To ensure reliable and scalable communication among various modules in the proposed platform, a customized application-level communication protocol is developed on top of the FD-CAN layer. The system employs two independent FD-CAN channels, each configured in standard identifier mode (11-bit) and operated at a data rate of 5Mbps using Bit Rate Switching (BRS) for enhanced throughput. Automatic retransmission and transmit pause features were enabled to improve communication reliability under time-critical control conditions.

To support modular and extensible communication, the 11-bit standard identifier was structured into three subfields including message type, source node ID, and target node ID. The message type, occupying the upper three bits of the identifier, defines the functional category of the data being transmitted. This message type was categorized into four classes: emergency (EMCY), synchronization (SYNC), Service Data Object (SDO), and Process Data Object (PDO).

This proposed communication protocol provides strong platform-level scalability, as new devices can be easily integrated by assigning respective node IDs without modifying the protocol structure. Message priority is inherently managed through the identifier field, allowing critical messages to be transmitted with higher arbitration precedence. This flexible and modular design enables seamless expansion and reliable coordination across distributed modules.

For achieving precise position control and transparent force control, it is essential to compensate for the inherent nonlinear friction present in the drive system. MD provides a GUI-based framework that enables direct measurement of friction forces across a wide range of operating velocities, from low-speed bidirectional motion to high-speed regimes. Using these measurements, users can identify an appropriate nonlinear friction model tailored to the actuator’s characteristics. The resulting model can be incorporated as a feedforward compensator when the velocity reference is known, or applied as a feedback-based compensation strategy when it is not. Through the GUI, users can intuitively tune key parameters such as the smoothness of the compensator, the remaining level of linear damping, and the compensator gain, allowing for easy design, adjustment, and validation of a customized friction-compensation controller.

For robots that physically interact with the external environment, particularly with humans, it is often essential not only to maintain robustness but also to shape the interaction dynamics in a controlled and intuitive manner. MD supports this capability through a corridor-based impedance controller, which allows the system’s impedance to vary smoothly as the actuator state moves within a user-defined corridor. By configuring the corridor’s shape and parameters, users can tailor how stiffness, damping, or interaction forces evolve across different regions. Moreover, time-varying corridors can be implemented to provide adaptive interaction behaviors, enabling the robot to deliver natural, responsive, and context-appropriate assistance during human–robot interaction.

In practical robotic applications, actuators inevitably experience external disturbances that degrade tracking performance and hinder precise control. To address this, MD incorporates a disturbance observer (DoB) that estimates and compensates for these disturbances based on identified model in real time. Using experimentally obtained system identification data, a nominal actuator model is constructed, and the DoB injects an additional compensation input so that the actual plant behaves as closely as possible to this nominal model, even in the presence of disturbances. Through MD’s GUI, users can perform system identification, configure the disturbance observer, and select an appropriate Q-filter cut-off frequency, enabling the design of a robust and disturbance-rejection-capable control architecture.