Search in Real Life

by Michael Wang

The biggest impact of the Internet is that it made information ubiquitously accessible. When the Internet was first invented (called Arpanet at the time), a relatively small number of research institutions shared information over connected servers. But as the amount of information grew exponentially over the decades, a more efficient way of indexing and searching for data was needed. Companies like Google and Yahoo jumped on this opportunity. They realized that without an effective search engine, the ever-expanding interconnected databases would be too unwieldy to access and use. Figure (a) shows the digital architecture of the Internet today. The search layer serves as the connecting layer that enables web applications to efficiently access and manipulate the information on the data layer. None of the Internet applications we use today (Facebook, News, Netflix, Email, weather, etc.) would be possible without effective searching.

Up until recently, most devices connected to the Internet are computers and smartphones, which are mainly used to store and process digital information. As sensors and wireless communication chipsets continue to improve in cost and performance through Moore’s Law, a new class of Internet-connected devices are being created: physical objects (or things). These Internet-of-Things (IoT) devices are not general purpose computers used to access digital information stored on the internet. Instead, they consist of sensors and actuators designed to make the physical objects themselves connected, accessible, and controllable from the Internet. By 2020, there will be 50 billion such devices connected to the Internet, more than 7 times the number of human beings in the world.

The value of having so many connected devices is actually quantified by Metcalfe's law, which states that the value of a telecommunication network is proportional to the square of the number of nodes. The challenge of searching through this rapidly increasing network of connected things requires a new form of search engine, consisting of both a physical layer and a digital layer, as shown in Figure (b).

These physical objects can be broadly divided into two categories: high-value items and low-value items. High-value items include things such as cars & other vehicles, buildings, expensive electronics, and even human beings. To connect these objects to the Internet, they require a physical search layer, consisting of sensors and actuators, for each individual object. For example, a smart car contains an array of sensors that serves to monitor the state of each individual car; a person wears a Fitbit, which serves to monitor the health of each individual person. On the other hand, low-value items include things such as individual pieces of merchandise in a grocery or retail store. It is not cost-effective to individually tag each low-value object with a set of sensors and actuators. Instead, sensor modules can be designed to identify a collection of items.

A truly effective search engine for IoT devices is essential for unlocking the full potential of these connected devices. Every object should be searchable based on their precise location, state, behavior, and actuation capabilities. (pronounced “surreal”), which stands for Search In Real Life, aims to be the most accurate and efficient search engine for the physical world by integrating novel indoor positioning technology,  low-cost yet sophisticated sensor fusion, with a user-friendly full-stack web platform. Imagine a manufacturing company being able to track the position and state of all the inventories they have in their supply chain in real time; imagine a smart car being able to determine dynamically who is sitting in which seats, and being able to adjust the safety and comfort setting individually for each person; imagine a grocery store customer finding precisely where a specific product is without having to fumble around manually.