Benchmark RadarJITOMA-Bench
JITOMA-Bench is a suite for long-horizon multi-tasking and multi-step reasoning in robotics, focusing on just-in-time scene graph growth to combat perceptual saturation.
- Released
- 2026-07-14
- Readiness
- Paper only
- Primary field
- Robotics & Autonomous Systems
Why it matters
The benchmark supports the JITOMA framework's evaluation, but its primary purpose is to demonstrate the framework's advantages rather than serve as a standalone comparison benchmark.
Motivation
While 3D Scene Graphs (3DSGs) provide crucial structured representations for embodied agents, conventional Ahead-of-Time, build-everything-then-filter pipelines conflict with the real-time, low-latency demands of edge platforms, inducing a perceptual saturation effect via severe observation redundancy.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.