Design and Validation of a Low-Level Controller for Hierarchically Controlled Exoskeletons

Area of Science:

• Robotics
• Control Systems Engineering
• Biomedical Engineering

Background:

• Hierarchically controlled exoskeletal systems often separate hardware (sensor collection, motor control) from complex high-level algorithms.
• Series Elastic Actuators (SEAs) are crucial components in these systems, requiring efficient low-level control for sensor feedback and actuator commands.
• Existing solutions may lack generalized hardware for real-time data acquisition and synchronization in complex robotic systems.

Purpose of the Study:

• To present a generalized low-level controller for sensor collection, motor input, and networking with high-level controllers in exoskeletal systems.
• To detail the hardware design of printed circuit boards (PCBs) for conditioning sensor feedback from SEA subsystems and inertial measurement units (IMUs).
• To provide an open-source solution for researchers in robotics, enabling direct data acquisition and low-level device control.

Main Methods:

• Designed two PCB prototypes for collecting and conditioning sensor data, including joint/motor encoders, motor current, and force sensors from SEAs.
• Implemented a generalized low-level controller capable of real-time sensor feedback processing and actuator command transmission.
• Developed a detailed networking strategy for high-level and low-level controller communication, including data frame structure and synchronization protocols.

Main Results:

• Successfully validated the low-level controller using a pendulum test bed with comprehensive sensor feedback, including IMU data.
• Demonstrated the controller's capability to handle real-time data acquisition and synchronization between high-level and low-level systems.
• The developed hardware and software designs are made available as open-source resources.

Conclusions:

• The presented generalized low-level controller effectively addresses the hardware requirements for sensor collection, motor control, and networking in hierarchically controlled robotic systems.
• The open-source nature of the design facilitates adoption and further development in various robotic applications, including humanoid robots, rehabilitation devices, and surgical robots.
• This work provides a foundational solution for real-time data acquisition and control in systems utilizing Series Elastic Actuators.