Subsumption Architecture: Fundamentals and Applications for Behavior Based Robotics and Reactive Control

By Fouad Sabry

What Is Subsumption Architecture

The term "subsumption architecture" refers to a type of reactive robotic architecture that is closely linked to the field of behavior-based robotics, which enjoyed a great deal of success in the 1980s and 1990s. In 1986, Rodney Brooks and his colleagues were the ones who first coined the word. The concept of subsumption has had a significant impact on autonomous robots as well as other areas of real-time artificial intelligence.

How You Will Benefit

(I) Insights, and validations about the following topics:

Chapter 1: Subsumption architecture

Chapter 2: Behavior-based robotics

Chapter 3: Rodney Brooks

Chapter 4: Decentralised system

Chapter 5: Model-based reasoning

Chapter 6: Intelligent agent

Chapter 7: Action selection

Chapter 8: Nouvelle AI

Chapter 9: Hierarchical control system

Chapter 10: Genghis (robot)

(II) Answering the public top questions about subsumption architecture.

(III) Real world examples for the usage of subsumption architecture in many fields.

Who This Book Is For

Professionals, undergraduate and graduate students, enthusiasts, hobbyists, and those who want to go beyond basic knowledge or information for any kind of subsumption architecture.

What Is Artificial Intelligence Series

The artificial intelligence book series provides comprehensive coverage in over 200 topics. Each ebook covers a specific Artificial Intelligence topic in depth, written by experts in the field. The series aims to give readers a thorough understanding of the concepts, techniques, history and applications of artificial intelligence. Topics covered include machine learning, deep learning, neural networks, computer vision, natural language processing, robotics, ethics and more. The ebooks are written for professionals, students, and anyone interested in learning about the latest developments in this rapidly advancing field.
The artificial intelligence book series provides an in-depth yet accessible exploration, from the fundamental concepts to the state-of-the-art research. With over 200 volumes, readers gain a thorough grounding in all aspects of Artificial Intelligence. The ebooks are designed to build knowledge systematically, with later volumes building on the foundations laid by earlier ones. This comprehensive series is an indispensable resource for anyone seeking to develop expertise in artificial intelligence.

• Language: English
• Publisher: One Billion Knowledgeable
• Release date: Jun 21, 2023

Book preview

Chapter 1: Subsumption architecture

The subsumption architecture was widely used in the 1980s and 1990s and is a reactive robotic design closely linked to the field of behavior-based robotics. Rodney Brooks and his coworkers coined the phrase in 1986. Subsumption's impact on autonomous robots and other areas of real-time artificial intelligence has been substantial.

To counter classical artificial intelligence, the subsumption architecture was suggested, or GOFAI.

As opposed to using abstract concepts to direct actions,, subsumption architecture couples sensory information to action selection in an intimate and bottom-up fashion.: 8–12, 15–16

When compared to conventional AI, subsumption architecture takes a very different approach to the intelligence issue.

Disappointed with the results of programs attempting to mimic the human mind conceptually, such as Shakey the robot, Rodney Brooks began developing robots with an alternative theory of intelligence, mimicking the workings of the subconscious.

Instead of simulating intelligence by playing with symbols,, this approach is aimed at real-time interaction and viable responses to a dynamic lab or office environment.: 130–131

There were four main concepts that guided the mission::

Situatedness — A core tenet of situated AI is the concept of a human-like response time for a robot to its surroundings. According to Brooks, a mobile robot in a particular location shouldn't build its own symbolic representation of the environment and then act on that representation. He argues that instead of trying to model the environment, it is better to engage with it directly via well-designed perception-to-action sets since the reality is its own best model. However, at a low level, near the sensory signals, each module/behavior still models the environment. These basic models depend on immediate sensory input as much as possible rather than relying on memory to anticipate the behavior of the world, but they still require hardcoded assumptions about the environment that are contained in the algorithms themselves.

Brooks claims that there are two goals that may be met by creating an embodied agent. The first is that it mandates the development and testing of a fully physical control system, as opposed to theoretical models or virtual robots that may not perform as expected in the real world. The second is that by tying sense-data to meaningful behaviors, it may address the symbol grounding problem, a philosophical obstacle faced by many conventional AIs. The robot's perception of the external environment directly roots regress, and the robot's internal relations among its behavioral layers.

Intellect - Brooks believes that the development of perceptual and movement abilities is a required basis for human-like intelligence by looking at evolutionary progress. Furthermore, it seems that intelligence is defined by the dynamics of interaction with the environment by dismissing top-down representations as a feasible starting point for AI.

Conventional Emergence, Separate modules do not constitute intelligence.

The dynamic between these modules, assessed by taking into account the agent's surroundings, that is generally agreed upon as clever (or not).

Thus, Intelligence, is in the eye of the observer.: 165–170

A discussion over the nature of intelligence and the best way to advance robotics and artificial intelligence, including the principles discussed above, continues to this day.

A series of augmented finite-state machines (AFSM) processors make up each layer; these AFSMs include additional variables called instance variables that may store any programmed data structures. A layer is a subsystem that implements a certain activity, like wander about. Within and among these behavioral components, there is no unified command. All AFSMs are continually and asynchronously connected to the appropriate sensors and provide feedback to the corresponding actuators (or other AFSMs). If an input signal is not read before a new one arrives, the unread signal is deleted. It's typical