Online Hamilton-Jacobi Reachability Onboard the Robot Control Loop: A Real-Time GPU Solver for Safety-Critical Autonomy
Overview
This paper introduces a CUDA implementation of a canonical Hamilton-Jacobi-Isaacs solver capable of computing backward reachable tubes (BRTs) for 2D, 3D, and 4D dynamics fast enough to run inside a robot's control loop. On a single NVIDIA RTX 4000 Ada GPU using FP32, the system computes a 3D Air3D BRT at 112 Hz, significantly outperforming JAX and PyTorch reference implementations. The method also demonstrates compatibility with embedded hardware like the NVIDIA Jetson AGX Thor and includes an extension for multi-agent capture-basin tessellation in wildfire-response simulations.
Generated from attributed reports · Updated 1 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Control: reported autonomous
“Online Hamilton-Jacobi Reachability Onboard the Robot Control Loop: A Real-Time GPU Solver for Safety-Critical Autonomy”
Exact source · revision 1Source owner not reported
Reported quantity · Hz: 112 other · Basis not reported · Differing source assertions
“licies, a deterministic safety layer with formal guarantees is an essential complement. We present a CUDA implementation of the canonical WENO5 + TVD-RK3 + Lax-Friedrichs level-set Hamilton-Jacobi-Isaacs solver for 2D, 3D, and 4D dynamics that is fast enough to run inside a robot's control loop. In 32-bit floating point (FP32) on a single workstation GPU (NVIDIA RTX 4000 Ada), it computes the 3D Air3D BRT at 112 Hz,”
Exact source · revision 1Source owner not reported
Reported quantity · Hz: 66 other · Basis not reported · Differing source assertions
“x faster. On an embedded NVIDIA Jetson AGX Thor it computes Air3D at 66 Hz (N=41) and the 2D drifter at 54 Hz, 3.8-8.9x faster than hj_reachability on the same device. As a multi-agent application, an in-kernel label-propagation extension recovers the K-agent capture-basin Voronoi tessellation from a single combined HJ solve, independent of fleet size; it runs at 4.0 ms per tick in a closed-loop wildfire-response sim”
Exact source · revision 1Source owner not reported
Reported quantity · Hz: 54 other · Basis not reported · Differing source assertions
“x faster. On an embedded NVIDIA Jetson AGX Thor it computes Air3D at 66 Hz (N=41) and the 2D drifter at 54 Hz, 3.8-8.9x faster than hj_reachability on the same device. As a multi-agent application, an in-kernel label-propagation extension recovers the K-agent capture-basin Voronoi tessellation from a single combined HJ solve, independent of fleet size; it runs at 4.0 ms per tick in a closed-loop wildfire-response sim”
Exact source · revision 1Source owner not reported
Reported quantity · ms: 4 other · Basis not reported
“x faster. On an embedded NVIDIA Jetson AGX Thor it computes Air3D at 66 Hz (N=41) and the 2D drifter at 54 Hz, 3.8-8.9x faster than hj_reachability on the same device. As a multi-agent application, an in-kernel label-propagation extension recovers the K-agent capture-basin Voronoi tessellation from a single combined HJ solve, independent of fleet size; it runs at 4.0 ms per tick in a closed-loop wildfire-response sim”
Exact source · revision 1Source owner not reported
Reported quantity · percent: 0.04 other · Basis not reported
“a 4D pursuit-evasion BRT at 61 Hz, and a 2D wind-affected drifter BRT at 74-147 Hz, 5.9-17.5x faster than the JAX reference hj_reachability at matched precision and 1.7-16x faster than a torch.compile reimplementation of the same scheme. In 64-bit floating point (FP64) it agrees with hj_reachability on eight benchmark cases, reproduces Mitchell's Air3D BRT to within 0.04% of the reference cell count, and runs 2.0-4.6”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalOnline Hamilton-Jacobi Reachability Onboard the Robot Control Loop: A Real-Time GPU Solver for Safety-Critical Autonomy
This paper introduces a CUDA implementation of a canonical Hamilton-Jacobi-Isaacs solver capable of computing backward reachable tubes (BRTs) for 2D, 3D, and 4D dynamics fast enough to run inside a robot's control loop. On a single NVIDIA RTX 4000 Ada GPU using FP32, the system computes a 3D Air3D BRT at 112 Hz, significantly outperforming JAX and PyTorch reference implementations. The method also demonstrates compatibility with embedded hardware like the NVIDIA Jetson AGX Thor and includes an extension for multi-agent capture-basin tessellation in wildfire-response simulations.
Event coverage history
There is not enough continuous observation data to show a trend.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).