The Mechatronic Arc: Bridging the Gap Between Physical Design and Computational Intelligence

About the Workshop

This workshop is organized in conjunction with the IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM 2026), which will be held in Genova, Italy, from July 7th to 10th, 2026. For registration and details about the main conference, please visit the official AIM 2026 Website.

The design philosophy underlying modern robotics is not monolithic; rather, it evolves dynamically in relation to the scale and complexity of the problem addressed. Advanced robotic systems rely on a continuous dialogue between their physical embodiment and their algorithmic intelligence. On one hand, intelligence does not reside solely in software but is encoded within the robot’s mechanics—a form of Embodied Intelligence that defines intrinsic capabilities in perception, action, and safety. On the other hand, the Algorithmic Intelligence of software expands these abilities, enabling versatility and complex behaviors that cannot be realized through matter alone.

In robotics, a fundamental dichotomy is often faced: should a specific function be delegated to the hardware or to the software?

This workshop proposes to explore this duality by tracing a "Mechatronic Arc," a narrative path that guides participants from a paradigm of component-level trade-offs to one of system-level synthesis. By fostering a direct comparison between leading researchers who approach similar challenges from opposing paradigms, the workshop is structured into two distinct Acts:

Act I

The Functional Dichotomy – Matter vs. Math

In the realm of component design, the engineer is frequently tasked with selecting the optimal tool for a singular purpose. This section investigates the functional trade-offs where physical laws and control laws compete for supremacy, debating whether the task is better achieved through material properties or mathematical models.

We will analyze how contact perception can be achieved through dedicated sensorization versus software estimators; how compliance can be an intrinsic material property versus a control-simulated behavior; and how safety can be guaranteed by passive mechanical constraints versus active algorithmic barriers.

Sessions in Act I:

Session 1 Sensing – Physical Transduction vs. Virtual Estimation Session 2 Actuation – Intrinsic Compliance vs. Variable Impedance Session 3 Safety – Mechanical Design vs. Algorithmic Supervision

Act II

The Necessary Symbiosis – When Bodies and Controllers Co-Evolve

Transitioning to high-complexity systems, the separation between “controller” and “plant” becomes an obstacle rather than a design choice. This section illustrates that for sophisticated robots—and especially for human-centered mechatronic devices—intelligence cannot simply be programmed “into” a machine but must be physically grounded “within” its morphology and interfaces.

By examining robot body co-design and rehabilitation/assistive robotics, it becomes evident that the most ambitious challenges cannot be solved by choosing between hardware and software, but only through their profound symbiosis. In these domains, mechanical architecture, physical coupling, and control policies must evolve jointly.

Sessions in Act II:

Session 4 Rehab – The Human–Machine Symbiosis Session 5 Body – Co-Designing Robots

Program Timetable

Time	Session / Activity	Slot Details
Act I: The Functional Dichotomy - Matter vs. Math
11:15 – 11:25	Welcome & Introduction The Mechatronic Arc: Bridging the Gap Between Physical Design and Computational Intelligence	10 min
11:25 – 12:00	Session 1: Sensing - Physical Transduction vs. Virtual Estimation ▼ F. Masiero (Click to view details)	35 min
Session Overview This session questions the necessity of complex hardware for perception. On one side, the Embodied Sensing approach advocates for the integration of dedicated, high-fidelity sensory hardware—such as distributed tactile surfaces or specialized force/torque sensors—to grant robots a direct, physical sensitivity to their environment. On the opposing side, the Algorithmic Sensing perspective champions the use of advanced observers and data-driven techniques to transform the robot’s existing actuation system into a "virtual sensor." This approach focuses on estimating external forces and contacts through proprioception and mathematical models, eliminating the need for additional, often fragile, instrumentation. Full Perspective (HW & SW) Federico Masiero “New Approaches to Magnetic Methods for Robotics” 🕒 30 min + 5 min Q&A
12:00 – 12:50	Session 2: Actuation - Intrinsic Compliance vs. Variable Impedance ▼ M. Garabini \| M. Saveriano (Click to view details)	50 min
Session Overview Adaptability to the environment is crucial, but where should it originate? The Passive Compliance perspective focuses on designing actuation systems with intrinsic elasticity—ranging from soft fluidic actuators to mechanically variable stiffness mechanisms. In this view, compliance is a physical property inherent to the device, ensuring passive safety and energy efficiency at the source. In contrast, the Active Compliance approach (Variable Impedance Control) relies on stiff, high-bandwidth actuators governed by control loops that simulate virtual springs and dampers, offering programmable versatility that allows the robot's behavior to be radically altered without mechanical modification. Hardware (HW) Manolo Garabini “Modular Multimodal Robots: the mechatronics behind the next robot generation” 🕒 20 min + 5 min Q&A Software (SW) Matteo Saveriano “The Role of Energy and Geometry in Contact-Rich Manipulation” 🕒 20 min + 5 min Q&A
13:00 – 14:00	🍽️ Lunch Break & Poster Session	60 min
14:00 – 14:35	Session 3: Safety - Mechanical Design vs. Algorithmic Supervision ▼ R. Kirschner (Click to view details)	35 min
Session Overview When humans and robots share space, safety is non-negotiable. The Passive Safety perspective emphasizes intrinsic mechanism design, advocating for lightweight structures, mechanical joint limits, and impact-mitigating geometries. This philosophy relies on physical constraints to ensure that injury is physically impossible or drastically mitigated, providing a fail-safe layer that persists even in the event of a total control failure. Conversely, the Active Safety approach employs formal methods and real-time algorithmic filters (such as control barrier functions, reachable set analysis or safety filters). This methodology wraps the system in a "mathematical safety shield," allowing high-performance hardware to operate aggressively while providing certifiable guarantees that unsafe states remain unreachable via continuous software monitoring. Full Perspective (HW & SW) Robin Kirschner “Certifiable Safe Human–Robot Interaction Starts with Holistic Physical Design” 🕒 30 min + 5 min Q&A
14:35 – 15:10	Roundtable Act I (featuring Special Guest Oussama Khatib) Act I: The Functional Dichotomy - Matter vs. Math	35 min
15:10 – 16:00	Session 4: Rehab - The Human–Machine Symbiosis ▼ E. Trigili \| F. Tessari (Click to view details)	50 min
Session Overview The final session addresses the ultimate integration: the merging of robotic systems with the human body. In rehabilitation and assistive robotics, hardware and software must not only coexist but cooperate intimately with a biological partner, where the human becomes an integral component of the system dynamics. The Hardware Perspective focuses on the mechatronic design of exoskeletons and prosthetic devices. It argues that the physical interface—including kinematic compatibility, back-drivability, compliance, and ergonomic actuation—is a primary determinant of interaction quality. Here, “intelligence” is embodied in the mechanism’s ability to be transparent or supportive through its physical structure, providing an intrinsic baseline of safety and usability even under imperfect control. Conversely, the Software Perspective centers on Human-in-the-Loop control and intention estimation. This approach tackles the algorithmic challenge of inferring the user’s intent (via EMG, EEG, or interaction forces) and modulating assistance in real time (“assistance-as-needed”). Hardware (HW) Emilio Trigili “Engineering human-robot interaction in upper-limb exoskeletons” 🕒 20 min + 5 min Q&A Software (SW) Federico Tessari “Complexity is a Feature, not a Bug: Rethinking Human Coordination for Computationally Intelligent Machines” 🕒 20 min + 5 min Q&A
16:00 – 16:30	☕ Afternoon Coffee Break	30 min
Act II: The Necessary Symbiosis - When Bodies and Controllers Co-Evolve
16:30 – 17:20	Session 5: Body - Co-Designing Robots ▼ M. Russo \| M. Trumić (Click to view details)	50 min
Session Overview In high-dimensional robots such as soft robots, the distinction between the "plant" and the "controller" becomes blurred. The Hardware Perspective focuses on advanced mechatronic architecture, illustrating how the physical design—including mass distribution, kinematic structure, actuator placement and mechanical stiffness—fundamentally shapes the robot's capabilities and pre-defines the solution space for the control problem. Complementing this, the Software Perspective focuses on high-level controllers such as data-driven (soft robots) or whole-body control (humanoids and quadrupeds) and planning, demonstrating how algorithms must orchestrate the massive kinematic and dynamic redundancy provided by the hardware to achieve robust locomotion, balance, and manipulation in unstructured environments. Hardware (HW) Matteo Russo “Robot design: Reclaiming hardware in the age of AI” 🕒 20 min + 5 min Q&A Software (SW) Maja Trumić “Tackling control challenges in soft robotics through model-based approaches” 🕒 20 min + 5 min Q&A
17:20 – 17:45	Roundtable Act II Act II: The Necessary Symbiosis - When Bodies and Controllers Co-Evolve	25 min
17:45 – 17:55	Closing Remarks Wrap-up and farewells	10 min

The Mechatronic Arc:

Bridging the Gap Between Physical Design and Computational Intelligence

About the Workshop

The Functional Dichotomy – Matter vs. Math

Sessions in Act I:

The Necessary Symbiosis – When Bodies and Controllers Co-Evolve

Sessions in Act II:

Organizers

Speakers

Special Guest

Program Timetable

Session Overview

Federico Masiero

Session Overview

Manolo Garabini

Matteo Saveriano

Session Overview

Robin Kirschner

Session Overview

Emilio Trigili

Federico Tessari

Session Overview

Matteo Russo

Maja Trumić