Connection between the JETSON and the T-motors with CAN transceiver using Dupont ConnectorsConnection with CAN transceiver.
Connection of the body wiring in the hip exoskeleton.
The overarching goal of this research is to develop a highly effective torque control strategy for a hip exoskeleton. This strategy aims to achieve a dual objective: first, to minimize the metabolic cost for the user during locomotion, thereby improving energy efficiency and reducing user exertion; and second, to maximize user preference by ensuring the exoskeleton assistance feels natural, comfortable, and effectively supports their intended movements. Achieving this balance is crucial for creating wearable robotic systems that are not only energy-efficient but also seamlessly integrated and accepted by users, promoting wider adoption and real-world impact in areas such as rehabilitation, mobility assistance for elderly or individuals with disabilities, or even human performance augmentation in industrial/logistics settings. This goal necessitates a deep understanding of biomechanics, robust data acquisition, and advanced control algorithms.
A critical challenge identified in the initial phase of this research was suboptimal marker placement for motion capture data acquisition. Specifically, the initial marker set lacked a strategically positioned marker on the thigh segment. This omission proved to be a significant limitation because it resulted in inaccurate estimation of thigh segment kinematics, leading to errors in joint angle calculation (specifically hip and knee angles), and consequently, compromised the fidelity of the motion capture data used for offline training of the torque control model. Accurate motion capture data is foundational for developing effective biomechanical models and control algorithms. Therefore, addressing this marker placement issue was a prerequisite for achieving reliable and physiologically relevant torque control.
This phase of the research heavily relied on the following skills:
🌟Biomechanics Principles: Applied a strong understanding of human biomechanics, particularly lower limb kinematics, joint angle definitions, gait cycle analysis, motion capture data acquisition principles, and torque calculations. This knowledge was essential for identifying the limitations of the initial marker placement and understanding its impact on data accuracy.
🌟Motion Capture Data Processing and Cleaning: Developed proficiency in processing and cleaning motion capture data. This involved using motion capture software (Qualisys Motion Capture Track Manager Software), identifying and addressing common motion capture artifacts (e.g., marker swaps, gaps), filtering noisy data, and implementing custom algorithms for data sorting and cleaning as detailed in "My Contribution". Ensuring data quality is paramount for reliable biomechanical analysis and model training.
Recognizing the limitations imposed by the initial marker placement, my primary contribution was to address the suboptimal marker configuration proactively. This involved two key actions:
1. Marker Placement Optimization: Based on biomechanical principles and literature, Marker Placement Reliability and Objectivity for Biomechanical Cohort Study by Malus J and others, I identified the need for and implemented the addition of a strategically placed marker on the thigh segment. This refined marker set significantly improved the accuracy of thigh segment kinematics capture.
2. Development of a Data Sorting and Cleaning Algorithm: To further enhance data quality and address potential artifacts or noise inherent in motion capture data, I designed and implemented a custom data sorting and cleaning algorithm. This algorithm coded in Python interpolation methods, filtering techniques, outlier detection methods within the algorithm]. This algorithm ensured that the motion capture data was of sufficient quality for subsequent biomechanical analysis and model training.
Enhanced Hip Exoskeleton Efficiency through Electrical System Optimization.
Excessive Circuit Wiring Weight Limiting Exoskeleton Mobility and Efficiency.
Redesigned and simplified exoskeleton wiring harness, achieving a 30% circuit weight reduction through optimized routing and component selection.
Significantly Improved Exoskeleton Mobility and Responsiveness.
Connection between the JETSON and the T-motors with CAN transceiver using Dupont ConnectorsConnection with CAN transceiver.
Connection of the body wiring in the hip exoskeleton.