Autonomous digital twin

As explained in Digital Twins in Action

An autonomous digital twin is a digital twin that independently evaluates changing conditions and takes or recommends real-world actions beyond fixed preprogrammed rules.

An autonomous digital twin closes the loop between a physical system and its digital representation. It uses information, predictions, or simulations to take real-world action or recommend a decision. Action can mean opening a pressure-release valve when tank pressure is predicted to become unsafe, or sending a notification to a person who can respond. section 1.3.5 section 1.7

Autonomy is not the same as fixed-rule automation. A conventional control system responds to sensor readings according to preprogrammed logic. An autonomous twin can instead weigh conflicting objectives, anticipate future states, and adapt its proposed control strategy. It need not use agentic AI: reinforcement learning, model predictive control, and multi-objective optimization are also possible approaches. The test is whether it independently evaluates tradeoffs and acts or recommends action beyond fixed rules. section 1.3.5

Home temperature

A basic thermostat turns heating or cooling on to reach a configured temperature. A smart thermostat offers a more concrete view of adaptive action: presence sensing through household phones can enable Eco mode when occupants are away. The useful question is not whether the device is branded “smart,” but what information it uses to decide when to change its behavior. section 1.3.5

Figure 1.12 The interface to a Nest smart thermostat showing how it takes action based on your physical location. Eco mode is enabled by presence sensing through the location of mobile phones in the household. Image © Thomas Smailus, Ph.D., P.E. Reproduced with permission.

The thermostat interface shows Eco mode enabled through presence sensing based on household phone locations—a visible link between a changing real-world condition and an action. section 1.3.5

Start with the intended outcome before choosing an autonomy technique. A digital representation may progress from static description to live updates, prediction, and simulation; closing the loop adds decisions that affect the physical world. Decide whether those decisions should be recommendations, electronic actions, or direct actuation, and assess the skills and build-or-buy approach needed to deliver them. section 1.3.5 section 1.7

How Digital Twins in Action covers it

How deep it goes

Defines autonomy through adaptive tradeoff evaluation, contrasts fixed-rule control, and names possible techniques without detailing their implementation.

Examples

Uses a pressure-release valve, device notification, and presence-aware home thermostat to illustrate different responses to changing conditions.

What it assumes

Requires a digital representation and information about the physical system; no particular AI technique is required for autonomy.

What you can do after

Evaluate/understand it by distinguishing adaptive decisions and recommendations from responses governed only by preprogrammed rules.

How it connects

Connects digital twins to predictive modeling, simulation, process control, reinforcement learning, model predictive control, and multi-objective optimization.

liveBookmanning.com

Read Digital Twins in Action on liveBook Get the book at manning.com

Related concepts

• continuous improvementcontrasts with · Continuous improvement is an ongoing process of evaluating outcomes and making successive changes to improve a system, process, or result.
• digital twinkind of · A digital twin is a computational representation of a physical entity, process, or system that models relevant characteristics, state, or behavior.
• AI agentuses · An AI agent is a computational system that uses artificial intelligence to interpret information, select actions, and pursue goals through interaction with an environment.
• sensor datauses · Sensor data consists of measurements produced by sensing devices about physical conditions, events, equipment, or environments.
• reinforcement learninguses · Reinforcement learning is a machine-learning paradigm in which an agent learns actions or a policy through interaction with an environment and evaluative feedback, such as rewards.

This explainer was generated by Manning's knowledge graph from the text of Digital Twins in Action. The section links point to the parts of the book it draws on.