Often when we are thinking about self-driving vehicles or delivery robots, we imagine machines with powerful sensors and powerful processors that are guided by complex algorithms. What NASA is proposing here, a Data and Reasoning Fabric or (DRF), is a way of distributing some of the sensors and computing power into a digital infrastructure. The key to this is what NASA is calling an edge node. Many edge nodes networked together across a city create a digital net that gathers information, about weather for example, and can provide direction to drones or air taxies.
Navigating a city is a complex matter of adjusting and adapting to the environment in one’s immediate vicinity while simultaneously moving toward a destination. People accomplish this with apparent ease. However, even for people information about the surroundings beyond immediate perception, such as an incoming storm, traffic, or an accident can change and improve how they negotiate the city. For instance, depending on someone’s reasons for getting to a certain location, the hospital say, the added information that rain is imminent, or that the travel time is greatly increased, may mean not going at all or going earlier to ensure a timely arrival.
Some aspects of this proposed solution should be familiar to all of us. The phone in your pocket depends on a networked structure like this. That network is composed of cells, hence why American’s use the term cellphone. Like a digital beehive structure each cell is covered by a tower that receives and transmits data, a cell tower. An edge node is like a cell tower, but with a significant difference: it isn’t passive, it’s active. It is in these nodes that the ‘reasoning’ from the name Data Reasoning Fabric is supposed to take place.
As the name implies the idea is that some parts of the logistical reasoning are distributed across the edge node network, or fabric as it is called in this case. Autonomous drones will, undoubtedly, have some onboard processing power as well. They will also have sensors. But here too the DRF model predicts distributed sensors, envisioning a smart city marketplace where companies put their data, from traffic to weather, out for services like the DRF to utilize.
The key difference between DRF and the cellular data technology we are familiar with is the distributed computing process. This alleviates some of the bottleneck for vehicles, potentially reducing cost, and sets the groundwork for the safe, effective, and efficient coordination at significant scale, which would be needed if we are to have flocks of drones in tight city areas.