Artificial intelligence (AI) is a technology that focuses on creating intelligent machines that are able to imitate human intelligence in processes. The original question is to define more futuristic visions to be implemented in the next decade.....
I had the chance to attend a keynote talk at IEEE PerCom SmartEdge'19 by Prof. Archan Misra and have his insight at possible answers to this question.
Knowing that number of IoT devices is growing exponentially, one very interesting idea for future infrastructures is to focus on cooperative cognitive abilities. By inferring knowledge through collaborative processing in clusters of IoT devices, a single machine's level of "intelligence" could reach the result of combined learning from an entire network.
For example, in the context of a Smart Home, it is plausible to envision that a single house could contain 1.000 sensing and actuating nodes in 2030. In this scenario, it can be very interesting to implement applications that can communicate and cooperate autonomously without the need for a permanent Internet connection. Performing learning at the edge could enable such systems to react upon human interaction in real-time. And delegation of heavy computing tasks could still be possible but only when necessary and desirable.
A second aspect would be to program these IoT devices in a way that allows them to infer autonomously what services other devices nearby can perform. This learning process would augment even further the features provided in the given scenario, as adding any new IoT device in the cluster would immediately allow the rest to know what new capability is available.
an interesting project to integrate IoT and AI would be:
Conception of solutions for the integration of sensors and actuators and the use of Smart Data (AI) from Big Data Analytics to reproduce human perception in distributed systems from sensors and actuators.
Nasri Nejah There have been a few indoor air quality sensors that link to devices, e.g. the Foobot. Related to some work I did with sensing indoor environmental parameters, I think it would be interesting to use sensors/occupancy preferences/IoT means to intelligently control HVAC in households. There are several 'low-hanging-fruit' applications I can think of...
An interesting project would be to develop an AI algorithm to control traffic lights. This is already a large area of study: "Smart" traffic light systems.
IoT - The collection of vehicle count sensors via Wireless Sensor Network (WSN) by a network coordinating device (gateway) + wireless traffic light control and monitoring = IoT application
AI - With an AI algorithm present in the network coordinating device (gateway), a device that has all traffic light data and vehicle count sensors. Thus, the times of traffic lights could be optimized by the algorithm in real time.