Embedded Files
Journal manuscripts based on this research are currently in preparation. This portfolio presents a high-level overview; complete methods and experimental results will be shared following publication.
Project Summary
Project Summary
Over the past two years, I have worked on developing a lower-limb exoskeleton from the ground up for assistance and rehabilitation applications. This experience strengthened my understanding of research, systematic product development, mechanical design, control systems, manufacturing, system integration, and practical implementation.
Overview
Overview
Developed a bilateral lower-limb exoskeleton for rehabilitation and mobility assistance. The system combines mechanical design, embedded electronics, intelligent control, and machine-learning-based personalized gait generation.
Project details
Project details
Role: Graduate Research Assistant
Duration: 2 years
Institution: ARM Lab, Northern Illinois University
Application: Rehabilitation and walking assistance
Status: Prototype fabricated and experimentally tested
My Key Contributions
My Key Contributions
• Actuator mechanism design and integration
• Adjustable wearable structure development
• Electronics and safety-system integration
• Exoskeleton control-system development
• Personalized gait-trajectory generation implementation
• Prototype manufacturing and experimental validation
System Specifications
System Specifications
Configuration: Bilateral lower-limb exoskeleton
Active joints: Hip and knee joints
Passive joints: Ankle joints
Control: Personalized trajectory tracking
Table of Contents
Table of Contents
Lower Limb Exoskeleton
Lower Limb Exoskeleton
To begin, I will explain what a lower-limb exoskeleton is.
A lower limb exoskeleton (LLE) is a wearable robotic device that is designed to work in synchrony with the human lower body and aid the movements of the main lower limb joints (hip, knee and ankle).
The function of such devices is intended to augment, assist, or partially substitute lower-limb function during mobility-related tasks such as walking, stair negotiation, and sit-to-stand movement.
A significant number of people are suffering from mobility limitations. This increases the demand for assistive robotic devices. Spinal cord injury, traumatic brain injury, stroke and neuromuscular problems can significantly impair independent walking and diminish quality of life.
Human Biomechanics
Human Biomechanics
Before beginning the development of an exoskeleton, it is essential to understand the biomechanics of the human body because the device is designed to be worn by and interact directly with the user.
Human biomechanics examines how bones, joints, muscles, and external forces work together to produce movement. Human motion occurs in three anatomical planes: sagittal, frontal, and transverse. Walking primarily takes place in the sagittal plane.
The human lower body has far more than six anatomical degrees of freedom, where the hip and knee perform flexion and extension, while the ankle performs dorsiflexion and plantarflexion, which produces the forward walking movement.
The gait cycle is what explains how a human should walk and classifies all the stages. It begins when one foot contacts the ground and ends when the same foot contacts the ground again. It consists of two main phases: the stance phase, which accounts for approximately 60% of the cycle and supports the body’s weight, and the swing phase, which accounts for approximately 40% and advances the leg forward.
Actuator Mechanism Design
Actuator Mechanism Design
The first step in the development process was to design and fabricate the actuator mechanism. Several actuation approaches were considered based on the exoskeleton’s objectives. Passive mechanisms typically use components such as springs and dampers to store, release, or dissipate energy, whereas active mechanisms use powered systems such as electric motors or hydraulic actuators to generate joint movement.
For this project, an electric motor coupled with a gear-drive mechanism was selected as the most suitable solution because it provided the required torque, speed, control, and compactness to meet the exoskeleton’s design objectives.
Dynamic Model
Dynamic Model
Determining the actuator specifications required an analysis of the torque, speed, and power demands of both the user and the exoskeleton. For this purpose, a three-degree-of-freedom (3-DoF) planar dynamic model of the exoskeleton system was developed, as shown in the figure. The equations of motion were derived using the Euler–Lagrange method while accounting for the effects of ground reaction forces.
Final Actuator Assembly
Final Actuator Assembly
The final CAD actuator has a reported total axial length of 94 mm and a mass of approximately 1.6 kg. The disk-shaped motor and harmonic drive are arranged concentrically, while the bearing and lower-ear structure create the output support. The result is a self-contained joint module that can be repeated across the bilateral hip and knee joints with different transmission ratios.
Exoskeleton Structure and Human-Interface Design
Exoskeleton Structure and Human-Interface Design
The next engineering problem was to transform those four compact actuator modules into a wearable lower-limb structure that could transmit joint torque to the user while remaining adjustable, structurally safe, and practical to assemble. This section focuses on the mechanical architecture surrounding the actuators: telescopic leg links, the waist/back support, user interfaces, structural analysis, and full-system integration.
Adjustable Link Design
Adjustable Link Design
The leg link was designed as a circular telescopic tube assembly fabricated primarily from 6063 aluminum telescoping round tubes. The assembly contains seven functional components and provides an adjustable length from 360 mm to 460 mm. This 100 mm adjustment range allows the joint-to-joint spacing to be adapted to the wearer rather than forcing the wearer to conform to one fixed frame geometry. The reported total link mass is approximately 650 g.
Back-Support Design
Back-Support Design
The bilateral leg assemblies require a common upper structure at the pelvis. The back-support system was developed to provide a rigid connection between the hip actuator assemblies and the user’s waist/back, maintain overall alignment, and transmit loads between the two sides of the exoskeleton and the torso.
Human Interface
Human Interface
A mechanically strong frame is not sufficient for a wearable robot unless actuator torque can be transferred comfortably to the body. The leg interface therefore uses knee braces connected to the exoskeleton links through custom adjustable 3D-printed mounts. These mounts allow the brace position to be adapted to the wearer and connect the rigid exoskeleton link to a broader, softer contact area on the leg.
Ankle–Foot Assembly
Ankle–Foot Assembly
A passive ankle–foot mechanism was developed to support the exoskeleton’s weight and transfer loads safely to the ground. The ankle provides one degree of freedom, allowing plantarflexion and dorsiflexion instead of restricting the joint with a fixed connection. The foot assembly integrates force-sensitive resistors (FSRs), which work with the IMUs to detect the user’s movement intention and gait phase. The rigid foot components were fabricated from PLA for prototype evaluation, while flexible TPU components, shown in blue, were incorporated to improve comfort and absorb impact.
Complete Exoskeleton
Complete Exoskeleton
Personalized Gait Trajectory Generation
Personalized Gait Trajectory Generation
A key innovation of this project is the use of machine learning to generate personalized gait trajectories from user-specific anatomical parameters. The model was trained using gait data from 245 participants and generated personalized hip and knee trajectories. On the test dataset, the predicted trajectories achieved joint-angle RMSE values ranging from approximately 1.5° to 3.0°.
Control System Design
Control System Design
Master Control Unit: NVIDIA Jetson Orin Nano
Master Control Unit: NVIDIA Jetson Orin Nano
Executes the high-level control program
Processes sensor and trajectory data
Communicates with the motor controllers
Motor Controller: Maxon EPOS4
Motor Controller: Maxon EPOS4
Interfaces between the Jetson and EC motors
Drives the motors at each active joint
Quantity: 4 controllers
Control Strategy
Control Strategy
Test Setup
Test Setup
The test setup provides a safe and stable environment for evaluating the exoskeleton during walking. It consists of a treadmill, an 8 ft × 6 ft × 6 ft aluminum support frame, and ratchet straps that suspend and secure the exoskeleton during testing.
Results
Results
Human-subject testing demonstrated that the exoskeleton successfully tracked the reference gait trajectories throughout the walking cycle. The hip joints achieved high tracking accuracy, with RMSE values of 1.30° for the right hip and 1.27° for the left hip. The knee joints showed slightly larger deviations during rapid flexion and extension, with RMSE values of 4.39° and 3.74° for the right and left knees, respectively. Overall, the results demonstrate stable and consistent trajectory tracking across all four actuated joints.
Project Conclusion
Project Conclusion
This project resulted in the successful development and preliminary validation of a bilateral lower-limb exoskeleton for assistance and rehabilitation research. The final prototype integrates compact actuators, an adjustable wearable structure, embedded electronics, personalized gait generation, and real-time joint control into a single functional system. Human-subject testing demonstrated accurate hip and knee trajectory tracking, confirming the feasibility of the proposed design and control framework. The system now provides a research platform for further multi-subject testing, control refinement, and future rehabilitation studies.
Page updated
Google Sites
Report abuse