Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Robot Safety Monitoring using Programmable Light Curtains

Karnik Ram, Shobhit Aggarwal, Robert Tamburo, Siddharth Ancha, Srinivasa Narasimhan

TL;DR

This work tackles safety in human–robot collaborative manufacturing by replacing fixed, costly laser curtains with programmable light curtains (PLCs) that adaptively envelop robots in real time. The authors develop an optimization-based instrumentation method to place multiple PLCs for maximal robot-coverage, design dynamic safety curtains that tightly follow robot motion, and use curtain sweeps to generate high-resolution 3D scene reconstructions. Key contributions include a RANSAC-like PLC placement algorithm with exponential brute-force comparison avoided, real-time curtain construction via convex hulls and ray-tracing, and an intrusion-detection pipeline that links detections to specific robots with persistence-based safety gating. In a four-robot testbed, the system demonstrates fast, accurate intrusion detection and scalable coverage with few PLCs at a fraction of traditional safety-system costs, enabling fence-less, safe collaboration in industrial settings.

Abstract

As factories continue to evolve into collaborative spaces with multiple robots working together with human supervisors in the loop, ensuring safety for all actors involved becomes critical. Currently, laser-based light curtain sensors are widely used in factories for safety monitoring. While these conventional safety sensors meet high accuracy standards, they are difficult to reconfigure and can only monitor a fixed user-defined region of space. Furthermore, they are typically expensive. Instead, we leverage a controllable depth sensor, programmable light curtains (PLC), to develop an inexpensive and flexible real-time safety monitoring system for collaborative robot workspaces. Our system projects virtual dynamic safety envelopes that tightly envelop the moving robot at all times and detect any objects that intrude the envelope. Furthermore, we develop an instrumentation algorithm that optimally places (multiple) PLCs in a workspace to maximize the visibility coverage of robots. Our work enables fence-less human-robot collaboration, while scaling to monitor multiple robots with few sensors. We analyze our system in a real manufacturing testbed with four robot arms and demonstrate its capabilities as a fast, accurate, and inexpensive safety monitoring solution.

Robot Safety Monitoring using Programmable Light Curtains

TL;DR

This work tackles safety in human–robot collaborative manufacturing by replacing fixed, costly laser curtains with programmable light curtains (PLCs) that adaptively envelop robots in real time. The authors develop an optimization-based instrumentation method to place multiple PLCs for maximal robot-coverage, design dynamic safety curtains that tightly follow robot motion, and use curtain sweeps to generate high-resolution 3D scene reconstructions. Key contributions include a RANSAC-like PLC placement algorithm with exponential brute-force comparison avoided, real-time curtain construction via convex hulls and ray-tracing, and an intrusion-detection pipeline that links detections to specific robots with persistence-based safety gating. In a four-robot testbed, the system demonstrates fast, accurate intrusion detection and scalable coverage with few PLCs at a fraction of traditional safety-system costs, enabling fence-less, safe collaboration in industrial settings.

Abstract

As factories continue to evolve into collaborative spaces with multiple robots working together with human supervisors in the loop, ensuring safety for all actors involved becomes critical. Currently, laser-based light curtain sensors are widely used in factories for safety monitoring. While these conventional safety sensors meet high accuracy standards, they are difficult to reconfigure and can only monitor a fixed user-defined region of space. Furthermore, they are typically expensive. Instead, we leverage a controllable depth sensor, programmable light curtains (PLC), to develop an inexpensive and flexible real-time safety monitoring system for collaborative robot workspaces. Our system projects virtual dynamic safety envelopes that tightly envelop the moving robot at all times and detect any objects that intrude the envelope. Furthermore, we develop an instrumentation algorithm that optimally places (multiple) PLCs in a workspace to maximize the visibility coverage of robots. Our work enables fence-less human-robot collaboration, while scaling to monitor multiple robots with few sensors. We analyze our system in a real manufacturing testbed with four robot arms and demonstrate its capabilities as a fast, accurate, and inexpensive safety monitoring solution.
Paper Structure (14 sections, 10 figures, 3 tables, 1 algorithm)

This paper contains 14 sections, 10 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Robot Safety Monitoring
  4. Programmable Light Curtains
  5. Background on Prog. Light Curtains
  6. Method
  7. PLC instrumentation to maximize robot coverage
  8. Safety curtain design
  9. Intrusion Detection
  10. Scene Reconstruction
  11. Experiments
  12. Setup
  13. Results
  14. Conclusion

Figures (10)

  • Figure 1: Overview. (a) A human worker walks into a robot workspace to adjust the work-piece on its end-effector. (b) A programmble light curtain (PLC) forms a tight convex hull enveloping the two robots. This acts as a flexible "safety shield" for collision monitoring that adapts to the robot's motion in real-time. (c) The intrusion detected by 3D points on the light curtain intersecting the obstacle are shown in green. This triggers the robot to stop. (d) The full intensity image captured by the PLC shows the outline of the detected worker.
  • Figure 2: (a) Our PLC prototype consists of a near-infrared (NIR) light sheet laser reflected by a rotating galvomirror, an NIR rolling shutter camera, and an additional RGB helper camera for visualization. (b) The light sheet laser rotates in synchrony with the rays of the rolling shutter camera. Only the line (green) at the intersection of the illumination plane (red) from the laser and the imaging plane (blue) of the camera is sensed at a given instant. By controlling the rotation of the laser, the green line can be made to follow a user-specified surface. Therefore, the PLC is able exclusively image 3D points on the user-specified surface at a high resolution. Figures adapted from bartels2019agileancha2020eccv.
  • Figure 3: Optimized configuration of PLCs (red circles) for an $8m \times 8m$ layout containing three robots (green squares). Our algorithm computes the configuration that maximizes the sum of the angles subtended by the robots at the PLCs (rays denoted by ---, field-of-view denoted by ---) over many samples of joint PLC configurations.
  • Figure 4: Testbed area used in our experiments.
  • Figure 5: Network setup used in the testbed. The PLC sensors are connected to the robots over ROS using Ethernet cables. The robots are monitored and controlled for safety directly by the PLC sensors. The workstation is used for visualization and additional control.
  • ...and 5 more figures