Rickerism

Law and expectations

The expectation of most customers, either stated or implied, is to create a replicated state of the physical realm within the digital realm. All efforts in sensor fusion and SODA are inexorably pulled toward fulfilling this expectation. The expectation is a bit like wanting a perpetual motion machine. Within the engineering community, it is common knowledge that the laws of thermodynamics prevent one from creating a perpetual motion machine. Unfortunately, there is no such equivalent knowledge of the limits of computer science. We see young engineers pursuing efforts with great energy that are, by their nature, impossible by the laws of science.

When dealing with computer science, there are laws just as stringent as those of thermodynamics. All those pursuing sensor fusion should be aware of them.

• Turing’s First Law, or the Law of Algorithms. A Turing machine (i.e., a digital computer) can only execute an algorithm.
• Turing’s Second Law, or the Law of Inherent Compatibility. A Turing machine of enough complexity can emulate any other Turing machine.
• Gödel’s Law, or the Law of Incompleteness. A set of axioms can be consistent or complete, but it cannot be both.
• Miller’s Law, or the Law of Seven, Plus or Minus Two. All humans have inherent natural cognitive limitations on how much information they can process.

By being aware of these laws and knowing the customer’s expectations, we can achieve practical results. We may not deliver a perpetual motion machine, but we can deliver a highly efficient Carnot engine.

Vision

I want to present to you a concept of something that I call quintessential space, or Q-space.

Quintessence is “the fifth and highest element in medieval philosophy that permeates all nature; the essence of a thing in its purest and most concentrated form; the most typical example of representation.”

The Q-space could be described as a non-visual virtual reality. It is a subset of the Internet, or digital realm, that contains within it a representation of the state of certain things in the physical realm. For each object of interest in the physical realm there is a corresponding representative in the Q-space. The Q-space does not represent all of reality; that would be impossible according to Gödel’s Law.

Avatars

We began looking at this challenge with DARPA and the challenges of the battlefield, so in the discussion to follow I will sometimes refer to the battlefield. One can easily substitute factory, emergency room, supply chain, etc.

How does one maintain the state of the Q-space current with the physical space (Φ-space)?

• Challenge: If maintaining the Q-space is a additional, secondary function, then it will not be maintained and it will never be an effective representation of the Φ-space.
• Solution: Make maintaining the Q-space an automatic part of normal battlefield functions.

If each person and system on the battlefield communicated with and through its representation in the Q-space, then the representation would have the most accurate state available of the person or system. We call the representation in the Q-space an avatar.

Suppose soldier A wanted to communicate with soldier B. With Q-space communication, soldier A would instruct his avatar to send a message to the avatar of soldier B. Soldier A sends all messages through his avatar. Soldier B receives all messages through his avatar.

An avatar is server software that communicates on multiple protocols. An avatar provides two types of connections: the entity and other avatars. An entity can be a person or a system such as a vehicle.

It essence, an avatar is a collection of services in the digital realm that are associated with a particular entity in the physical realm.

Sensor fusion

Avatars only provide a state of people and systems that communicate. Using only avatars, the Q-space would only represent “friendly” objects. For full representation, we need non-communicating objects such as terrain, buildings and enemy objects.

• Transducer space creates digitized interpretation of the state of the physical space
• Monad space creates a normalized representation of the transducer space
• Agents act upon the monad space to create the Q-space
• Q-space provides feature representation of the physical space to provide humans awareness so that they can create decisions
• Humans act to change the physical space

Transducer space is the normalization of all sensors over the Internet. T-space is a collection of XML and RTP data streams over the Internet.

Monad is a Greek term meaning “ultimate, indivisible unit”. The monad is a specific measurement value at a specific location at a specific time, such as the red spectrum intensity read by a specific pixel at a specific time observed along a specific vector from a specific point.

The M-space creates a massive, homogeneous representation of all transducer readings M-space is a distributed terabyte database. Picture a giant, rolling wheat field. Each grain is a monad, a specific uniform record in the terabyte database.

Agents scourer the monads looking for patterns like bees looking for pollen. Agents create events when they detect a pattern. Other agents scourer agent events looking for patterns of patterns. A food chain of events occurs, eventually creating a feature, such as a hill or tank, that can fit in the Q-space.

Implications on SODA

The concept of avatars and monads has two direct effects on the design of SODA.

First, devices should be treated as avatars. An avatar is a collection of services associated with an entity in the physical realm. Remember that there is only data in the digital realm. It is through the Q-space that the human can associate the data with the physical realm.

Second, SODA must enable sensor fusion. Since only a fraction of the physical realm will actively communicate through avatars. We can treat the sensor entity as an avatar, but what about all of the entities that the sensor is measuring? In order to achieve the Q-space, all sensors must enable sensor fusion. The Q-space will not be created in a day. As programmers create better agents and algorithms, we will be able to identify and track more and more entities in the Q-space.