OPENTECHNICAL PUBLICATION
OT-2026-000000015

Dynamic Robot Stopping Contour for Speed and Separation Monitoring

Published: 2026-09-15T19:01:16.117552Z
Permanent URL: https://opentechnical.org/p/OT-2026-000000015
Categories: Robotics & Automation > Collaborative Robotics > Speed and Separation

------------------------------------------------------------------------

ABSTRACT

This disclosure describes a method and system for dynamic speed and separation
monitoring for robotic manipulators. It addresses the limitations of fixed
safety zones that either cause premature robot stops or necessitate
conservative sizing, leading to reduced operational throughput. The described
arrangement computes a predicted stopping contour based on real-time robot
kinematics, current speed, payload, and joint configuration. This contour
defines the dynamic safety boundary, allowing the robot to operate closer to
personnel while maintaining safety. The contour is continuously updated,
enabling adaptive speed control and improving operational efficiency in
collaborative environments. A defined fallback mechanism ensures safety
integrity if prediction data becomes unavailable.

1. Technical Field

[0001] The present disclosure relates to industrial automation and robotics,
specifically to safety systems designed for collaborative robot operation.
More particularly, it concerns methods and apparatus for monitoring speed and
separation between a robot manipulator and personnel within a shared
workspace.

2. Technical Context

[0002] In collaborative robotic applications, robots and human operators work
in proximity or share a workspace. To ensure human safety, regulatory
standards often define operational modes based on distance and speed.
Traditional approaches employ fixed safety zones around the robot or
workspace. If a human enters a predefined outer zone, the robot may slow down.
If a human enters a predefined inner zone, the robot is commanded to stop. The
size of these fixed zones is typically determined by the worst-case stopping
performance of the robot, considering maximum speed, maximum payload, and
worst-case joint configurations. This conservative sizing often results in
large safety zones, leading to premature robot slowdowns or stops, even when
the robot could safely continue operating or stop in a much smaller area. This
reduces the effective throughput and efficiency of the robotic system,
diminishing the benefits of collaboration. A need exists for a more adaptive
safety monitoring system that dynamically adjusts safety boundaries based on
real-time robot state.

3. System Overview

[0003] The disclosed system provides dynamic speed and separation monitoring
by computing a predicted stopping contour for a robot manipulator (18). This
contour represents the physical envelope the robot would occupy if it were to
initiate an emergency stop from its current state. By continuously calculating
and updating this contour, the system can determine safe separation distances
in real time, allowing the robot to operate closer to personnel than with
fixed safety zones, thereby improving operational efficiency. The system
comprises a human presence sensor (10), a robot controller (12), a prediction
module (14), and a safety controller (16). The human presence sensor (10)
provides spatial data regarding human location. The robot controller (12)
supplies real-time robot state information, including joint positions,
velocities, accelerations, and payload data. The prediction module (14) uses
this data to compute the predicted stopping contour. The safety controller
(16) receives both the predicted contour data and human presence data,
determines required robot actions, and issues speed commands to the robot
controller (12). This interaction is illustrated in FIG. 1.

4. Components

[0004] The system consists of several interconnected components:

[0005]
  - Human Presence Sensor (10): This component is responsible for detecting the presence and position of personnel within the shared workspace. It may comprise a single sensor or an array of sensors. Examples include lidar scanners, stereo vision cameras, 3D time-of-flight cameras, or radar sensors. The sensor (10) outputs spatial coordinates or occupancy grids indicating the location of humans relative to the robot's workspace. Its update rate is typically 10 Hz to 30 Hz, providing sufficient temporal resolution for human movement.
  - Robot Controller (12): This is the primary control unit for the robot manipulator (18). It manages the robot's motion, kinematics, and dynamics. The robot controller (12) provides real-time data streams including:
  - Current joint angles and velocities.
  - Current Cartesian pose and velocity of the end-effector (20).
  - Commanded joint torques or motor currents.
  - Robot configuration parameters, such as link lengths, masses, and inertia tensors.
  - Current payload mass and its center of gravity.
  - Operating mode and commanded speed limits.

[0006] The robot controller (12) also receives speed and motion commands from
the safety controller (16) and executes them.

[0007]
  - Prediction Module (14): This module is a computational unit responsible for calculating the predicted stopping contour (22). It receives robot state data from the robot controller (12) and may also receive data from the human presence sensor (10) for context. Its primary function is to simulate the robot's trajectory and physical extent during an emergency stop. This module typically runs on a dedicated processor or within the safety controller (16) itself.
  - Safety Controller (16): This is a safety-rated programmable logic controller (PLC) or a dedicated safety computation unit. It receives the predicted stopping contour (22) from the prediction module (14) and human presence data from the human presence sensor (10). It continuously compares the predicted contour (22) with the detected human presence (24) to determine if a minimum separation distance is maintained. Based on this comparison, it issues appropriate speed reduction or stop commands to the robot controller (12). The safety controller (16) operates according to defined safety functions and has a high level of integrity.
  - Robot Manipulator (18): This is the physical robotic arm, comprising multiple links and joints, designed to perform tasks within a workspace. It includes an end-effector (20) for interacting with the environment or workpieces.

5. Operation

[0008] The system operates continuously during collaborative tasks. The
operational sequence is as follows:

[0009] 1.## Technical Field The present disclosure relates to industrial
automation and robotics, specifically to safety systems designed for
collaborative robot operation. More particularly, it concerns methods and
apparatus for monitoring speed and separation between a robot manipulator and
personnel within a shared workspace.

6. Technical Context

[0010] In collaborative robotic applications, robots and human operators work
in proximity or share a workspace. To ensure human safety, regulatory
standards often define operational modes based on distance and speed.
Traditional approaches employ fixed safety zones around the robot or
workspace. If a human enters a predefined outer zone, the robot may slow down.
If a human enters a predefined inner zone, the robot is commanded to stop. The
size of these fixed zones is typically determined by the worst-case stopping
performance of the robot, considering maximum speed, maximum payload, and
worst-case joint configurations. This conservative sizing often results in
large safety zones, leading to premature robot slowdowns or stops, even when
the robot could safely continue operating or stop in a much smaller area. This
reduces the effective throughput and efficiency of the robotic system,
diminishing the benefits of collaboration. A need exists for a more adaptive
safety monitoring system that dynamically adjusts safety boundaries based on
real-time robot state.

7. System Overview

[0011] The disclosed system provides dynamic speed and separation monitoring
by computing a predicted stopping contour for a robot manipulator (18). This
contour represents the physical envelope the robot would occupy if it were to
initiate an emergency stop from its current state. By continuously calculating
and updating this contour, the system can determine safe separation distances
in real time, allowing the robot to operate closer to personnel than with
fixed safety zones, thereby improving operational efficiency. The system
comprises a human presence sensor (10), a robot controller (12), a prediction
module (14), and a safety controller (16). The human presence sensor (10)
provides spatial data regarding human location. The robot controller (12)
supplies real-time robot state information, including joint positions,
velocities, accelerations, and payload data. The prediction module (14) uses
this data to compute the predicted stopping contour. The safety controller
(16) receives both the predicted contour data and human presence data,
determines required robot actions, and issues speed commands to the robot
controller (12). This interaction is illustrated in FIG. 1.

8. Components

[0012] The system consists of several interconnected components:

[0013]
  - Human Presence Sensor (10): This component is responsible for detecting the presence and position of personnel within the shared workspace. It may comprise a single sensor or an array of sensors. Examples include lidar scanners, stereo vision cameras, 3D time-of-flight cameras, or radar sensors. The sensor (10) outputs spatial coordinates or occupancy grids indicating the location of humans relative to the robot's workspace. Its update rate is typically 10 Hz to 30 Hz, providing sufficient temporal resolution for human movement.
  - Robot Controller (12): This is the primary control unit for the robot manipulator (18). It manages the robot's motion, kinematics, and dynamics. The robot controller (12) provides real-time data streams including:
  - Current joint angles and velocities.
  - Current Cartesian pose and velocity of the end-effector (20).
  - Commanded joint torques or motor currents.
  - Robot configuration parameters, such as link lengths, masses, and inertia tensors.
  - Current payload mass and its center of gravity.
  - Operating mode and commanded speed limits.

[0014] The robot controller (12) also receives speed and motion commands from
the safety controller (16) and executes them.

[0015]
  - Prediction Module (14): This module is a computational unit responsible for calculating the predicted stopping contour (22). It receives robot state data from the robot controller (12) and may also receive data from the human presence sensor (10) for context. Its primary function is to simulate the robot's trajectory and physical extent during an emergency stop. This module typically runs on a dedicated processor or within the safety controller (16) itself.
  - Safety Controller (16): This is a safety-rated programmable logic controller (PLC) or a dedicated safety computation unit. It receives the predicted stopping contour (22) from the prediction module (14) and human presence data from the human presence sensor (10). It continuously compares the predicted contour (22) with the detected human presence (24) to determine if a minimum separation distance is maintained. Based on this comparison, it issues appropriate speed reduction or stop commands to the robot controller (12). The safety controller (16) operates according to defined safety functions and has a high level of integrity.
  - Robot Manipulator (18): This is the physical robotic arm, comprising multiple links and joints, designed to perform tasks within a workspace. It includes an end-effector (20) for interacting with the environment or workpieces.

9. Operation

[0016] The system operates continuously during collaborative tasks. The
operational sequence is as follows:

[0017] Start-up: Upon system initialization, the robot controller (12)
performs self-diagnostics. The human presence sensor (10) is activated and
begins monitoring the workspace. The prediction module (14) initializes its
kinematic and dynamic models of the robot. The safety controller (16) enters a
safe state, typically preventing robot motion until all safety functions are
verified. Once verified, the robot may be enabled for motion at a safe,
reduced speed or remain stationary depending on the initial state of the
workspace.

[0018] Steady Running:

[0019]
  - Data Acquisition: The robot controller (12) continuously provides real-time robot state data, including joint positions, velocities, and payload information, to the prediction module (14). Simultaneously, the human presence sensor (10) provides updated spatial data on human presence (24) within the workspace to the safety controller (16).
  - Contour Computation: The prediction module (14) uses the received robot state data to compute the predicted stopping contour (22). This computation involves:
  - Kinematic Model: Accessing the robot's kinematic model to determine the spatial extent of all robot links and the end-effector (20) for the current joint configuration.
  - Dynamic Model: Utilizing the robot's dynamic model, including joint inertias, friction parameters, and motor torque limits, to simulate the deceleration profile of each joint under a stop command. The current payload mass and its center of gravity are incorporated to accurately model inertia.
  - Stopping Trajectory Prediction: For each joint, the module predicts the path and final position if an emergency stop were initiated. This accounts for the time delay from detection to brake engagement, and the deceleration time due to motor braking and friction.
  - Contour Generation: The individual predicted stopping trajectories of all robot links and the end-effector (20) are combined to form a three-dimensional envelope, which is the predicted stopping contour (22). This contour represents the maximum physical space the robot would sweep through and occupy during a stop.
  - Update Rate: The predicted stopping contour (22) is recomputed and updated at a rate suitable for dynamic environments, typically between 50 Hz and 100 Hz. This frequency ensures that changes in robot speed, configuration, and payload are rapidly reflected in the safety boundary.
  - Separation Monitoring: The safety controller (16) receives the updated predicted stopping contour (22) from the prediction module (14) and the human presence data (24) from the human presence sensor (10). It then calculates the minimum separation distance between any point on the predicted stopping contour (22) and any detected part of the human (24). This is illustrated in FIG. 2.
  - Speed Adjustment: Based on the calculated separation distance and predefined safety parameters (e.g., minimum permissible separation distance, safety margins), the safety controller (16) determines the appropriate robot speed. If the separation distance decreases and approaches a threshold, the safety controller (16) commands the robot controller (12) to reduce the robot's speed. If the separation distance falls below a critical threshold, a monitored stop or emergency stop command is issued.

[0020] Fault or Interruption:

[0021]
  - Prediction Unavailable: If the prediction module (14) fails, or if data from the robot controller (12) required for contour computation becomes unavailable or untrustworthy, the system defaults to a safe state. This fallback mechanism typically involves the safety controller (16) issuing an immediate monitored stop command to the robot controller (12). The robot then decelerates to a standstill. Alternatively, a pre-defined worst-case fixed safety zone, sized for maximum speed and payload, may be engaged, forcing the robot to operate with significantly increased separation distances or to stop if a human is within this conservative zone.
  - Sensor Failure: If the human presence sensor (10) fails or reports invalid data, the safety controller (16) will command a stop. In some configurations, redundant sensing or a default maximum robot speed limit may be invoked until the sensor is restored.
  - Safety Controller Malfunction: If the safety controller (16) detects an internal fault or loss of communication, it will immediately issue a safe stop command to the robot controller (12), typically an emergency stop, bringing the robot to a halt.

[0022] Relationship to Monitored Stop: The predicted stopping contour (22) is
intrinsically linked to the concept of a monitored stop (often referred to as
a Safe Monitored Stop, SMS). When the safety controller (16) determines that
the predicted stopping contour (22) is about to intersect with the detected
human presence (24) within a defined safety margin, it initiates a monitored
stop. This means the robot's motion is controlled to bring it to a standstill
safely and predictably, typically by decelerating along its current path. The
system continually monitors the actual stopping performance against the
predicted contour. If the robot's actual stopping behavior deviates from the
predicted behavior in an unsafe manner (e.g., it travels further than
predicted), the safety controller (16) would escalate the response,
potentially to an emergency stop.

10. Example Embodiments

[0023] Embodiment 1 In this embodiment, the prediction module (14) is
implemented as a software component running on the robot controller (12)
itself. The robot controller (12) has sufficient processing power and safety
integrity level to perform both motion control and real-time stopping contour
calculations. The human presence sensor (10) is a single 3D lidar scanner
positioned to cover the entire shared workspace. The safety controller (16) is
integrated as a safety PLC within the robot controller (12), communicating
directly with the motion control kernel. The predicted stopping contour (22)
is represented as a series of bounding boxes around each robot link,
calculated at each control cycle.

[0024] Embodiment 2 This embodiment utilizes a dedicated external computation
unit for the prediction module (14). This unit receives robot state data via a
high-speed industrial Ethernet connection from the robot controller (12). The
human presence sensor (10) consists of multiple overhead stereo vision cameras
providing a dense point cloud of the workspace. The safety controller (16) is
a standalone safety relay system that receives safety-rated signals from the
external prediction module (14) indicating safe/unsafe separation and issues
hardwired stop commands to the robot controller (12) via safety relays. The
predicted stopping contour (22) is generated as a detailed triangulated mesh,
offering higher geometric precision.

[0025] Embodiment 3 In this embodiment, the prediction module (14) and safety
controller (16) are combined into a single safety-certified hardware module.
This module directly interfaces with the robot's drive units to read motor
encoder feedback and issue direct torque commands for deceleration. The human
presence sensor (10) employs radar technology, capable of operating
effectively in dusty or low-light environments. The robot controller (12)
primarily handles path planning and trajectory generation, while the safety
module (14, 16) manages all real-time safety-critical functions, including the
dynamic stopping contour prediction and safety supervision.

11. Alternative Configurations

[0026]
  - Sensor Fusion for Human Presence: Instead of a single sensor type, multiple types of human presence sensors (10) could be fused (e.g., lidar and vision) to improve robustness and accuracy of human detection, especially in environments with occlusions or varying lighting conditions.
  - Predictive Human Motion: The system could incorporate a module that predicts human motion, not just current position. This would allow for more proactive robot speed adjustments and potentially larger safety margins for predicted human paths, further enhancing safety without necessarily reducing throughput.
  - Learning-Based Prediction Models: The dynamic and kinematic models used by the prediction module (14) could be augmented or replaced by machine learning models trained on extensive robot stopping data to refine prediction accuracy under various load and environmental conditions.
  - Adjustable Safety Margins: The safety controller (16) could implement dynamic safety margins that adapt based on the confidence level of the sensor data, the predictability of human movement, or the specific task being performed, allowing for more flexible operation.

12. Parameter Variations

[0027]
  Parameter | Realistic Range | Effect of Increasing Value | Effect of Decreasing Value
  Prediction Module Update Rate | 50 Hz to 200 Hz | Faster response to robot state changes; higher CPU load. | Slower response; contour lags real state; larger safety margins needed.
  Human Presence Sensor Latency | 20 ms to 100 ms | Larger safety margins required due to delayed detection. | Smaller safety margins possible; more reactive system.
  Robot Max Payload | 0.5 kg to 50 kg | Longer stopping distances; larger predicted contour. | Shorter stopping distances; smaller predicted contour.
  Robot Max Joint Speed | 100 deg/s to 360 deg/s | Longer stopping distances; larger predicted contour. | Shorter stopping distances; smaller predicted contour.
  Safety Margin Distance | 50 mm to 300 mm | Increased separation; more conservative operation. | Reduced separation; higher throughput; less conservative.
  Robot Braking Deceleration | 5 rad/s² to 20 rad/s² | Shorter stopping time/distance; higher mechanical stress. | Longer stopping time/distance; smoother stops.

13. Additional Implementations

[0028]
  - Mobile Manipulators: The described system can be extended to mobile manipulator platforms. In this case, the prediction module (14) would also account for the stopping characteristics of the mobile base, including its velocity, braking capabilities, and potential for slippage on the floor surface. The predicted stopping contour (22) would then encompass both the manipulator and the mobile platform.
  - Multi-Robot Systems: For systems involving multiple collaborative robots sharing a workspace, each robot could have its own prediction module (14) and safety controller (16), or a central safety controller (16) could coordinate the predicted stopping contours of all robots. This allows for dynamic collision avoidance between robots as well as between robots and humans.
  - Virtual Fixtures and Haptic Feedback: The predicted stopping contour (22) can be used to generate virtual fixtures that guide human operators away from unsafe regions or to provide haptic feedback through a human-robot interface, warning the operator when they are approaching the robot's dynamic safety boundary.
  - Process Optimization: Beyond safety, the real-time separation data derived from the predicted stopping contour (22) can be used to optimize robot task execution. For example, the robot could automatically increase its speed when sufficient separation from personnel is detected, or adjust its path to maintain maximum throughput while respecting safety constraints.

14. Figures

FIG. 1
    This block diagram illustrates the interconnections between system
    components for generating and utilizing a dynamic stopping contour.

FIG. 2
    This elevation view depicts a robot manipulator operating within a
    workspace, showing a dynamically predicted stopping contour.

FIGURES

FIG. 1  This block diagram illustrates the interconnections between system components for generating and utilizing a dynamic stopping contour.
FIG. 2  This elevation view depicts a robot manipulator operating within a workspace, showing a dynamically predicted stopping contour.

------------------------------------------------------------------------

OpenTechnical publication OT-2026-000000015
Published 2026-09-15T19:01:16.117552Z
https://opentechnical.org/p/OT-2026-000000015