4 a2

Introduction

  • AI Robotics Overview

    • Exploring the necessity and hierarchy of intelligence in robots.

Learning Objectives

  • Definitions and Design Process

    • Apply Levis' definition of architectures in intelligent robot software organizing.

    • Familiarize with layers in a canonical operational architecture: behavioral, deliberative, and interface.

    • Describe the four primitives of robotic intelligence: SENSE, PLAN, ACT, LEARN.

    • Recognize vulnerabilities leading to failures in canonical operational architecture.

Architectures Overview

  • Types of Architectures

    • Operational Architecture: Describes what the system does at a high level.

    • Systems Architecture: Details how major subsystems of a system work.

    • Technical Architecture: Covers implementation details including programming languages.

Importance of Organizational Structure

  • Software Architecture

    • An overall design style termed "architecture" that dictates control system organization and interaction among components.

    • Supports software organization and influences problem-solving approaches.

Processing Architecture

  • Three Tier Hybrid Architecture

    • Characteristics of operational, biological, and process architectures.

  • Primitives

    • Enable robots to SENSE, PLAN, ACT, and LEARN.

Reaction Type Framework

  • Behavioral Robotics

    • Instinctive SENSE-ACT couplings without a planning phase.

    • Many behaviors acting on concurrent stimuli.

    • Fast and modular but limited predictive capability and coordination challenges.

Learning and Adaptation

  • LEARN

    • Incorporates diverse forms of learning, capable of enhancing the architecture.

Functional Considerations

  • Functional Observations

    • Factory automation functions like generating, implementing, and monitoring require integrating primitives.

    • Struggles exist in knowledge structures and implementation due to monolithic routines.

Layered Programming Benefits

  • Programming Styles

    • Different layers allow for specialization in programming languages suited for each task type, improving modularity.

    • Enhances reusability and adaptation capabilities, accommodating evolving requirements.

Decision Making and Models

  • Planning Horizons and Time Scales

    • Establishing the type and detail of models (local or global), dictates algorithm updates and decisions.

Conclusion and Future Considerations

  • Autonomy Limits

    • Questions arise on the trustworthiness of robot autonomy and its interaction critically with human oversight.

  • The evolving programming capabilities aim to enhance robot autonomy and improve function efficiency.