NKK Switches in Automation: AMRs, Teach Pendants, Cobots and Humanoids

<strong>NKK Switches in Automation: AMRs, Teach Pendants, Cobots and Humanoids</strong>

Emergency Stops in Real-World Robotic Environments

January | February 2026

Robots operating outside traditional industrial cells demand a different approach to emergency stop design. When systems work in close proximity to people, whether navigating shared environments, assisting in care settings, performing tasks alongside human operators, or enabling emerging humanoid interaction models, stop authority and interface controls must be immediate, visually clear, and physically integrated into the local interface.

This spotlight highlights how NKK emergency stop solutions are designed into a range of people-adjacent robotic applications. The examples span autonomous mobile robots in warehousing and logistics, patient-facing cobots, autonomous vehicles used in transportation and agriculture, and emerging humanoid platforms that combine a dedicated E-Stop with complementary interface elements such as YB2 Pushbuttons and Indicators. Across each use case, the E-Stop is specified early in the design process to balance safety requirements, interface clarity, and proportional integration within increasingly refined system enclosures.

For a quick overview and video, visit our Emergency Stop Switches page.

Red emergency stop pushbutton shown alongside the headline “E-Stop in Robotics: AMR Placement,” with subtext referencing placement tied to risk and real approach paths, presented on a light gray technical background.

Autonomous mobile robots (AMRs) are no longer isolated machines; they are nodes in human-centric, software-orchestrated environments where AI, computer vision, and digital twins coordinate fleets alongside conveyors, lifts, and manual work. In that context, safety is a design requirement, and E-Stop placement is one of the clearest expressions of that intent.

No Universal Layout

Risk sets the pattern

Current AMR guidance requires an E-stop on each vehicle, but it does not define a universal layout. Placement is driven by vehicle geometry, approach paths, and application hazards. A risk assessment is the most reliable way to translate these variables into device count and location.

Landscape infographic

Strong E-Stop Placement

A simple rule of thumb

Many platforms follow a simple E-Stop pattern: low profile latching switches at opposite ends of the AMR. This geometry-based approach keeps a physical stop within arm’s length for people working alongside the robot.Rule of Thumb: if someone must bend, reach across a load zone, or step into the AMR’s path to press the button, the placement likely needs revision.

Landscape infographic

See the Top 20 AMR Archetypes infographic for a quick snapshot of the market landscape, including several that feature NKK E-Stops. Across this category, NKK is specified on platforms where teams prioritize safety interfaces.

FF01: Built for AMRs in Motion

Shock and vibration performance

AMRs live with constant vibration, impact risk, and shifting payloads, so that the E-Stop has to behave as a stable safety input. Rated for 10–500Hz vibration (amplitude 0.35mm, acceleration 50m/s²) and 1,000m/s² shock (150m/s² malfunction), NKK’s E-Stop supports mobile platforms where weakness in the safety chain is unacceptable.

FF01: Low Profile E-Stop by Design

Ready for dual channel safety

DPST normally closed contacts support dual channel safety, while the Ø16mm body with 13.6mm behind panel depth fits recessed AMR sidewalls and other tight designs. NKK’s focus on quality and stable field performance makes our E-Stops deliberate safety-path choices.

For additional system-level context on designing autonomous mobile robots, including sensing, navigation, and hardware architecture considerations, see Qualcomm’s technical overview:

Autonomous Mobile Robots: What Do I Need to Know to Design One?

 

Annotated diagram of a 30 mm emergency stop actuator showing key features and callouts.

Close-up of a red emergency stop pushbutton mounted on a panel, shown as a physical hardware example.

Red emergency stop pushbutton shown alongside the headline “E-Stop in Robotics: AMR Placement,” with subtext referencing placement tied to risk and real approach paths, presented on a light gray technical background.

In patient facing cobots, local stop authority is a human-factors requirement at the point of service. Collision sensing, force limiting, and safety-rated stop functions reduce risk, but they do not eliminate the need for a direct, local E-Stop. Because the operator station is the care zone, the E-Stop must live in that same visual and physical field, with zero search time and no dependence on a pendant.

Control Stack Evolution

Closed-loop force control

Patient-contact platforms are shifting from scripted motion to sensor-feedback, closed-loop contact control, where perception and force sensing continuously regulate interaction. As autonomy increases during physical interaction, emergency stop access must remain intuitive: obvious at a glance and immediately actionable.

Design-In Context

Body mapping drives the session

One robotics team has developed an on-demand, fully automated massage robot built around two cobot arms. Before each session, the system performs an overhead infrared body scan to generate a 3D model with over 1.1 million data points, which the platform uses to locate the body, identify target areas, and generate repeatable robotic stroke paths tailored to that individual.

Head-end Stop Authority

At the point of care

Users control the session from a head-end touch screen, selecting a program and adjusting pressure in real time. The platform pairs 3D body mapping with user preferences to execute targeted contact paths, detects unexpected interaction to automatically disengage and reset, and uses AI and machine learning on saved profile history to refine personalization over time.

Minimal Interface & Clear Stop

NKK’s no legend actuator and low profile design

Why the switch fits: The OEM required immediate, unambiguous stop authority at the head-end operator station without disturbing a minimal, calming interface. They specified NKK’s E-Stop with a 25mm diameter, no legend actuator, and the 13.6mm behind panel depth made integration feasible in a shallow enclosure. The actuator’s form and placement still read clearly as an emergency stop.

Diagram showing the process of creating a 3D body model.
Diagram showing the process of generating a session plan.

Patient-Contact Control Stack Shifts

As physical interaction becomes more software-controlled, more decisions move into perception and control algorithms. Emergency stop authority must stay local,
immediate and hardware-based.

3D Body Model:
Overhead vision and table-integrated sensors generate a body reference used to define target regions and define constraint limits.

Model-Based Planning:
Trajectory planning is relative to the body model, not a fixed coordinate frame.

Closed-Loop Contact:
Force feedback regulates contact pressure in real time during interaction.

Personalization:
Settings persist and are reused to fine-tune future sessions.

An image of NKK's low profile, short behind panel design with NC latchdown.
Patient facing interface with touch screen and integrated E-Stop 25mm actuator with no legends.

Banner image reading “Autonomous Driving Systems – E-Stop in Supervised Testing” with the subtitle “Preserving human stop authority during public-road ADS testing,” alongside images of red emergency stop pushbutton switches.

Public-road automated driving systems (ADS) testing places humans and automated systems in the same operating envelope. With perception, planning, and motion executing live, intervention must remain immediate, deterministic, and independent of software state.

Human-Facing Control

Road oversight and real-time logging

During supervised ADS testing and mapping of geofences, operators monitor perception and planning through a touch screen teach pendant. One person maintains situational awareness of the road environment while another logs behavior and anomalies in real time. A safety driver disengages when needed, and the engineer captures deviations from intended behavior. Each intervention is logged to improve system performance.

An E-Stop on the teach pendant preserves immediate stop authority.

See the visual overview of robotaxi testing and safety validation process.

Accountability in Practice

Testing data is safety evidence

NHTSA does not approve ADS testing in advance, but it does impose accountability through mandatory ADS crash reporting under Standing General Order 2021-01. Robust event logging and clear intervention markers are foundational to incident review and root-cause analysis. For additional context, NHTSA outlines this approach to automated driving system safety in its voluntary guidance for ADS development and testing.

Why the Switch Fits

FF01 Series: Designed for a pendant interface

One robotaxi developer specifies NKK’s FF01 Series E-Stop for a human-facing, touch screen test controller used during supervised ADS road testing and geofence mapping.

Limited panel space and shallow behind panel clearance made integration challenging, so the Ø 25mm actuator and short behind panel depth fit cleanly alongside the touch screen HMI.

Technical drawing labeled “Typical Switch Dimensions” showing side and rear views of a DPST emergency stop switch with a 25.0 mm cap, including M16 P1 threaded bushing, anti-rotation tab, terminal layout, and dimensional callouts in millimeters and inches.

The switch is supplied with a custom pre-assembled, ready to install wire harness, and the DPST configuration provides redundant stop circuits for vehicle stop authority and system monitoring.

Shock resistant internal construction helps mitigate vibration and contact chattering, while a sliding latch mechanism mechanically maintains the contacts in the OFF state after actuation.

Close-up of a teach pendant showing a touchscreen interface with autonomy status indicators and a red emergency stop pushbutton mounted on the side of the housing.
Concept image of a dedicated autonomous vehicle platform showing a sensor-equipped vehicle with 360-degree awareness visualization and an interior cabin without traditional driver controls.