Towards Cloud Native Context Aware Autonomous Robotics for IoT Applications

Sensor networks are widely deployed around the world to measure different environmental variables over time, this general acceptance lowered the cost of installing such systems, but sometimes the requirement of the measuring makes it expensive or impossible to do. For instance, deploying sensors in a hard-to-reach area or just measuring for a brief short of time. This is where a new requirement is manifested namely automatic deployment of environmental sensor networks adjusted to the particular requirements of the end user. There is extensive research in IoT applications for environmental monitoring, autonomous robots and vehicles, context awareness, cloud computing, sensor implementation and sensor deployment strategies., but there are missing proposals and prototypes of complete and autonomous systems that interconnect all these topics focused on minimizing human intervention from a cloud-native platform. This paper proposes a basic prototype of a cloud-native context-aware system that semi-autonomously controls a line following robot for environmental monitoring as a first step towards Cloud-Native Context-Aware Autonomous Robotics for IoT Applications. A real-life experiment is made, having a line follower robot transport a sensor to three rooms inside a university, measure environmental variables such as temperature, dew point, pressure, humidity, altitude and illuminance and provide recommendations for user comfort and energy savings, all of this in an semi-autonomous way from a pure cloud backend. The performance and cost of the system is evaluated showing the advantages of a cloud platform in terms of cost, scalability and capabilities including an analysis of the latencies of the different requests. The proposed prototype also incorporates real-time sensor data analysis and provides personalized recommendations, resulting in a more efficient and user-centric monitoring solution as well as an energy saving solution for building management. This research proves that using cloud as a tool to integrate different already well-established systems under a simple architecture with easy to fulfill requirements is possible in terms of cost, capabilities and deployment time and can be scaled into big management and measuring systems for municipalities or buildings