Autonomous Systems Overview

Principles

Definition

Autonomous systems is an interdisciplinary research field focused on the principles, design, and analysis of entities capable of operating independently in dynamic environments. It investigates the methodologies for creating systems that perceive, decide, and act autonomously, characterized by properties such as self-governance, adaptability, and decision-making under uncertainty, holding significance for advanced automation and operation in complex or inaccessible settings.

Ontological type

Core Methods

Verification and Validation

Architectures

Key developments

Key Figures

Cooperative Real-Time Autonomy era

Steven M. LaValle [1] is a central figure in cooperative real-time autonomy during this era, with affiliations at Stanford University [3] and Université Joseph Fourier [4]. The Randomized Kinodynamic Planning paper [6] introduced a scalable, sampling-based approach to kinodynamic motion planning that enabled real-time control under dynamic constraints, shaping how autonomous systems manage concurrent objectives in uncertain environments. James Kuffner [2], associated with Stanford University [3] and Carnegie Mellon University [5], contributed to these ideas by promoting kinodynamic planning concepts in real-time autonomy contexts. His collaborative work on randomized kinodynamic planning [6] helped advance robust, real-time motion generation for multi-robot and mobile platforms operating under uncertainty in this era.

Integrated Perception-Planning Systems era

Sebastian Thrun [1] is associated with Stanford University [3] and the University of Pittsburgh [4] during the Integrated Perception-Planning era. His key contributions in this era include Stanley: The robot that won the DARPA Grand Challenge [7] and Towards fully autonomous driving: Systems and algorithms [8], which collectively advanced end-to-end perception–planning pipelines and catalyzed practical demonstrations of autonomous driving. Roland Siegwart [2] is affiliated with École Polytechnique Fédérale de Lausanne [5] and ETH Zurich [6] during this era. His contributions include Keyframe-based visual–inertial odometry using nonlinear optimization [9] and Monocular‐SLAM–based navigation for autonomous micro helicopters in GPS‑denied environments [10], advancing robust state estimation and navigation for integrated perception–planning systems.

Planning-First Autonomy era

Daniela Rus[1] has associations with Massachusetts Institute of Technology[3] and Vassar College[4] during the Planning-First Autonomy era. Her key contributions in this era include the paper Planning and Decision-Making for Autonomous Vehicles[7], which advanced planning-centric autonomy by integrating decision-making with trajectory generation for on-board real-time systems. Shinpei Kato[2] has associations with Carnegie Mellon University[5] and University of Illinois Urbana-Champaign[6] during the Planning-First Autonomy era. His key contributions include Autoware on Board: Enabling Autonomous Vehicles with Embedded Systems[8], which propelled embedded-system deployments for autonomous vehicles and bridged perception, planning, and control in hardware-constrained settings.