
Author: Jenny He
Senior Systems Engineer, Hydrix
Translating regulatory recommendations into engineering practice
When a surgical device incorporates a robotic arm, the design challenge extends beyond reaching the intended position accurately. How will the arm move safely? What happens if communication is lost or the operator loses control? Developers need to answer these questions and demonstrate that the complete system performs as intended, with its risks adequately controlled.
In this article, I explore how developers can translate relevant recommendations from FDA’s new draft guidance on robotically-assisted surgical devices into practical engineering decisions. Using communication loss as an example, I show how a safety expectation can shape requirements, system architecture and the testing needed to demonstrate the complete system’s response.
Issued on 25 September 2026, Robotically-Assisted Surgical Devices – Premarket Submissions provides recommendations on non-clinical testing, clinical data and labeling for certain robotically-assisted surgical devices (RASDs). FDA’s stated aim is to promote consistency and facilitate efficient review of submissions. The guidance remains a draft for comment and contains non-binding recommendations.
Although its scope is limited to certain RASDs, some of its engineering principles may also offer useful insights for other medical devices with robotic features. The opportunity is to consider relevant recommendations early, while they can inform the design and help teams plan the evidence they will need.


Assesses the robotic arm’s movement and positioning, observing how the system behaves during testing.
Looking at the complete system
A robotic feature depends on mechanical, electronic and software elements working together to turn user input or a subsystem command into controlled movement. Its safety and performance depend on those interactions. The draft guidance recommends testing subassemblies and the complete system, including relevant accessories and compatible devices. This raises a practical question: what needs to be demonstrated at each level and in each configuration?
Testing an actuator can establish how that component performs. It does not, by itself, show how the arm behaves while carrying a tool or payload and responding to commands through the complete control chain.
Taking a system view helps the team define boundaries and interfaces, assign responsibilities and decide what to verify at each level. Component and subsystem tests provide supporting evidence. Complete-system testing establishes whether the integrated design meets the specified performance requirements.
From recommendations to engineering practice
A useful starting point is to ask what each relevant recommendation means for the device being developed. Which functions and use scenarios does it concern? What assumptions need to be resolved? What evidence would support the performance claim?
The answers help the team define testable system requirements, with justified limits and clearly specified conditions. The system architecture then sets out how those requirements will be met and assigns responsibilities across mechanical, electrical and software design.
Risk analysis informs these decisions throughout development. It helps identify how faults or degraded performance could create hazardous situations, which controls are needed and which conditions require particular attention during testing.
Verification planning establishes how the team will demonstrate that the requirements and risk controls have been met. It identifies the test methods, configurations and system levels needed to build that evidence. Validation confirms that the device meets user needs and its intended use.
This is the systems engineering approach we bring to development at Hydrix. We connect requirements, architecture, risk management and verification and validation, so that design decisions and the evidence supporting them develop together.
A lost connection, what does “fail-safe” mean?
The FDA draft guidance recommends fail-safe operation when teleoperation faults occur. But what does that mean for a particular device? The protective response needs to address the fault without introducing unacceptable risks of its own.
Consider a loss of communication between the main controller and the arm controller while the arm is moving. If the arm continues to execute the last received command, the operator may lose effective control, exposing the patient to unintended movement or excessive force. Yet simply stopping the arm or removing power can also introduce hazards. The following scenarios illustrate why the response needs careful consideration.
Hazard
Foreseeable sequence of events
Hazardous situation
Mechanical energy from device motion
Communication is lost; the fault is not addressed promptly; the arm continues executing the last received command
The patient is exposed to unintended device movement that the operator cannot stop using the normal controls
Mechanical forces during deceleration
Communication is lost; the protective response stops the arm too abruptly.
The patient is exposed to excessive stopping forces transmitted through the device.
Gravity
Communication is lost; drive power is removed; braking or support is insufficient; the arm or load falls.
The patient or operator is exposed to falling components or crushing forces.
Examining these scenarios helps the team assess the risks and select appropriate controls. Those decisions define what fail-safe operation means for this device: how the arm should respond, and which safe condition it must reach and maintain.

Clinical observation of a surgical robotic system, bringing engineering development into the context of its intended operating environment.
Turning the protective response into requirements
“Enter a safe state” is not yet a testable requirement. What counts as invalid communication? How quickly must the fault be detected and the protective response completed? What limits apply to residual movement, stopping forces and the ability to hold the load? The system requirements need to answer these questions.
For example, a system requirement could state:
If valid control communication is unavailable for more than [T milliseconds], the system shall perform [the defined protective response] within [T_response milliseconds] after fault detection and maintain [the specified safe condition] under [defined operating conditions], without exceeding [specified movement and force limits].
The combined detection and response time must stay within the overall limit established for the protective response. Requirements also need to specify how the user is notified and what must happen before motion can resume, including whether a deliberate user action is required.
Assigning and implementing the risk controls
These requirements shape the architecture. The team needs to decide how communication loss will be detected and how the arm will stop and hold its load without further commands from the main controller. The protective response must remain available when the communication link fails.
The architecture provides the basis for allocating system requirements to subsystems and components. The following table illustrates how responsibilities for the risk controls can be assigned and how each function can be verified.
Design Element
Supporting Requirement
Test level
Verification Approach
Communication Interface
Specify how the controller detects missing messages, rejects outdated commands and identifies invalid data.
Controller subsystem integration testing
Simulate missing messages, outdated commands and invalid data, and verify that the controller detects or rejects them as specified.
Local Control Firmware
Detect the communication fault and initiate the protective response within the allocated time. Define the conditions for recovery.
Firmware unit testing;
Controller subsystem integration testing
Test timeout boundaries, fault-state transitions and recovery logic.
On the controller hardware, verify that the protective outputs are issued within the allocated time.
Drive & Holding Mechanism
Perform the specified stopping response and hold the load under defined operating and power conditions.
Drive and mechanical assembly testing
Measure stopping behaviour and load holding under the specified loads, configurations and power conditions.
User Interface
Notify the user of the fault and permit restart only when the specified recovery conditions are met.
UI software testing;
Subsystem integration testing
Verify fault indication and restart handling. Check that restart is blocked until the required conditions are satisfied.
Complete System
Reach and maintain the defined safe condition within specified time, movement and force limits.
System verification testing
Interrupt communication during representative and justified worst-case motions. Measure total time from communication loss to the defined safe state, residual movement, stopping forces and load holding. Restore communication and verify that motion can resume only when the specified recovery conditions are met.
Depending on the risks, motion monitoring or an emergency-stop function may also be needed to complement the automatic response to communication loss.
Subsystem testing verifies the individual functions. Complete-system verification demonstrates that they work together to produce the required physical response and maintain the defined safe condition.
What does this mean for development teams?
Reviewing the draft guidance early helps teams identify missing requirements, clarify interfaces and plan the evidence they need. Resolving these questions before the design is finalized can reduce late design changes, repeated verification and gaps discovered while preparing the submission.
For devices within the guidance’s scope, this approach can support a more complete and coherent premarket submission. That may reduce requests for additional information and help FDA review the submission more efficiently. The practical starting point is to ask which recommendations are relevant to your device, what they mean for the design and how you will demonstrate that the resulting requirements and risk controls have been met.
Source
FDA. Robotically-Assisted Surgical Devices – Premarket Submissions. Draft guidance issued 25 September 2026. https://www.fda.gov/media/194987/download