Home | AIMRL

Director’s Message: A Vision for Integrated Intelligence

For over forty years, the Advanced Intelligent Mechatronics Research Laboratory (AIMRL) has operated at the intersection of perception, actuation, and bio-inspired design. Our work is driven by a singular, unified vision: the transition from treating sensors and actuators as separate, add-on components to a paradigm of Integrated Intelligence.

In this framework, perception and actuation are not just connected; they are physically and functionally coupled. By treating physical fields—electromagnetic, optical, and mechanical—as high-dimensional information sources, we enable machines to "see" through materials, reconstruct complex geometries, and adapt to the compliant reality of living systems. Whether automating biosecure food production or designing anatomically-informed exoskeletons for stroke recovery, our goal remains the same: creating a seamless interface between machine logic and the physical world.

Professor Kok-Meng Lee, Ph.D.

The Emergence of Intelligent Mechatronics (1985–)

Agricultural Automation & Biosecurity (1995–)

Intelligent Manufacturing & Field Perception (2005–)

Human-Centered Design & M3C (2010–)

Mission: Integrated Human-Like Intelligence

Our mission is to redefine the relationship between machine logic and the physical world. We advance a paradigm of Integrated Intelligence, where perception and actuation are no longer separate components, but are physically and functionally coupled within a single coherent architecture. By treating multi-physics fields as high-dimensional information sources, we strive to create machines that interact with living systems and industrial environments with the grace, safety, and adaptability of biological entities.

Research Vision

AIMRL’s research follows a unifying progression:

Signals & Systems → Information → Knowledge → Human-Like Intelligence

This progression reflects a fundamental paradigm shift—from point sensing to field-based perception, from isolated measurements to physics-guided inference, and from data processing to intelligent decision-making and control.

A Unified Paradigm for Sensing and Motion

AIMRL’s research advances a paradigm where human-like intelligence is embedded into the very architecture of the machine. This philosophy has evolved into three distinct pillars of innovation

Pillar I (Field)

Multi-Physics Field-Based Sensing (9)

Reconstructing complex states from EM, Optical, and Mechanical fields(12).

Pillar II (System)

Integrated Multi-DOF Actuation System (10)

Transitioning from cascaded joints to unified spherical architectures(13)

Pillar III (Intelligence)

Multi-Physics Field-Based Sensing (9)

Bridging machine logic with the compliant reality of living systems(14)

Core Contributions: Transforming Society through Integrated Intelligence

The legacy of AIMRL is defined by three paradigm-shifting shifts in mechatronic research, each shaped by the critical societal challenges of the 20th and 21st centuries:

Field-Based Machine Perception & Biosecurity: We advanced sensing from discrete point measurements to the high-dimensional field reconstruction of physical states from electromagnetic, optical, and mechanical fields. Originally driven by post-2001 biosecurity needs, this technology enabled the "soft-touch" automation required for gentle, reliable handling of fragile biological subjects.

Resilient Systems & Infrastructure Safety: We transitioned mechatronics from cascaded mechanical joints to unified, multi-DOF spherical motor architectures. In response to the 2007 I-35W bridge collapse, this research enabled Flexonic Mobile Sensing Nodes that traverse complex steel geometries to deliver high-fidelity perception for structural health monitoring and resilience.

Human-Centered M3C & Clinical Support: We established the Mind-Motor-Motion-Control (M3C) paradigm to synchronize human intent with robotic assistance. By integrating Spine-Equivalent-Beam (SEB) models, pantographic exoskeletons, and physics-informed neural networks, we provide co-adaptive, home-based rehabilitation—addressing caregiver strain and biosecurity risks highlighted during the COVID-19 pandemic.

AIMRL at a GLANCE

Field-Based Machine Perception

beyond conventional point sensing

Physics-Guided Learning

combining first principles with data-driven intelligence

Human-Centered Mechatronics

for rehabilitation, assistance, and interaction

Multi-DOF Actuation & Robotics

Robotics with embedded sensing and control.

Hierarchical Clustering

FIELD

Physical fields, physical laws, and distributed phenomena

Optical / Vision Fields

Image formation and projection geometry
Optical flow, structured light, interferometry
Color, illumination, and reflectance fields
Vision–sensor interaction models

Thermal / Strain / Impedance Fields

Temperature fields in machining and processes
Strain, stress, and impedance distributions
Coupled multiphysics field interactions

Electromagnetic / Magnetic / Eddy-current Fields

Distributed current source (DCS) & multipole models
Motion-induced, boundary, and end-effect fields
Magnetic dipole / Lorentz-force formulations
Field observability, sensitivity, identifiability

Mechanical Deformation & Compliance Fields

Beam, plate, flexure, and shell deformation fields
Large-deflection and nonlinear compliance
Bio-inspired anatomical structures
Contact, cutting, and interaction mechanics

SYSTEM

Integrated mechatronic realizations built on fields

Field-based Sensing Systems

Embedded magnetic / eddy-current sensing
Optical, strain, impedance-based sensors
Distributed and mobile sensing architectures
Sensor fusion and calibration systems

Manufacturing & Process Monitoring Systems

Machining vibration, temperature, and force systems
Thin-wall and duplex machining monitoring
Structural health monitoring (SHM)
Automated inspection and defect detection

Multi-DOF Actuation Systems

Spherical motors, VR actuators, magnetic bearings
Tiltable stages and micro-positioning platforms
Variable-stiffness and compliant actuators
Co-designed actuation–sensing systems

Biomedical / Human-centered Systems

Rehabilitation, exoskeletons, and assistive robots
Wearable and anatomical sensing systems
Biomechanics-based devices
Human–machine physical interfaces

INTELLIGENCE

Inference, control, learning, and decision layers emerging from systems

Estimation & Reconstruction Intelligence

Inverse field problems
State, parameter, and geometry estimation
Sparse and distributed sensing inference
Field-to-information reconstruction pipelines

Physics-guided Learning

Hybrid analytical–data-driven models
Physics-informed neural networks (PINNs)
Model-constrained learning architectures
Data-efficient learning from physical fields

Model-based Control & Optimization

Field-aware and physics-consistent control
Adaptive, optimal, and back-stepping control
Performance–robustness tradeoffs
Multi-DOF coordinated control

Human-centered Intelligence

Biomechanical parameter inference
Perception-driven assistance and adaptation
Decision-support for rehabilitation and interaction
Human-in-the-loop intelligent systems

AIMRL