The ENABLE Lab integrates dynamic systems, controls, biomechanics, neuroscience, and artificial intelligence to develop intelligent wearable robotic systems that enhance mobility, balance, and rehabilitation outcomes, particularly for older adults and individuals with neurological conditions. The lab’s work spans human–machine interaction, neuromuscular modeling, and reinforcement learning–based control, with the goal of advancing personalized, accessible, and effective technologies that promote healthy aging and independence.

Research directions

  • Dynamic system control
  • Assistive/rehabilitative robotics
  • Wearable robotics
  • Human-robot interaction
  • Neuromuscular modeling and control
  • Functional electrical stimulation
  • Human motion intent detection
  • Lyapunov-based nonlinear control/adaptive control
  • Machine-learning-based control
  • Surface electromyography/ultrasound imaging processing
  • Human movement biomechanics

Neuromuscular signals fusion-based finger movement prediction and control of prosthetic hands

1. Machine learning (transfer learning) for real-time hand gesture detection;
2. Online ultrasound imaging processing and analysis;
3. Human-machine-interface development for prosthetic hand in VR/DT system;
4. Real-time control of prosthetic hands in the physical world.

Modular cable-driven pelvic robot development for rehabilitation training

1. Mechatronics design of the platform;
2. Closed-loop assistance force trajectory control investigation;
3. Hierarchical control framework for personalization control;
4. Walking biomechanics analysis with and without waist assistance from the cable robotic device.

Reinforcement learning-based robotic assistance personalization control

1. Lightweight bilateral hip exoskeleton design with high wearability and comfort;
2. Closed-loop output torque trajectory tracking control;
3. Hierarchical control framework for assistance personalization with multiple control objectives;
4. Human/patient-in-the-loop experimental validation of the proposed device and control.

Hybrid neuroprosthesis for lower-limb rehabilitation training

1. Integration of the motorized tricycle and functional electrical stimulation (FES); 
2. Collaborative nonlinear control design for cycling speed regulation and tracking;
3. Adaptive optimal control (reinforcement learning) design for cycling speed regulation and tracking;
4. Model predictive control (MPC) design for cycling speed regulation and tracking.

Fused ultrasound imaging and sEMG-based control of a cable-driven ankle exoskeleton

1. Assist-as-needed control framework design by using adaptive impedance control;
2. Online sensor fusion between sEMG and ultrasound echo intensity;
3. Human volitional ankle plantarflexion torque prediction via Hill-type neuromuscular model with sensor fusion. 

Ultrasound Imaging-based Closed-Loop Control of Functional Electrical Stimulation for Drop Foot Correction

1. Sampled-data observer design;
2. FES delay compensation design;
3. Virtual constraint-based online reference trajectory generation;
4. PID-type dynamic surface control design.

​Neuromuscular model-based and model-free human ankle joint motion intent or muscle strength prediction

Neuromuscular features extraction from both surface electromyography (sEMG) and ultrasound imaging under isometric and dynamic conditions.

1. Temporal analysis of sEMG
2. Spatiotemporal analysis of ultrasound imaging
 – Gray-scale analysis;
 – Adaptive optical flow analysis;
 – Cross-correlation-based speckle tracking analysis.

Multiple robotic platforms for the development of control approaches

Robotic Baxter manipulator

Robotic Kinova arm 

Robotic quadcopters