and scale all infrastructure needed to ingest streaming video data from millions of devices. IoT durably stores, encrypt and index videos data in your stream and allow you to access your data through an easy to use API. Kinesis video streams enable you to play back video for live or on demand viewing. Quickly build applications that take advantage of computer vision and video analytics through integration with Amazon recognition video and libraries for machine learning frameworks such as Tensorflow and OpenCV. Kinesis video stream is also supporting WebRTC. This is an open source project that enable real time media streaming and interaction between web browsers and mobile application and connected robots via simple API. Case video stream supports media ingestions over WeBRC connection for secure storage, playback and analytics processing in things architecture the user want to use a simulation to test the same robot application in different simulation walls, world one and world two. And here I would like to mention three of the core benefits we have heard from customer for using simulation in robotics development. First, the ability to reproduce and test exact scenarios that have triggered unexpected behavior in the past, including edgy cases and unsafe condition. This is difficult when testing with physical robots. Second, you can speed up the clock and run simulation faster than real time, producing results in a fraction of time it would take on physical robots. Third, you can expand test coverage by programmatically testing many scenarios as here we are testing simulation one and simulation two and this can be multiplied using parapterized and repeatable simulation. Robomaker is a cloud service that make IoT easy to build, test and deploy robotics application. In this architecture, we want to build an application that can work with robots from system a and also robots from system b where you have different robots from various vendors and you want them to work seamlessly with each other. So AWS IoT robo Runner make it easy to build application for optimizing fleets of diverse robots. IoT Robo Runner provides central data repository for storing and using data from different robots management system and enterprise system. Once robots are connected, you can use sample application and software development libraries to build management application on top of the centralized data repository. IoT Roborunner help you to build complex management application that require robot's interoperability such as task orchestration and view information in a single unified display. As you can see here, you can use a robots runner for designing a shared place. You can define the entry points and the exit points. Robots wait at the entry points if they are not cleared for entry. By the time they get there, robots notify system of exiting the space at exit points. In this architecture we will walk through a sample solution that implements CI CD pipeline with automated scenarios based testing and simulation. First, AWS code pipeline is a fully managed continuous delivery service. You can connect with your code repositories in GitHub, GitLab, BitBucket and automate a set of build and test actions at each stage of your release. Once the code pipeline is set up, developers work in agile sprints and build new functionality in a feature branch. When ready, they submit a board request with new code. The board request gets reviewed and eventually merged into integration branch. This starts first two stages in the CI CD pipeline. First, the source of code is copied into a build server running in AWS code. Build a fully managed continuous integration tool that will run the ROS and Docker build command. Then copy build container into Amazon Elastic container registry, a center repository of container images. Once the container images are built and published, the next step is to run batch of tests. Automated tests in robomaker simulation in this solution, we use AWS step function, a low code solution for building state machine in the cloud to track and send notification on the progress of the simulation. Rotest results are automatically copied from a robomaker simulation to Amazon SS three bucket where they can be queried, analyzed and visualized using Amazon Athena and Amazon quicksight. Once a test in the simulation pass, the new container images can then be automatically deployed to real robots in a test area using AWS IoT green grass. Results from the test on the real robots can also be uploaded to Amazon SS three, then can combined with simulation test results. After all of the test validation check tasks, the code can be merged into the main branch, then it be deployed over the air to production fleet. In things architecture, we are using a spot developed by Boston Dynamics for industrial facilities inspection with AWS services. So a spot is a robot developed by Boston Dynamics and we are using it to run two inference deployed by AWS IoT greengrass. So we are running AWS it greengrass which is hosting two interfaces. One is running locally to detect a valve and a second one is running remotely to detect the state of the valve if the valve is open or closed, and the spot will be an auto walk mission where it detects the two valves and uploads the data to Amazon SS three. Once the data uploaded to Amazon SS three, we are using lambda function to read the data and update DynamDb and Cloudwatch through GraphQL API by AWS Appsync we're also hosting a dashboard dashboard which displays the state of the vault represented in Amazon Dynamodb and also in Cloudwatch. Let's watch a real life demo that being presented on reinvent this weekend. Right now, spot is detecting the valve over there. It will go in can auto OC mission to detect the second valve over there and update the state of the valve on a dashboard that being hosted on a monitor here. So it detects the valve image. It's going to send the image itself to a Tensorflow model and based on that one it was updated a file to Amazon SS three where a lambda function will process the file, update the lost state. It will go back to where the mission started. And this is just like a pose from a spot to indicate it's uploading data and eventually the data will be updated on the dashboard. Robots are doing amazing things and AWS is helping the builders behind the robots you see in this video innovate in our Amazon fulfillment centers, you see robots helping make our environment safe for people and deliver value to Amazon customer. This is also true for many of our industrial customer. Other customers are focused on using robots to solve some of humanity's biggest challenges, such as making our world more sustainable. So we are just getting started but see the incredible potential in the space. It's one small step per man, one diaphragm permanent. Let's get started today. So I would encourage you to look at AWS robotics blogs. Where are there many blogs talking about robots connectivity, robots simulation, robots integration from different vendors using apprunner and robomaker. Also have a look at IoT greengrass as we've been talking before, it's a core component that can help you run a smart software at the edge. IoT can help you with over the air update and running interfaces at the edge. Also, there are open source sets available in AWS samples and AWS robotics. Thank you very much for joining my session and please give me your feedback.